Patent Application
20140348258
1. Field
The present invention relates generally to data communication, and more specifically to a multiple-input multiple-output (MIMO) wireless local area network (WLAN) communication system.
2. Background
Wireless communication systems are widely deployed to provide various types of communication such as voice, packet data, and so on. These systems may be multiple-access systems capable of supporting communication with multiple users sequentially or simultaneously by sharing the available system resources. Examples of multiple-access systems include Code Division Multiple Access (CDMA) systems, Time Division Multiple Access (TDMA) systems, and Frequency Division Multiple Access (FDMA) systems.
Wireless local area networks (WLANs) are also widely deployed to enable communication among wireless electronic devices (e.g., computers) via wireless link. A WLAN may employ access points (or base stations) that act like hubs and provide connectivity for the wireless devices. The access points may also connect (or “bridge”) the WLAN to wired LANs, thus allowing the wireless devices access to LAN resources.
In a wireless communication system, a radio frequency (RF) modulated signal from a transmitter unit may reach a receiver unit via a number of propagation paths. The characteristics of the propagation paths typically vary over time due to a number of factors, such as fading and multipath. To provide diversity against deleterious path effects and improve performance, multiple transmit and receive antennas may be used. If the propagation paths between the transmit and receive antennas are linearly independent (i.e., a transmission on one path is not formed as a linear combination of the transmissions on the other paths), which is generally true to at least an extent, then the likelihood of correctly receiving a data transmission increases as the number of antennas increases. Generally, diversity increases and performance improves as the number of transmit and receive antennas increases.
A MIMO system employs multiple (NT) transmit antennas and multiple (NR) receive antennas for data transmission. A MIMO channel formed by the NT transmit and NR receive antennas may be decomposed into NS spatial channels, with NS≦min{NT, NR}. Each of the NS spatial channels corresponds to a dimension. The MIMO system can provide improved performance (e.g., increased transmission capacity and/or greater reliability) if the additional dimensionalities created by the multiple transmit and receive antennas are utilized.
The resources for a given communication system are typically limited by various regulatory constraints and requirements and by other practical considerations. However, the system may be required to support a number of terminals, provide various services, achieve certain performance goals, and so on.
There is, therefore, a need in the art for a MIMO WLAN system capable of supporting multiple users and providing high system performance.
A multiple-access MIMO WLAN system having various capabilities and able to achieve high performance is described herein. In an embodiment, the system employs MIMO and orthogonal frequency division multiplexing (OFDM) to attain high throughput, combat deleterious path effects, and provide other benefits. Each access point in the system can support multiple user terminals. The allocation of downlink and uplink resources is dependent on the requirements of the user terminals, the channel conditions, and other factors.
A channel structure supporting efficient downlink and uplink transmissions is also provided herein. The channel structure comprises a number of transport channels that may be used for a number of functions, such as signaling of system parameters and resource assignments, downlink and uplink data transmissions, random access of the system, and so on. Various attributes of these transport channels are configurable, which allows the system to easily adapt to changing channel and loading conditions.
Multiple rates and transmission modes are supported by the MIMO WLAN system to attain high throughput when supported by the channel conditions and the capabilities of the user terminals. The rates are configurable based on estimates of the channel conditions and may be independently selected for the downlink and uplink. Different transmission modes may also be used, depending on the number of antennas at the user terminals and the channel conditions. Each transmission mode is associated with different spatial processing at the transmitter and receiver and may be selected for use under different operating conditions. The spatial processing facilitates data transmission from multiple transmit antennas and/or data reception with multiple receive antennas for higher throughput and/or diversity.
In an embodiment, the MIMO WLAN system uses a single frequency band for both the downlink and uplink, which share the same operating band using time division duplexing (TDD). For a TDD system, the downlink and uplink channel responses are reciprocal. Calibration techniques are provided herein to determine and account for differences in the frequency responses of the transmit/receive chains at the access point and user terminals. Techniques are also described herein to simplify the spatial processing at the access point and user terminals by taking advantage of the reciprocal nature of the downlink and uplink and the calibration.
A pilot structure with several types of pilot used for different functions is also provided. For example, a beacon pilot may be used for frequency and system acquisition, a MIMO pilot may be used for channel estimation, a steered reference (i.e., a steered pilot) may be used for improved channel estimation, and a carrier pilot may be used for phase tracking.
Various control loops for proper system operation are also provided. Rate control may be exercised independently on the downlink and uplink. Power control may be exercised for certain transmissions (e.g., fixed-rate services). Timing control may be used for uplink transmissions to account for different propagation delays of user terminals located throughout the system.
Random access techniques to allow user terminals to access the system are also provided. These techniques support system access by multiple user terminals, fast acknowledgment of system access attempts, and quick assignment of downlink/uplink resources.
The various aspects and embodiments of the invention are described in further detail below.
The features and nature of the present invention will become more apparent from the detailed description set forth below when taken in conjunction with the drawings in which like reference characters identify correspondingly throughout and wherein:
The word “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any embodiment or design described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments or designs.
User terminals 120 may be dispersed throughout the system. Each user terminal may be a fixed or mobile terminal that can communicate with the access point. A user terminal may also be referred to as a mobile station, a remote station, an access terminal, a user equipment (UE), a wireless device, or some other terminology. Each user terminal may communicate with one or possibly multiple access points on the downlink and/or uplink at any given moment. The downlink (i.e., forward link) refers to transmission from the access point to the user terminal, and the uplink (i.e., reverse link) refers to transmission from the user terminal to the access point.
In
The MIMO WLAN system is based on a centralized controller network architecture. Thus, a system controller 130 couples to access points 110 and may further couple to other systems and networks. For example, system controller 130 may couple to a packet data network (PDN), a wired local area network (LAN), a wide area network (WAN), the Internet, a public switched telephone network (PSTN), a cellular communication network, and so on. System controller 130 may be designed to perform a number of functions such as (1) coordination and control for the access points coupled to it, (2) routing of data among these access points, (3) access and control of communication with the user terminals served by these access points, and so on.
The MIMO WLAN system may be able to provide high throughput with significantly greater coverage capability than conventional WLAN systems. The MIMO WLAN system can support synchronous, asynchronous, and isochronous data/voice services. The MIMO WLAN system may be designed to provide the following features:
The MIMO WLAN system may be operated in various frequency bands (e.g., the 2.4 GHz and 5.x GHz U-NII bands), subject to the bandwidth and emission constraints specific to the selected operating band. The system is designed to support both indoor and outdoor deployments, with typical maximum cell size of 1 km or less. The system supports fixed terminal applications, although some operating modes also support portable and limited mobility operation.
MIMO, MISO, and SIMO
In a specific embodiment and as described throughout the specification, each access point is equipped with four transmit and receive antennas for data transmission and reception, where the same four antennas are used to transmit and to receive. The system also supports the case where the transmit and receive antennas of the device (e.g. access point, user terminal) are not shared, even though this configuration normally provides lower performance than when the antennas are shared. The MIMO WLAN system may also be designed such that each access point is equipped with some other number of transmit/receive antennas. Each user terminal may be equipped with a single transmit/receive antenna or multiple transmit/receive antennas for data transmission and reception. The number of antennas employed by each user terminal type may be dependent on various factors such as, for example, the services to be supported by the user terminal (e.g., voice, data, or both), cost considerations, regulatory constraints, safety issues, and so on.
For a given pairing of multi-antenna access point and multi-antenna user terminal, a MIMO channel is formed by the NT transmit antennas and NR receive antennas available for use for data transmission. Different MIMO channels are formed between the access point and different multi-antenna user terminals. Each MIMO channel may be decomposed into NS spatial channels, with NS≦min {NT, NR}. NS data streams may be transmitted on the NS spatial channels. Spatial processing is required at a receiver and may or may not be performed at a transmitter in order to transmit multiple data streams on the NS spatial channels.
The NS spatial channels may or may not be orthogonal to one another. This depends on various factors such as (1) whether or not spatial processing was performed at the transmitter to obtain orthogonal spatial channels and (2) whether or not the spatial processing at both the transmitter and the receiver was successful in orthogonalizing the spatial channels. If no spatial processing is performed at the transmitter, then the NS spatial channels may be formed with NS transmit antennas and are unlikely to be orthogonal to one another.
The NS spatial channels may be orthogonalized by performing decomposition on a channel response matrix for the MIMO channel, as described below. Each spatial channel is referred to as an eigenmode of the MIMO channel if the NS spatial channels are orthogonalized using decomposition, which requires spatial processing at both the transmitter and the receiver, as described below. In this case, NS data streams may be transmitted orthogonally on the NS eigenmodes. However, an eigenmode normally refers to a theoretical construct. The NS spatial channels are typically not completely orthogonal to one another due to various reasons. For example, the spatial channels would not be orthogonal if (1) the transmitter has no knowledge of the MIMO channel or (2) the transmitter and/or receiver have imperfect estimate of the MIMO channel. For simplicity, in the following description, the term “eigenmode” is used to denote the case where an attempt is made to orthogonalize the spatial channels using decomposition, even though the attempt may not be fully successful due to, for example, an imperfect channel estimate.
For a given number of (e.g., four) antennas at the access point, the number of spatial channels available for each user terminal is dependent on the number of antennas employed by that user terminal and the characteristics of the wireless MIMO channel that couples the access point antennas and the user terminal antennas. If a user terminal is equipped with one antenna, then the four antennas at the access point and the single antenna at the user terminal form a multiple-input single-output (MISO) channel for the downlink and a single-input multiple-output (SIMO) channel for the uplink.
The MIMO WLAN system may be designed to support a number of transmission modes. Table 1 lists the transmission modes supported by an exemplary design of the MIMO WLAN system.
For simplicity, the term “diversity” refers to transmit diversity in the following description, unless noted otherwise.
The transmission modes available for use for the downlink and uplink for each user terminal are dependent on the number of antennas employed at the user terminal Table 2 lists the transmission modes available for different terminal types for the downlink and uplink, assuming multiple (e.g., four) antennas at the access point.
For the downlink, all of the transmission modes except for the spatial multiplexing mode may be used for single-antenna user terminals, and all transmission modes may be used for multi-antenna user terminals. For the uplink, all transmission modes may be used by multi-antenna user terminals, while single-antenna user terminals use the SIMO mode to transmit data from the one available antenna. Receive diversity (i.e., receiving a data transmission with multiple receive antennas) may be used for the SIMO, diversity, and beam-steering modes.
The MIMO WLAN system may also be designed to support various other transmission modes, and this is within the scope of the invention. For example, a beam-forming mode may be used to transmit data on a single eigenmode using both the amplitude and phase information for the eigenmode (instead of only the phase information, which is all that is used by the beam-steering mode). As another example, a “non-steered” spatial multiplexing mode can be defined whereby the transmitter simply transmits multiple data streams from multiple transmit antennas (without any spatial processing) and the receiver performs the necessary spatial processing to isolate and recover the data streams sent from the multiple transmit antennas. As yet another example, a “multi-user” spatial multiplexing mode can be defined whereby the access point transmits multiple data streams from multiple transmit antennas (with spatial processing) to multiple user terminals concurrently on the downlink. As yet another example, a spatial multiplexing mode can be defined whereby the transmitter performs spatial processing to attempt to orthogonalize the multiple data streams sent on the multiple transmit antennas (but may not be completely successful because of an imperfect channel estimate) and the receiver performs the necessary spatial processing to isolate and recover the data streams sent from the multiple transmit antennas. Thus, the spatial processing to transmit multiple data streams via multiple spatial channels may be performed (1) at both the transmitter and receiver, (2) at only the receiver, or (3) at only the transmitter. Different spatial multiplexing modes may be used depending on, for example, the capabilities of the access point and the user terminals, the available channel state information, system requirements, and so on.
In general, the access points and user terminals may be designed with any number of transmit and receive antennas. For clarity, specific embodiments and designs are described below whereby each access point is equipped with four transmit/receive antennas, and each user terminal is equipped with four or less transmit/receive antennas.
OFDM
In an embodiment, the MIMO WLAN system employs OFDM to effectively partition the overall system bandwidth into a number of (NF) orthogonal subbands. These subbands are also referred to as tones, bins, or frequency channels. With OFDM, each subband is associated with a respective subcarrier that may be modulated with data. For a MIMO system that utilizes OFDM, each spatial channel of each subband may be viewed as an independent transmission channel where the complex gain associated with each subband is effectively constant across the subband bandwidth.
In an embodiment, the system bandwidth is partitioned into 64 orthogonal subbands (i.e., NF=64), which are assigned indices of −32 to +31. Of these 64 subbands, 48 subbands (e.g., with indices of ±{1, . . . , 6, 8, . . . , 20, 22, . . . , 26}) are used for data, 4 subbands (e.g., with indices of ±{7, 21}) are used for pilot and possibly signaling, the DC subband (with index of 0) is not used, and the remaining subbands are also not used and serve as guard subbands. This OFDM subband structure is described in further detail in a document for IEEE Standard 802.11a and entitled “Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications: High-speed Physical Layer in the 5 GHz Band,” September 1999, which is publicly available and incorporated herein by reference. Different numbers of subbands and various other OFDM subband structures may also be implemented for the MIMO WLAN system, and this is within the scope of the invention. For example, all 53 subbands with indices from −26 to +26 may be used for data transmission. As another example, a 128-subband structure, a 256-subband structure, or a subband structure with some other number of subbands may be used. For clarity, the MIMO WLAN system is described below with the 64-subband structure described above.
For OFDM, the data to be transmitted on each subband is first modulated (i.e., symbol mapped) using a particular modulation scheme selected for use for that subband. Zeros are provided for the unused subbands. For each symbol period, the modulation symbols and zeros for all NF subbands are transformed to the time domain using an inverse fast Fourier transform (IFFT) to obtain a transformed symbol that contains NF time-domain samples. The duration of each transformed symbol is inversely related to the bandwidth of each subband. In one specific design for the MIMO WLAN system, the system bandwidth is 20 MHz, NF=64, the bandwidth of each subband is 312.5 KHz, and the duration of each transformed symbol is 3.2 pec.
OFDM can provide certain advantages, such as the ability to combat frequency selective fading, which is characterized by different channel gains at different frequencies of the overall system bandwidth. It is well known that frequency selective fading causes inter-symbol interference (ISI), which is a phenomenon whereby each symbol in a received signal acts as distortion to subsequent symbols in the received signal. The ISI distortion degrades performance by impacting the ability to correctly detect the received symbols. Frequency selective fading can be conveniently combated with OFDM by repeating a portion of (or appending a cyclic prefix to) each transformed symbol to form a corresponding OFDM symbol, which is then transmitted.
The length of the cyclic prefix (i.e., the amount to repeat) for each OFDM symbol is dependent on the delay spread of the wireless channel. In particular, to effectively combat ISI, the cyclic prefix should be longer than the maximum expected delay spread for the system.
In an embodiment, cyclic prefixes of different lengths may be used for the OFDM symbols, depending on the expected delay spread. For the specific MIMO WLAN system described above, a cyclic prefix of 400 nsec (8 samples) or 800 nsec (16 samples) may be selected for use for the OFDM symbols. A “short” OFDM symbol uses the 400 nsec cyclic prefix and has a duration of 3.6 μpec. A “long” OFDM symbol uses the 800 nsec cyclic prefix and has a duration of 4.0 μpec. Short OFDM symbols may be used if the maximum expected delay spread is 400 nsec or less, and long OFDM symbols may be used if the delay spread is greater than 400 nsec. Different cyclic prefixes may be selected for use for different transport channels, and the cyclic prefix may also be dynamically selectable, as described below. Higher system throughput may be achieved by using the shorter cyclic prefix when possible, since more OFDM symbols of shorter duration can be transmitted over a given fixed time interval.
The MIMO WLAN system may also be designed to not utilize OFDM, and this is within the scope of the invention.
Layer Structure
The upper layers includes various applications and protocols, such as signaling services 212, data services 214, voice services 216, circuit data applications, and so on. Signaling is typically provided as messages and data is typically provided as packets. The services and applications in the upper layers originate and terminate messages and packets according to the semantics and timing of the communication protocol between the access point and the user terminal. The upper layers utilize the services provided by Layer 2.
Layer 2 supports the delivery of messages and packets generated by the upper layers. In the embodiment shown in
Layer 1 comprises physical layer 240 and supports the transmission and reception of radio signals between the access point and user terminal. The physical layer performs coding, interleaving, modulation, and spatial processing for various transport channels used to send messages and packets generated by the upper layers. In this embodiment, the physical layer includes a multiplexing sublayer 242 that multiplexes processed PDUs for various transport channels into the proper frame format. Layer 1 provides data in units of frames.
Transport Channels
A number of services and applications may be supported by the MIMO WLAN system. Moreover, other data required for proper system operation may need to be sent by the access point or exchanged between the access point and user terminals. A number of transport channels may be defined for the MIMO WLAN system to carry various types of data. Table 3 lists an exemplary set of transport channels and also provides a brief description for each transport channel.
As shown in Table 3, the downlink transport channels used by the access point includes the BCH, FCCH, and FCH. The uplink transport channels used by the user terminals include the RACH and RCH. Each of these transport channels is described in further detail below.
The transport channels listed in Table 3 represent a specific embodiment of a channel structure that may be used for the MIMO WLAN system. Fewer, additional, and/or different transport channels may also be defined for use for the MIMO WLAN system. For example, certain functions may be supported by function-specific transport channels (e.g., pilot, paging, power control, and sync channel channels). Thus, other channel structures with different sets of transport channels may be defined and used for the MIMO WLAN system, and this is within the scope of the invention.
Frame Structures
A number of frame structures may be defined for the transport channels. The specific frame structure to use for the MIMO WLAN system is dependent on various factors such as, for example, (1) whether the same or different frequency bands are used for the downlink and uplink and (2) the multiplexing scheme used to multiplex the transport channels together.
If only one frequency band is available, then the downlink and uplink may be transmitted on different phases of a frame using time division duplexing (TDD), as described below. If two frequency bands are available, then the downlink and uplink may be transmitted on different frequency bands using frequency division duplexing (FDD).
For both TDD and FDD, the transport channels may be multiplexed together using time division multiplexing (TDM), code division multiplexing (CDM), frequency division multiplexing (FDM), and so on. For TDM, each transport channel is assigned to a different portion of a frame. For CDM, the transport channels are transmitted concurrently but each transport channel is channelized by a different channelization code, similar to that performed in a code division multiple access (CDMA) system. For FDM, each transport channel is assigned a different portion of the frequency band for the link.
Table 4 lists the various frame structures that may be used to carry the transport channels. Each of these frame structures is described in further detail below. For clarity, the frame structures are described for the set of transport channels listed in Table 3.
Each TDD frame is partitioned into a downlink phase and an uplink phase. The downlink phase is further partitioned into three segments for the three downlink transport channels—the BCH, FCCH, and FCH. The uplink phase is further partitioned into two segments for the two uplink transport channels—the RCH and RACH.
The segment for each transport channel may be defined to have either a fixed duration or a variable duration that can change from frame to frame. In an embodiment, the BCH segment is defined to have a fixed duration, and the FCCH, FCH, RCH, and RACH segments are defined to have variable durations.
The segment for each transport channel may be used to carry one or more protocol data units (PDUs) for that transport channel. In the specific embodiment shown in
Frame structure 300a represents a specific arrangement of the various transport channels within a TDD frame. This arrangement can provide certain benefits such as reduced delays for data transmission on the downlink and uplink. The BCH is transmitted first in the TDD frame since it carries system parameters that may be used for the PDUs of the other transport channels within the same TDD frame. The FCCH is transmitted next since it carries channel assignment information indicative of which user terminal(s) are designated to receive downlink data on the FCH and which user terminal(s) are designated to transmit uplink data on the RCH within the current TDD frame. Other TDD-TDM frame structures may also be defined and used for the MIMO WLAN system, and this is within the scope of the invention.
As shown in
In the specific embodiment shown in
As shown in
Other frame structures may also be defined for the downlink and uplink transport channels, and this is within the scope of the invention. Moreover, it is possible to use different types of frame structure for the downlink and uplink. For example, a TDM-based frame structure may be used for the downlink and a CDM-based frame structure may be used for the uplink.
In the following description, the MIMO WLAN system is assumed to use one frequency band for both downlink and uplink transmissions. For clarity, the TDD-TDM frame structure shown in
The transport channels are used to send various types of data and may be categorized into two groups: common transport channels and dedicated transport channels. Because the common and dedicated transport channels are used for different purposes, different processing may be used for these two groups of transport channels, as described in further detail below.
Common Transport Channels
The common transport channels include the BCH, FCCH, and RACH. These transport channels are used to send data to or receive data from multiple user terminals. For improved reliability, the BCH and FCCH are transmitted by the access point using the diversity mode. On the uplink, the RACH is transmitted by the user terminals using the beam-steering mode (if supported by the user terminal). The BCH is operated at a known fixed rate so that the user terminals can receive and process the BCH without any additional information. The FCCH and RACH support multiple rates to allow for greater efficiency. As used herein, each “rate” or “rate set” is associated with a particular code rate (or coding scheme) and a particular modulation scheme.
Dedicated Transport Channels
The dedicated transport channels include the FCH and RCH. These transport channels are normally used to send user-specific data to or by specific user terminals. The FCH and RCH may be dynamically allocated to the user terminals as necessary and as available. The FCH may also be used in a broadcast mode to send overhead, page, and broadcast messages to the user terminals. In general, the overhead, page, and broadcast messages are transmitted prior to any user-specific data on the FCH.
For clarity, the transport channels are described for the specific TDD-TDM frame structure shown in
Broadcast Channel (BCH)—Downlink
The BCH is used by the access point to transmit a beacon pilot, a MIMO pilot, and system parameters to the user terminals. The beacon pilot is used by the user terminals to acquire system timing and frequency. The MIMO pilot is used by the user terminals to estimate the MIMO channel formed by the access point antennas and their own antennas. The beacon and MIMO pilots are described in further detail below. The system parameters specify various attributes of the downlink and uplink transmissions. For example, since the durations of the FCCH, FCH, RACH, and RCH segments are variable, the system parameters that specify the length of each of these segments for the current TDD frame are sent in the BCH.
A preamble may be used to send one or more types of pilot and/or other information. A beacon pilot comprises a specific set of modulation symbols that is transmitted from all transmit antennas. A MIMO pilot comprises a specific set of modulation symbols that is transmitted from all transmit antennas with different orthogonal codes, which then allows the receivers to recover the pilot transmitted from each antenna. Different sets of modulation symbols may be used for the beacon and MIMO pilots. The generation of the beacon and MIMO pilots is described in further detail below.
The BCH message carries system configuration information. Table 5 lists the various fields for an exemplary BCH message format.
The Frame Counter value may be used to synchronize various processes at the access point and user terminals (e.g., the pilot, scrambling codes, cover code, and so on). A frame counter may be implemented with a 4-bit counter that wraps around. This counter is incremented at the start of each TDD frame, and the counter value is included in the Frame Counter field. The Net ID field indicates the identifier (ID) of the network to which the access point belongs. The AP ID field indicates the ID of the access point within the network ID. The AP Tx Lvl and AP Rx Lvl fields indicate the maximum transmit power level and the desired receive power level at the access point, respectively. The desired receive power level may be used by the user terminal to determine the initial uplink transmit power.
The FCCH Length, FCH Length, and RCH Length fields indicate the lengths of the FCCH, FCH, and RCH segments, respectively, for the current TDD frame. The lengths of these segments are given in units of OFDM symbols. The OFDM symbol duration for the BCH is fixed at 4.0 μsec. The OFDM symbol duration for all other transport channels (i.e., the FCCH, FCH, RACH, and RCH) is variable and depends on the selected cyclic prefix, which is specified by the Cyclic Prefix Duration field. The FCCH Rate field indicates the rate used for the FCCH for the current TDD frame.
The RACH Length field indicates the length of the RACH segment, which is given in units of RACH slots. The duration of each RACH slot is given by the RACH Slot Size field, in units of OFDM symbols. The RACH Guard Interval field indicates the amount of time between the last RACH slot and the start of the BCH segment for the next TDD frame. These various fields for the RACH are described in further detail below.
The Page Bit and Broadcast Bit indicate whether or not page messages and broadcast messages, respectively, are being sent on the FCH in the current TDD frame. These two bits may be set independently for each TDD frame. The RACH Acknowledgment Bit indicates whether or not acknowledgments for PDUs sent on the RACH in prior TDD frames are being sent on the FCCH in the current TDD frame.
The CRC field includes a CRC value for the entire BCH message. This CRC value may be used by the user terminals to determine whether the received BCH message is decoded correctly (i.e., good) or in error (i.e., erased). The Tail Bits field includes a group of zeros used to reset the convolutional encoder to a known state at the end of the BCH message.
As shown in Table 5, the BCH message includes a total of 120 bits. These 120 bits may be transmitted with 10 OFDM symbols using the processing described in detail below.
Table 5 shows a specific embodiment of the format for the BCH message. Other BCH message formats with fewer, additional, and/or different fields may also be defined and used, and this is within the scope of the invention.
Forward Control Channel (FCCH)—Downlink
In an embodiment, the access point is able to allocate resources for the FCH and RCH on a per frame basis. The FCCH is used by the access point to convey the resource allocation for the FCH and RCH (i.e., the channel assignments).
In an embodiment, the FCCH can be transmitted using four possible rates. The specific rate used for the FCCH PDU in each TDD frame is indicated by the FCCH Phy Mode field in the BCH message. Each FCCH rate corresponds to a particular code rate and a particular modulation scheme and is further associated with a particular transmission mode, as shown in Table 26.
An FCCH message may include zero, one, or multiple information elements (IEs). Each information element may be associated with a specific user terminal and may be used to provide information indicative of the assignment of FCH/RCH resources for that user terminal Table 6 lists the various fields for an exemplary FCCH message format.
The N_IE field indicates the number of information elements included in the FCCH message sent in the current TDD frame. For each information element (IE) included in the FCCH message, the IE Type field indicates the particular type of this IE. A number of IE types are defined for use to allocate resources for different types of transmissions, as described below.
The MAC ID field identifies the specific user terminal for which the information element is intended. Each user terminal registers with the access point at the start of a communication session and is assigned a unique MAC ID by the access point. This MAC ID is used to identify the user terminal during the session.
The Control Fields are used to convey channel assignment information for the user terminal and are described in detail below. The Padding Bits field includes a sufficient number of padding bits so that the overall length of the FCCH message is an even number of OFDM symbols. The FCCH CRC field includes a CRC value that may be used by the user terminals to determine whether the received FCCH message is decoded correctly or in error. The Tail Bits field includes zeros used to reset the convolutional encoder to a known state at the end of the FCCH message. Some of these fields are described in further detail below.
A number of transmission modes are supported by the MIMO WLAN system for the FCH and RCH, as indicated in Table 1. Moreover, a user terminal may be active or idle during a connection. Thus, a number of types of IE are defined for use to allocate FCH/RCH resources for different types of transmissions. Table 7 lists an exemplary set of IE types.
For IE types 0, 1 and 4, resources are allocated to a specific user terminal for both the FCH and RCH (i.e., in channel pairs). For IE type 2, minimal resources are allocated to the user terminal on the FCH and RCH to maintain up-to-date estimate of the link. An exemplary format for each IE type is described below. In general, the rates and durations for the FCH and RCH can be independently assigned to the user terminals.
IE Type 0, 4—Diversity/Beam-Steering Mode
IE type 0 and 4 are used to allocate FCH/RCH resources for the diversity and beam-steering modes, respectively. For fixed low-rate services (e.g., voice), the rate remains fixed for the duration of the call. For variable rate services, the rate may be selected independently for the FCH and RCH. The FCCH IE indicates the location of the FCH and RCH PDUs assigned to the user terminal Table 8 lists the various fields of an exemplary IE Type 0 and 4 information element.
The FCH and RCH Offset fields indicate the time offset from the beginning of the current TDD frame to the start of the FCH and RCH PDUs, respectively, assigned by the information element. The FCH and RCH Rate fields indicate the rates for the FCH and RCH, respectively.
The FCH and RCH Preamble Type fields indicate the size of the preamble in the FCH and RCH PDUs, respectively. Table 9 lists the values for the FCH and RCH Preamble Type fields and the associated preamble sizes.
The RCH Timing Adjustment field includes two bits used to adjust the timing of the uplink transmission from the user terminal identified by the MAC ID field. This timing adjustment is used to reduce interference in a TDD-based frame structure (such as the one shown in
The RCH Power Control field includes two bits used to adjust the transmit power of the uplink transmission from the identified user terminal. This power control is used to reduce interference on the uplink. Table 11 lists the values for the RCH Power Control field and the associated actions.
The channel assignment for the identified user terminal may be provided in various manners. In an embodiment, the user terminal is assigned FCH/RCH resources for only the current TDD frame. In another embodiment, the FCH/RCH resources are assigned to the terminal for each TDD frame until canceled. In yet another embodiment, the FCH/RCH resources are assigned to the user terminal for every n-th TDD frame, which is referred to as “decimated” scheduling of TDD frames. The different types of assignment may be indicated by an Assignment Type field in the FCCH information element.
IE Type 1—Spatial Multiplexing Mode
IE type 1 is used to allocate FCH/RCH resources to user terminals using the spatial multiplexing mode. The rate for these user terminals is variable, and may be selected independently for the FCH and RCH. Table 12 lists the various fields of an exemplary IE type 1 information element.
For IE type 1, the rate for each spatial channel may be selected independently on the FCH and RCH. The interpretation of the rates for the spatial multiplexing mode is general in that it can specify the rate per spatial channel (e.g., for up to four spatial channels for the embodiment shown in Table 12). The rate is given per eigenmode if the transmitter performs spatial processing to transmit data on the eigenmodes. The rate is given per antenna if the transmitter simply transmits data from the transmit antennas and the receiver performs the spatial processing to isolate and recover the data (for the non-steered spatial multiplexing mode).
The information element includes the rates for all enabled spatial channels and zeros for the ones not enabled. User terminals with less than four transmit antennas set the unused FCH/RCH Spatial Channel Rate fields to zero. Since the access point is equipped with four transmit/receive antennas, user terminals with more than four transmit antennas may use them to transmit up to four independent data streams.
IE Type 2—Idle Mode
IE type 2 is used to provide control information for user terminals operating in an Idle state (described below). In an embodiment, when a user terminal is in the Idle state, steering vectors used by the access point and user terminal for spatial processing are continually updated so that data transmission can start quickly if and when resumed. Table 13 lists the various fields of an exemplary IE type 2 information element.
IE Type 3—RACH Quick Acknowledgment
IE type 3 is used to provide quick acknowledgment for user terminals attempting to access the system via the RACH. To gain access to the system or to send a short message to the access point, a user terminal may transmit an RACH PDU on the uplink. After the user terminal sends the RACH PDU, it monitors the BCH to determine if the RACH Acknowledgement Bit is set. This bit is set by the access point if any user terminal was successful in accessing the system and an acknowledgment is being sent for at least one user terminal on the FCCH. If this bit is set, then the user terminal processes the FCCH for acknowledgment sent on the FCCH. IE Type 3 information elements are sent if the access point desires to acknowledge that it correctly decoded the RACH PDUs from the user terminals without assigning resources. Table 14 lists the various fields of an exemplary IE Type 3 information element.
A single or multiple types of acknowledgment may be defined and sent on the FCCH. For example, a quick acknowledgment and an assignment-based acknowledgment may be defined. A quick acknowledgment may be used to simply acknowledge that the RACH PDU has been received by the access point but that no FCH/RCH resources have been assigned to the user terminal. An assignment-based acknowledgment includes assignments for the FCH and/or RCH for the current TDD frame.
The FCCH may be implemented in other manners and may also be transmitted in various ways. In one embodiment, the FCCH is transmitted at a single rate that is signaled in the BCH message. This rate may be selected, for example, based on the lowest signal-to-noise-and-interference ratios (SNRs) of all user terminals for which the FCCH is being sent in the current TDD frame. Different rates may be used for different TDD frames depending on the channel conditions of the recipient user terminals in each TDD frame.
In another embodiment, the FCCH is implemented with multiple (e.g., four) FCCH subchannels. Each FCCH subchannel is transmitted at a different rate and is associated with a different required SNR in order to recover the subchannel. The FCCH subchannels are transmitted in order from lowest rate to highest rate. Each FCCH subchannel may or may not be transmitted in a given TDD frame. The first FCCH subchannel (with the lowest rate) is transmitted first and can be received by all user terminals. This FCCH subchannel can indicate whether or not each of the remaining FCCH subchannels will be transmitted in the current TDD frame. Each user terminal can process the transmitted FCCH subchannels to obtain its FCCH information element. Each user terminal can terminate processing of the FCCH if any of the following occurs: (1) failure to decode the current FCCH subchannel, (2) reception of its FCCH information element in the current FCCH subchannel, or (3) all transmitted FCCH subchannels have been processed. A user terminal can terminate processing of the FCCH as soon as it encounters FCCH decoding failure because the FCCH subchannels are transmitted at ascending rates and the user terminal is unlikely to be able to decode subsequent FCCH subchannels transmitted at higher rates.
Random Access Channel (RACH)—Uplink
The RACH is used by the user terminals to gain access to the system and to send short messages to the access point. The operation of the RACH is based on a slotted Aloha random access protocol, which is described below.
In an embodiment, four different rates are supported for the RACH. The specific rate used for each RACH message is indicated by a 2-bit RACH data rate indicator (DRI), which is embedded in the preamble portion of the RACH PDU, as shown in
The RACH message carries short messages and access requests from the user terminal Table 16 lists the various fields of an exemplary RACH message format and the size of each field for each of the four different message sizes.
The Message Duration field indicates the size of the RACH message. The MAC PDU Type field indicates the RACH message type. The MAC ID field contains the MAC ID that uniquely identifies the user terminal sending the RACH message. During initial system access, a unique MAC ID has not been assigned to the user terminal. In this case, a registration MAC ID (e.g., a specific value reserved for registration purpose) may be included in the MAC ID field. The Slot ID field indicates the starting RACH slot on which the RACH PDU was sent (the RACH timing and transmission is described below). The Payload field includes the information bits for the RACH message. The CRC field contains a CRC value for the RACH message, and the Tail Bits field is used to reset the convolutional encoder for the RACH. The operation of the RACH in conjunction with the BCH and FCCH for system access is described in further detail below.
The RACH may also be implemented with a “fast” RACH (F-RACH) and a “slow” RACH (S-RACH). The F-RACH and S-RACH can be designed to efficiently support user terminals in different operating states. For example, the F-RACH may be used by user terminals that (1) have registered with the system, (2) can compensate for their round trip delays (RTDs) by properly advancing their transmit timing, and (3) achieve the required SNR for operation on the F-RACH. The S-RACH may be used by user terminals that cannot use the F-RACH for any reasons.
Different designs may be used for the F-RACH and S-RACH to facilitate rapid access to the system whenever possible and to minimize the amount of system resources needed to implement random access. For example, the F-RACH can use a shorter PDU, employ a weaker coding scheme, require F-RACH PDUs to arrive approximately time-aligned at the access point, and utilize a slotted Aloha random access scheme. The S-RACH can use a longer PDU, employ a stronger coding scheme, allow S-RACH PDUs to arrive non-aligned in time at the access point, and utilize an unslotted Aloha random access scheme.
For simplicity, the following description assumes that a single RACH is used for the MIMO WLAN system.
Forward Channel (FCH)—Downlink
The FCH is used by the access point to transmit user-specific data to specific user terminals and page/broadcast messages to multiple user terminals. The FCH may also be used to transmit pilot to user terminals. The FCH can be allocated on a per frame basis. A number of FCH PDU types are provided to accommodate different uses of the FCH. Table 17 lists an exemplary set of FCH PDU types.
FCH PDU Type 0 is used to send page/broadcast messages and user messages/packets on the FCH and only includes the message/packet. (Data for a specific user terminal may be sent as a message or a packet, and these two terms are used interchangeably herein.) FCH PDU Type 1 is used to send user packets and includes a preamble. FCH PDU Type 2 includes only the preamble and no message/packet, and is associated with Idle state FCH traffic.
A number of FCH Message types are provided to accommodate different uses of the FCH. Table 18 lists an exemplary set of FCH Message types.
A page message may be used to page multiple user terminals and is sent using FCH PDU Type 0. If the Page Bit in the BCH message is set, then one or more FCH PDUs with page messages (or “Page PDUs”) are sent first on the FCH. Multiple Page PDUs may be sent in the same frame. Page PDUs are transmitted using the diversity mode and the lowest rate of 0.25 bps/Hz to increase the likelihood of correct reception by the user terminals.
A broadcast message may be used to send information to multiple user terminals and is sent using FCH PDU Type 0. If the Broadcast Bit in the BCH message is set, then one or more FCH PDUs with broadcast messages (or “Broadcast PDUs”) are sent on the FCH immediately following any Page PDUs sent on the FCH. The Broadcast PDUs are also transmitted using the diversity mode and the lowest rate of 0.25 bps/Hz to increase the likelihood of correct reception.
A user packet may be used to send user-specific data, and may be sent using FCH PDU Type 1 or 2. User PDUs of Type 1 and 2 are sent on the FCH following any Page and Broadcast PDUs sent on the FCH. Each User PDU may be transmitted using the diversity, beam-steering, or spatial multiplexing mode. The FCCH information element specifies the rate and transmission mode used for each User PDU sent on the FCH.
A message or packet sent on the FCH comprises an integer number of PHY frames. In an embodiment and as described below, each PHY frame may include a CRC value that permits individual PHY frames in an FCH PDU to be checked and retransmitted if necessary. For asynchronous services, the RLP may be employed for segmentation, retransmission, and reassembly of PHY frames within a given FCH PDU. In another embodiment, a CRC value is provided for each message or packet, instead of each PHY frame.
Table 19 lists the various fields for an exemplary FCH PDU format for FCH PDU Types 0 and 1.
The FCH Message Type and FCH Message Length fields are sent in the header of the first PHY frame of the FCH PDU. The payload, CRC, and tail bits fields are included in each PHY frame. The payload portion of each FCH PDU carries the information bits for the page/broadcast message or user-specific packet. Pad bits are used to fill the last PHY frame of the FCH PDU, if required.
A PHY frame may also be defined to comprise some other number of OFDM symbols (e.g., one, two, four, eight, and so on). The PHY frame may be defined with even number of OFDM symbols because OFDM symbols are transmitted in pairs for the diversity mode, which may be used for the FCH and RCH. The PHY frame size may be selected based on the expected traffic such that inefficiency is minimized. In particular, if the frame size is too large, then inefficiency results from using a large PHY frame to send a small amount of data. Alternatively, if the frame size is too small, then the overhead represents a larger fraction of the frame.
Reverse Channel (RCH)—Uplink
The RCH is used by the user terminals to transmit uplink data and pilot to the access point. The RCH may be allocated on a per TDD frame basis. One or more user terminals may be designated to transmit on the RCH in any given TDD frame. A number of RCH PDU types are provided to accommodate different operating modes on the RCH. Table 20 lists an exemplary set of RCH PDU types.
RCH PDU Type 0 is used to send a message/packet on the RCH and does not include a preamble. RCH PDU Type 1 is used to send a message/packet and includes a preamble. RCH PDU Type 2 includes a preamble and a short message, and is associated with Idle state RCH traffic.
A packet sent on the RCH (for PDU Types 0 and 1) comprises an integer number of PHY frames. The structure for an RCH packet (for PDU Types 0 and 1) is shown in
Table 21 lists the various fields for an exemplary RCH PDU format for RCH PDU Types 0 and 1.
The RCH Message Type, RCH Message Length, and FCH Rate Indicator fields are sent in the header of the first PHY frame of the RCH PDU. The FCH Rate Indicator field is used to convey FCH rate information (e.g., the maximum rates supported by each of the spatial channels) to the access point.
Table 22 lists the various fields for an exemplary RCH PDU format for RCH PDU Type 2.
The RCH Request field is used by the user terminal to request additional capacity on the uplink. This short RCH message does not include a CRC and is transmitted in a single OFDM symbol.
Dedicated Channel Activity
Data transmission on the FCH and RCH can occur independently. Depending on the transmission modes selected for use for the FCH and RCH, one or multiple spatial channels (for the beam-steering and diversity modes) may be active and used for data transmission for each dedicated transport channel. Each spatial channel may be associated with a specific rate.
When only the FCH or only the RCH has all four rates set to zero, the user terminal is idle on that link. The idle terminal still transmits an idle PDU on the RCH. When both the FCH and RCH have all four rates set to zero, both the access point and user terminal are off and not transmitting. User terminals with less than four transmit antennas set the unused rate fields to zero. User terminals with more than four transmit antennas use no more than four spatial channels to transmit data. Table 23 shows the transmission state and channel activity when the rates on all four spatial channels of either the FCH or RCH (or both) are set to zero.
There may also be a situation where both the RCH and FCH are idle (i.e., not transmitting data) but still transmitting preamble. This is referred to as the Idle state. The control fields used to support a user terminal in the Idle state are provided in an FCCH IE Type 2 information element, which is shown in Table 13.
Alternative Designs
For clarity, specific PDU types, PDU structures, message formats, and so on, have been described for an exemplary design. Fewer, additional, and/or different types, structures, and formats may also be defined for use, and this is within the scope of the invention.
In the above description, the same OFDM subband structure is used for all of the transport channels. Improved efficiency may be achieved by using different OFDM subband structures for different transport channels. For example, a 64-subband structure may be used for some transport channels, a 256-subband structure may be used for some other transport channels, and so on. Moreover, multiple OFDM subband structures may be used for a given transport channel.
For a given system bandwidth of W, the duration of an OFDM symbol is dependent on the number of total subbands. If the total number of subbands is N, then the duration of each transformed symbol (without a cyclic prefix) is N/W μsec (if W is given in MHz). A cyclic prefix is added to each transformed symbol to form a corresponding OFDM symbol. The length of the cyclic prefix is determined by the expected delay spread of the system. The cyclic prefix represents overhead, which is needed for each OFDM symbol in order to combat a frequency selective channel. This overhead represents a larger percentage of the OFDM symbol if the symbol is short and a smaller percentage if the symbol is long.
Since different transport channels may be associated with different types of traffic data, a suitable OFDM subband structure may be selected for use for each transport channel to match the expected traffic data type. If a large amount of data is expected to be transmitted on a given transport channel, then a larger subband structure may be defined for use for the transport channel. In this case, the cyclic prefix would represent a smaller percentage of the OFDM symbol and greater efficiency may be achieved. Conversely, if a small amount of data is expected to be transmitted on a given transport channel, than a smaller subband structure may be defined for use for the transport channel. In this case, even though the cyclic prefix represents a larger percentage of the OFDM symbol, greater efficiency may still be achieved by reducing the amount of excess capacity by using a smaller size OFDM symbol. The OFDM symbol may thus be viewed as a “box car” that is used to send data, and the proper size “box car” may be selected for each transport channel depending on the amount of data expected to be sent.
For example, for the embodiment described above, the data on the FCH and RCH is sent in PHY frames, each of which comprises 6 OFDM symbols. In this case, another OFDM structure may be defined for use for the FCH and RCH. For example, a 256-subband structure may be defined for the FCH and RCH. A “large” OFDM symbol for the 256-subband structure would be approximately four times the duration of a “small” OFDM symbol for the 64-subband structure but would have four times the data-carrying capacity. However, only one cyclic prefix is needed for one large OFDM symbol, whereas four cyclic prefixes are needed for the equivalent four small OFDM symbols. Thus, the amount of overhead for the cyclic prefixes may be reduced by 75% by the use of the larger 256-subband structure.
This concept may be extended so that different OFDM subband structures may be used for the same transport channel. For example, the RCH supports different PDU types, each of which may be associated with a certain size. In this case, a larger subband structure may be used for a larger-size RCH PDU type, and a smaller subband structure may be used for a smaller-size RCH PDU type. A combination of different subband structures may also be used for a given PDU. For example, if one long OFDM symbol is equivalent to four short OFDM symbols, then a PDU may be sent using Nlarge large OFDM symbols and Nsmall small OFDM symbols, where Nlarge≧0 and 3≧Nsmall≧0.
Different OFDM subband structures are associated with OFDM symbols of different lengths. Thus, if different OFDM subband structures are used for different transport channels (and/or for the same transport channel), then the FCH and RCH offsets for the FCH and RCH PDUs would need to be specified with the proper time resolution, which is smaller than an OFDM symbol period. In particular, the time increment for the FCH and RCH PDUs may be given in integer numbers of cyclic prefix length, instead of OFDM symbol period.
The transport channels described above are used to send various types of data for various services and functions. Each transport channel may be designed to support one or more rates and one or more transmission modes.
Transmission Modes
A number of transmission modes are supported for the transport channels. Each transmission mode is associated with specific spatial processing at the transmitter and receiver, as described below. Table 24 lists the transmission mode(s) supported by each of the transport channels.
For the diversity mode, each data symbol is transmitted redundantly over multiple transmit antennas, multiple subbands, multiple symbol periods, or a combination thereof to achieve spatial, frequency, and/or time diversity. For the beam-steering mode, a single spatial channel is used for data transmission (typically the best spatial channel), and each data symbol is transmitted on the single spatial channel using full transmit power available for the transmit antennas. For the spatial multiplexing mode, multiple spatial channels are used for data transmission, and each data symbol is transmitted on one spatial channel, where a spatial channel may correspond to an eigenmode, a transmit antenna, and so on. The beam-steering mode may be viewed as a special case of the spatial multiplexing mode whereby only one spatial channel is used for data transmission.
The diversity mode may be used for the common transport channels (BCH and FCCH) for the downlink from the access point to the user terminals. The diversity mode may also be used for the dedicated transport channels (FCH and RCH). The use of the diversity mode on the FCH and RCH may be negotiated at call setup. The diversity mode transmits data on one “spatial mode” using a pair of antennas for each subband.
The beam-steering mode may be employed on the RACH by user terminals with multiple transmit antennas. A user terminal can estimate the MIMO channel based on the MIMO pilot sent on the BCH. This channel estimate may then be used to perform beam-steering on the RACH for system accesses. The beam-steering mode may also be used for the dedicated transport channels (FCH and RCH). The beam-steering mode may be able to achieve higher received signal-to-noise-and-interference ratio (SNR) at the receiver than the diversity mode by exploiting the gain of the antenna array at the transmitter. In addition, the preamble portion of the PDU may be reduced since the steered reference only includes symbols for a single “steered” antenna. The diversity mode may also be used for the RACH.
The spatial multiplexing mode may be used for the FCH and RCH to achieve greater throughput, when supported by the channel conditions. The spatial multiplexing and beam-steering modes are reference driven and require closed-loop control for proper operation. As such, a user terminal is allocated resources on both the FCH and RCH to support the spatial multiplexing mode. Up to four spatial channels may be supported on the FCH and RCH (limited by the number of antennas at the access point).
Coding and Modulation
A number of different rates are supported for the transport channels. Each rate is associated with a particular code rate and a particular modulation scheme, which collectively results in a particular spectral efficiency (or data rate). Table 25 lists the various rates supported by the system.
Each common transport channel supports one or more rates and one transmission mode (or possibly more, as may be the case for the RACH). The BCH is transmitted at a fixed rate using the diversity mode. The FCCH may be transmitted at one of four possible rates, as indicated by the FCCH Phy Mode field in the BCH message, using the diversity mode. In one embodiment, the RACH may be transmitted at one of four possible rates, as indicated by the RACH DRI embedded in the preamble of the RACH PDU, and each RACH message is one of four possible sizes. In another embodiment, the RACH is transmitted at a single rate. Table 26 lists the coding, modulation, and transmission parameters and the message sizes supported by each common transport channel.
The FCCH message is variable in size and given in even number of OFDM symbols.
The FCH and RCH support all of the rates listed in Table 25. Table 27 lists the coding, modulation, and transmission parameters and the message sizes supported by the FCH and RCH.
The PHY frame size in Table 27 indicates the number of code bits, modulation symbols, and OFDM symbols for each PHY frame. If 48 data subbands are used for data transmission, then each OFDM symbol includes 48 modulation symbols. For the diversity and beam-steering modes, one symbol stream is transmitted and the PHY frame size corresponds to the single rate employed for this symbol stream. For the spatial multiplexing mode, multiple symbol streams may be sent on multiple spatial channels, and the overall PHY frame size is determined by the sum of the PHY frame sizes for the individual spatial channels. The PHY frame size for each spatial channel is determined by the rate employed for that spatial channel.
As an example, suppose the MIMO channel is capable of supporting four spatial channels operating at spectral efficiencies of 0.5, 1.5, 4.5, and 5.5 bps/Hz. The four rates selected for the four spatial channels would then be as shown in Table 28.
The overall PHY frame size is then 144+432+1296+1584=3456 information bits or 288+576+1728+2304=4896 code bits. Even though each of the four spatial channels supports a different number of payload bits, the overall PHY frame can be transmitted in 6 OFDM symbols (e.g., 24 μsec, assuming 4 psec/OFDM symbol).
On the downlink, at access point 110x, a transmit (TX) data processor 710 receives traffic data (i.e., information bits) from a data source 708 and signaling and other information from a controller 730 and possibly a scheduler 734. These various types of data may be sent on different transport channels. TX data processor 710 “frames” the data (if necessary), scrambles the framed/unframed data, encodes the scrambled data, interleaves (i.e., reorders) the coded data, and maps the interleaved data into modulation symbols. For simplicity, a “data symbol” refers to a modulation symbol for traffic data, and a “pilot symbol” refers to a modulation symbol for pilot. The scrambling randomizes the data bits. The encoding increases the reliability of the data transmission. The interleaving provides time, frequency, and/or spatial diversity for the code bits. The scrambling, coding, and modulation may be performed based on control signals provided by controller 730 and are described in further detail below. TX data processor 710 provides a stream of modulation symbols for each spatial channel used for data transmission.
A TX spatial processor 720 receives one or more modulation symbol streams from TX data processor 710 and performs spatial processing on the modulation symbols to provide four streams of transmit symbols, one stream for each transmit antenna. The spatial processing is described in further detail below.
Each modulator (MOD) 722 receives and processes a respective transmit symbol stream to provide a corresponding stream of OFDM symbols. Each OFDM symbol stream is further processed to provide a corresponding downlink modulated signal. The four downlink modulated signals from modulator 722a through 722d are then transmitted from four antennas 724a through 724d, respectively.
At each user terminal 120, one or multiple antennas 752 receive the transmitted downlink modulated signals, and each receive antenna provides a received signal to a respective demodulator (DEMOD) 754. Each demodulator 754 performs processing complementary to that performed at modulator 722 and provides received symbols. A receive (RX) spatial processor 760 then performs spatial processing on the received symbols from all demodulators 754 to provide recovered symbols, which are estimates of the modulation symbols sent by the access point.
An RX data processor 770 receives and demultiplexes the recovered symbols into their respective transport channels. The recovered symbols for each transport channel may be symbol demapped, deinterleaved, decoded, and descrambled to provide decoded data for that transport channel. The decoded data for each transport channel may include recovered packet data, messages, signaling, and so on, which are provided to a data sink 772 for storage and/or a controller 780 for further processing.
The processing by access point 110 and terminal 120 for the downlink is described in further detail below. The processing for the uplink may be the same or different from the processing for the downlink.
For the downlink, at each active user terminal 120, RX spatial processor 760 further estimates the downlink to obtain channel state information (CSI). The CSI may include channel response estimates, received SNRs, and so on. RX data processor 770 may also provide the status of each packet/frame received on the downlink. A controller 780 receives the channel state information and the packet/frame status and determines the feedback information to be sent back to the access point. The feedback information is processed by a TX data processor 790 and a TX spatial processor 792 (if present), conditioned by one or more modulators 754, and transmitted via one or more antennas 752 back to the access point.
At access point 110, the transmitted uplink signal(s) are received by antennas 724, demodulated by demodulators 722, and processed by an RX spatial processor 740 and an RX data processor 742 in a complementary manner to that performed at the user terminal. The recovered feedback information is then provided to controller 730 and a scheduler 734.
Scheduler 734 uses the feedback information to perform a number of functions such as (1) selecting a set of user terminals for data transmission on the downlink and uplink, (2) selecting the transmission rate(s) and the transmission mode for each selected user terminal, and (3) assigning the available FCH/RCH resources to the selected terminals. Scheduler 734 and/or controller 730 further uses information (e.g., steering vectors) obtained from the uplink transmission for the processing of the downlink transmission, as described in further detail below.
A number of transmission modes are supported for data transmission on the downlink and uplink. The processing for each of these transmission modes is described in further detail below.
Diversity Mode—Transmit Processing
Within a TX data processor 710a, a framing unit 808 frames the data for each packet to be transmitted on the FCH or RCH. The framing need not be performed for the other transport channels. The framing may be performed as illustrated in
An encoder 812 receives and codes the scrambled data in accordance with a selected coding scheme to provide code bits. A repeat/puncture unit 814 then repeats or punctures (i.e., deletes) some of the code bits to obtain the desired code rate. In an embodiment, encoder 812 is a rate ½, constraint length 7, binary convolutional encoder. A code rate of ¼ may be obtained by repeating each code bit once. Code rates greater than ½ may be obtained by deleting some of the code bits from encoder 812. A specific design for framing unit 808, scrambler 810, encoder 812, and repeat/puncture unit 814 is described below.
An interleaver 818 then interleaves (i.e., reorders) the code bits from unit 814 based on a selected interleaving scheme. In an embodiment, each group of 48 consecutive code bits to be transmitted on a given spatial channel is spread over the 48 data-carrying subbands (or simply, data subbands) to provide frequency diversity. The interleaving is described in further detail below.
A symbol mapping unit 820 then maps the interleaved data in accordance with a particular modulation scheme to provide modulation symbols. As shown in Table 26, BPSK, 4 QAM, or 16 QAM may be used for the diversity mode, depending on the selected rate. In the diversity mode, the same modulation scheme is used for all data subbands. The symbol mapping may be achieved by (1) grouping sets of B bits to form B-bit values, where B≧1, and (2) mapping each B-bit value to a point in a signal constellation corresponding to the selected modulation scheme. Each mapped signal point is a complex value and corresponds to a modulation symbol. Symbol mapping unit 820 provides a stream of modulation symbols to a TX diversity processor 720a.
In an embodiment, the diversity mode utilizes space-time transmit diversity (STTD) for dual transmit diversity on a per subband basis. STTD supports simultaneous transmission of independent symbol streams on two transmit antennas while maintaining orthogonality at the receiver.
The STTD scheme operates as follows. Suppose that two modulation symbols, denoted as s1 and s2, are to be transmitted on a given subband. The transmitter generates two vectors, x1=[s1 s2]T and x2=[s2*−s1*]T, where “*” denotes the complex conjugate and “T” denotes the transpose. Each vector includes two elements that are to be transmitted from two transmit antennas in one symbol period (i.e., vector x1 is transmitted from two antennas in the first symbol period, and vector x2 is transmitted from two antennas in the next symbol period).
If the receiver is equipped with a single receive antenna, then the received symbols may be expressed as:
r
1
=h
1
s
1
+h
2
s
2
+n
1, and Eq. (1)
r
2
=h
1
s
2
*−h
2
s
1
*+n
2,
The receiver may then derive estimates of the two transmitted symbols, s1 and s2 as follows:
Alternatively, the transmitter may generate two vectors x1=[s1−s2*]T and x2=[s2 s1*]T and transmit the two vectors sequentially in two symbol periods from two transmit antennas. The received symbols may then be expressed as:
r
1
=h
1
s
1
−h
2
s
2
*+n
1, and
r
2
=h
1
s
2
+h
2
s
1
*+n
2.
The receiver may then derive estimates of the two transmitted symbols as follows:
The above description may be extended for a MIMO-OFDM system with two or more transmit antennas, NR receive antennas, and multiple subbands. Two transmit antennas are used for any given subband. Suppose that two modulation symbols, denoted as s1(k) and s2(k), are to be transmitted on a given subband k. The transmitter generates two vectors x1(k)=[s1(k) s2(k)]T and x2(k)=[s2*(k)−si*(k)]T or equivalently two symbol sets {xi(k)}={s1(k) s2*(k)} and {xj(k)}={s2(k)−s1*(k)}. Each symbol set includes two elements that are to be transmitted sequentially in two symbol periods from a respective transmit antenna on subband k (i.e., symbol set {xi(k)} is transmitted on subband k from antenna i in two symbol periods, and symbol set is {xj(k)} is transmitted on subband k from antenna j in the same 2-symbol period).
The vectors of received symbols at the receive antennas in the two symbol periods may be expressed as:
r
1(k)=hi(k)s1(k)+hj(k)s2(k)+ni(k), and
r
2(k)=hi(k)s2*(k)−hj(k)s1*(k)+n2(k),
where
The receiver may then derive estimates of the two transmitted symbols, s1(k) and s2(k), as follows:
Alternatively, the transmitter may generate two symbol sets {xi(k)}={s1(k) s2(k)} and {xj(k)}={−s2*(k) s1*(k)} and transmit these two symbol sets from two transmit antennas i and j. The vectors of received symbols may then be expressed as:
r
1(k)=hi(k)s1(k)−hj(k)s2*(k)+n1(k), and
r
2(k)=hi(k)s2(k)+hj(k)s1*(k)+n2(k),
The receiver may then derive estimates of the two transmitted symbols as follows:
The STTD scheme is described by S. M. Alamouti in a paper entitled “A Simple Transmit Diversity Technique for Wireless Communications,” IEEE Journal on Selected Areas in Communications, Vol. 16, No. 8, October 1998, pgs. 1451-1458. The STTD scheme is also described in commonly assigned U.S. patent application Ser. No. 09/737,602, entitled “Method and System for Increased Bandwidth Efficiency in Multiple Input-Multiple Output Channels,” filed Jan. 5, 2001, and U.S. patent application Ser. No. 10/179,439, entitled “Diversity Transmission Modes for MIMO OFDM Communication Systems,” filed Jun. 24, 2002.
The STTD scheme effectively transmits one modulation symbol per subband over two transmit antennas in each symbol period. However, the STTD scheme distributes the information in each modulation symbol over two successive OFDM symbols. Thus, the symbol recovery at the receiver is performed based on two consecutive received OFDM symbols.
The STTD scheme utilizes one pair of transmit antennas for each data subband. Since the access point includes four transmit antennas, each antenna may be selected for use for half of the 48 data subbands. Table 29 lists an exemplary subband-antenna assignment scheme for the STTD scheme.
As shown in Table 29, transmit antennas 1 and 2 are used for subbands with indices −26, −19, −13, and so on, transmit antennas 2 and 4 are used for subbands with indices −25, −18, −12, and so on, transmit antennas 1 and 3 are used for subbands with indices −24, −17, −11, and so on. There are six different antenna pairings with four transmit antennas. Each of the six antenna pairings is used for 8 data subbands, which are spaced approximately uniformly across the 48 data subbands. The antenna pairing to subband assignment is such that different antennas are used for adjacent subbands, which may provide greater frequency and spatial diversity. For example, antennas 1 and 2 are used for subband −26, and antennas 3 and 4 are used for subband −25.
The antenna-subband assignment in Table 29 is also such that all four transmit antennas are used for each code bit for the lowest rate of ¼, which may maximize spatial diversity. For rate ¼, each code bit is repeated and sent on two subbands (which is also referred to as dual subband repeat coding). The two subbands used for each code bit are mapped to different antenna pairs so that all four antennas are used to transmit that code bit. For example, bit indices 0 and 1 in Table 29 correspond to the same code bit for the diversity mode, where the bit with index 0 is transmitted from antennas 1 and 2 on subband −26 and the bit with index 1 is transmitted from antennas 3 and 4 on subband 1. As another example, bit indices 2 and 3 in Table 29 correspond to the same code bit, where the bit with index 2 is transmitted from antennas 1 and 3 in subband −17 and the bit with index 3 is transmitted from antennas 2 and 4 in subband 10.
The system may support other transmit diversity schemes, and this is within the scope of the invention. For example, the system may support a space-frequency transmit diversity (SFTD) that can achieve space and frequency diversity on a per-subband-pair basis. An exemplary SFTD scheme operates as follows. Suppose that two modulation symbols, denoted as s(k) and s(k+1), are generated and mapped to two adjacent subbands of an OFDM symbol. For SFTD, the transmitter would transmit symbols s(k) and s(k+1) from two antennas on subband k and would transmit symbols s*(k+1) and −s*(k) from the same two antennas on subband k+1. Adjacent subbands are used for the pair of modulation symbols because the channel response is assumed to be constant for the transmission of the two pairs of symbols. The processing at the receiver to recover the modulation symbols is the same as for the STTD scheme, except that the received symbols for two subbands instead of two OFDM symbol periods are processed.
Within each TX subband diversity processor 840, a demultiplexer 842 demultiplexes the modulation symbols for the subband into two symbol sequences, with each sequence having a symbol rate of (2TOFDM)−1. A space-time encoder 850 receives the two modulation symbol sequences and, for each 2-symbol period, uses two symbols s1 and s2 in the two sequences to form two symbol sets {xi}={s1 s2*} and {xj}={s2−s1*} for two transmit antennas. Each symbol set includes two symbols, one symbol from each of the two sequences. Symbol set {xi} is generated by providing symbol s1 first and symbol s2* next, where s1 is obtained via a switch 856a and s2 is obtained by taking the conjugate of s2 with a unit 852a and delaying the conjugated symbol by one symbol period with a delay unit 854a. Similarly, symbol set {xj} is generated by providing symbol s2 first and symbol −s1* next, where s2 is obtained via a switch 856b and −s1* is obtained by taking the negative conjugate of s1 with a unit 852b and delaying the negative conjugated symbol by one symbol period with a delay unit 854b. The two symbol sets {xi} and {xj} are to be transmitted from two antennas i and j assigned to the subband, as indicated in Table 29. Space-time encoder 850 provides the first symbol set {xi}={s1 s2*} to a buffer/multiplexer 870 for the first transmit antenna i and the second symbol set {xi}={s2−s1*} to another buffer/multiplexer 870 for the second transmit antenna j. The two symbols provided by space-time encoder 850 for each symbol period are referred to as STTD symbols.
Buffers/multiplexers 870a through 870d are used to buffer and multiplex the STTD symbols from all diversity processors 840. Each buffer/multiplexer 870 receives pilot symbols and STTD symbols from the appropriate TX subband diversity processors 840, as determined by Table 29. For example, buffer/multiplexer 870a receives modulation symbols for subbands −26, −24, −22, −19, and so on (i.e., all subbands mapped to antenna 1), buffer/multiplexer 870b receives modulation symbols for subbands −26, −23, −20, −19, and so on (i.e., all subbands mapped to antenna 2), buffer/multiplexer 870c receives modulation symbols for subbands −25, −24, −20, −18, and so on (i.e., all subbands mapped to antenna 3), and buffer/multiplexer 870d receives modulation symbols for subbands −25, −23, −22, −18, and so on (i.e., all subbands mapped to antenna 4).
Each buffer/multiplexer 870 then, for each symbol period, multiplexes four pilots, 24 STTD symbols, and 36 zeros for the four pilot subbands, 24 data subbands, and 36 unused subbands, respectively, to form a sequence of 64 transmit symbols for the 64 total subbands. Although there are a total of 48 data subbands, only 24 subbands are used for each transmit antenna for the diversity mode, and the effective total number of unused subbands for each antenna is thus 36 instead of 12. Each transmit symbol is a complex value (which may be zero for an unused subband) that is sent on one subband in one symbol period. Each buffer/multiplexer 870 provides a stream of transmit symbols xi(n) for one transmit antenna. Each transmit symbol stream comprises concatenated sequences of 64 transmit symbols, one sequence for each symbol period. Referring back to
For each transformed symbol, cyclic prefix generator 854 repeats a portion of the transformed symbol to form a corresponding OFDM symbol. As noted above, one of two different cyclic prefix lengths may be used. The cyclic prefix for the BCH is fixed and is 800 nsec. The cyclic prefix for all other transport channels is selectable (either 400 nsec or 800 nsec) and indicated by the Cyclic Prefix Duration field of the BCH message. For a system with a bandwidth of 20 MHz, a sample period of 50 nsec, and 64 subbands, each transformed symbol has a duration of 3.2 msec (or 64×50 nsec) and each OFDM symbol has a duration of either 3.6 msec or 4.0 msec depending on whether the 400 nsec or 800 nsec cyclic prefix is used for the OFDM symbol.
Cyclic prefix generator 854 provides a stream of OFDM symbols to a transmitter (TMTR) 856. Transmitter 856 converts the OFDM symbol stream into one or more analog signals, and further amplifies, filters, and frequency upconverts the analog signals to generate a modulated signal suitable for transmission from an associated antenna.
The baseband waveform for an OFDM symbol may be expressed as:
where
n denotes the symbol period (i.e., the OFDM symbol index);
k denotes the subband index;
NST is the number of pilot and data subbands;
cn(k) denotes the symbol transmitted on subband k of symbol period n; and
where
TCP is the cyclic prefix duration;
TS is the OFDM symbol duration; and
Δf is the bandwidth of each subband.
Spatial Multiplexing Mode—Transmit Processing
Within a TX data processor 710b, framing unit 808 frames the data for each FCH/RCH packet to generate one or more PHY frames for the packet. Each PHY frame includes the number of data bits that may be transmitted in all NE spatial channels within 6 OFDM symbols. Scrambler 810 scrambles the data for each transport channel. Encoder 812 receives and codes the scrambled data in accordance with a selected coding scheme to provide code bits. In an embodiment, a common coding scheme is used to code the data for all NE spatial channels, and different code rates for different spatial channels are obtained by puncturing the code bits with different puncturing patterns. Puncture unit 814 thus punctures the code bits to obtain the desired code rate for each spatial channel. The puncturing for the spatial multiplexing mode is described in further detail below.
A demultiplexer 816 receives and demultiplexes the code bits from puncture unit 814 to provide NE code bit streams for the NE spatial channels selected for use. Each code bit stream is provided to a respective interleaver 818, which interleaves the code bits in the stream across the 48 data subbands. The coding and interleaving for the spatial multiplexing mode are described in further detail below. The interleaved data from each interleaver 818 is provided to a respective symbol mapping unit 820.
In the spatial multiplexing mode, up to four different rates may be used for the four spatial channels, depending on the received SNRs achieved for these spatial channels. Each rate is associated with a particular modulation scheme, as shown in Table 25. Each symbol mapping unit 820 maps the interleaved data in accordance with a particular modulation scheme selected for the associated spatial channel to provide modulation symbols. If all four spatial channels are selected for use, then symbol mapping units 820a through 820d provide four streams of modulation symbols for the four spatial channels to a TX spatial processor 720b.
TX spatial processor 720b performs spatial processing for the spatial multiplexing mode. For simplicity, the following description assumes that four transmit antennas, four receive antennas, and 48 data subbands are used for data transmission. The data subband indices are given by the set K, where K=±{1, . . . , 6, 8 . . . , 20, 22, . . . 26} for the OFDM subband structure described above.
The model for a MIMO-OFDM system may be expressed as:
r
(k)=H(k)x(k)+n(k), for kεK, Eq. (5)
The channel response matrix H(k) for subband k may be expressed as:
where entry hij(k), for iε{1, 2, 3, 4} and jε{1, 2, 3, 4}, is the coupling (i.e., complex gain) between transmit antenna i and receive antenna j for subband k. For simplicity, it is assumed that the channel response matrices H(k), for kεK, are known or can be ascertained by both the transmitter and receiver.
The channel response matrix H(k) for each subband may be “diagonalized” to obtain the NS eigenmodes for that subband. This can be achieved by performing eigenvalue decomposition on the correlation matrix of H(k), which is R(k)=HH(k)H(k), where HH(k) denotes the conjugate transpose of H(k). The eigenvalue decomposition of the correlation matrix R(k) may be expressed as:
R
(k)=V(k)D(k)VH(k), for kεK, Eq. (7)
where
The channel response matrix H(k) may also be diagonalized using singular value decomposition, which may be expressed as:
H
(k)=U(k)Σ(k)VH(k), for kεK, Eq. (8)
where
The diagonal matrix D(k) for each subband contains non-negative real values along the diagonal and zeros everywhere else. The eigenvalues of R(k) are denoted as {λ1(k), λ2(k), λ3(k), λ4(k)} or {λi(k)} for iε{1, 2, 3, 4}.
The eigenvalue decomposition may be performed independently for the channel response matrix H(k) for each of the 48 data subbands to determine the four eigenmodes for that subband (assuming that each matrix H(k) is full rank). The four eigenvalues for each diagonal matrix D(k) may be ordered such that {λ1(k)≧λ2(k)≧λ3(k)≧λ4(k)}, where λ1(k) is the largest eigenvalue and λ4(k) is the smallest eigenvalue for subband k. When the eigenvalues for each diagonal matrix D(k) are ordered, the eigenvectors (or columns) of the associated matrix V(k) are also ordered correspondingly.
A “wideband” eigenmode may be defined as the set of same-order eigenmodes of all subbands after the ordering (i.e., wideband eigenmode m includes eigenmodes m of all subbands). Each wideband eigenmode is associated with a respective set of eigenvectors for all of the subbands. The “principal” wideband eigenmode is the one associated with the largest singular value in each of the matrices {circumflex over (Σ)}(k) after the ordering.
A vector dm may then be formed to include the m-th rank eigenvalue for all 48 data subbands. This vector dm may be expressed as:
d
m=[λm(−26) . . . λm(−22) . . . λm(22) . . . λm(26)], for m={1, 2, 3, 4}. Eq (9)
The vector d1 includes the eigenvalues for the best or principal wideband eigenmode. For a MIMO-OFDM system with four transmit antennas and four receive antennas (i.e., a 4×4 system), there are up four wideband eigenmodes.
If the noise variance σ2 at the receiver is constant across the operating band and known to the transmitter, then the received SNR for each subband of each wideband eigenmode may be determined by dividing the eigenvalue λm(k) by the noise variance σ2. For simplicity, the noise variance can be assumed to be equal to one (i.e., σ2=1).
For the spatial multiplexing mode, the total transmit power Ptotal available to the transmitter may be distributed to the wideband eigenmodes based on various power allocation schemes. In one scheme, the total transmit power Ptotal is distributed uniformly to all four wideband eigenmodes such that Pm=Ptotal/4, where Pm is the transmit power allocated to wideband eigenmode m. In another scheme, the total transmit power Ptotal is distributed to the four wideband eigenmodes using a water-filling procedure.
The water-filling procedure distributes power such that the wideband eigenmodes with higher power gains receive greater fractions of the total transmit power. The amount of transmit power allocated to a given wideband eigenmode is determined by its received SNR, which in turn is dependent on the power gains (or eigenvalues) for all of the subbands of that wideband eigenmode. The water-filling procedure may allocate zero transmit power to wideband eigenmodes with sufficiently poor received SNRs. The water-filling procedure receives β={β1, β2, β3, β4} for the four wideband eigenmodes, where βm is a normalization factor for wideband eigenmode m and may be expressed as:
The normalization factor βm keeps the transmit power allocated to wideband eigenmode m invariant after channel inversion is applied, as described below. As shown in equation (10), the normalization factor βm can be derived based on the eigenvalues in the vector dm and with the assumption of the noise variance being equal to one (i.e., σ2=1).
The water-filling procedure then determines the fraction αm of the total transmit power to allocate to each wideband eigenmode based on the set β such that spectral efficiency or some other criterion is optimized. The transmit power allocated to wideband eigenmode m by the water-filling procedure may be expressed as:
P
m=αmPtotal, for m={1, 2, 3, 4}. Eq. (11)
The power allocations for the four wideband eigenmodes may be given by α={α1, α2, α3, α4}, where
The spatial multiplexing mode may be selected for use if more than one value in set α is non-zero.
The procedure for performing water-filling is known in the art and not described herein. One reference that describes water-filling is “Information Theory and Reliable Communication,” by Robert G. Gallager, John Wiley and Sons, 1968, which is incorporated herein by reference.
For the spatial multiplexing mode, the rate for each spatial channel or wideband eigenmode may be selected based on the received SNR achieved by that spatial channel/wideband eigenmode with its allocated transmit power of Pm. For simplicity, the following description assumes data transmission on the wideband eigenmodes. The received SNR for each wideband eigenmode may be expressed as:
In one embodiment, the rate for each wideband eigenmode is determined based on a table that includes the rates supported by the system and a range of SNRs for each rate. This table may be obtained by computer simulation, empirical measurements, and so on. The particular rate to use for each wideband eigenmode is the rate in the table with a range of SNRs covering the received SNR for the wideband eigenmode. In another embodiment, the rate for each wideband eigenmode is selected based on (1) the received SNR for the wideband eigenmode, (2) an SNR offset used to account for estimation error, variability in the MIMO channel, and other factors, and (3) a table of supported rates and their required SNRs. For this embodiment, an average received SNR for each wideband eigenmode is first computed as described above or as an average of the received SNRs (in units of dBs) for all of the subbands of the wideband eigenmode. In any case, an operating SNR is next computed as the sum of the received SNR and the SNR offset (where both are given in units of dBs). The operating SNR is then compared against the required SNR for each of the rates supported by the system. The highest rate in the table with a required SNR that is less than or equal to the operating SNR is then selected for the wideband eigenmode. The rate for the transmit diversity mode and the beam-steering mode may also be determined in similar manner.
The transmit power Pm allocated for each wideband eigenmode may be distributed across the 48 data subbands of that wideband eigenmode such that the received SNRs for all subbands are approximately equal. This non-uniform allocation of power across the subbands is referred to as channel inversion. The transmit power Pm(k) allocated to each subband may be expressed as:
where βm is given in equation (10).
As shown in equation (13), the transmit power Pm is distributed non-uniformly across the data subbands based on their channel power gains, which is given by the eigenvalues λm(k), for kεK. The power distribution is such that approximately equal received SNRs are achieved at the receiver for all data subbands of each wideband eigenmode. This channel inversion is performed independently for each of the four wideband eigenmodes. The channel inversion per wideband eigenmode is described in further detail in commonly assigned U.S. patent application Ser. No. 10/229,209, entitled “Coded MIMO Systems with Selective Channel Inversion Applied Per Eigenmode,” filed Aug. 27, 2002.
The channel inversion may be performed in various manners. For full channel inversion, all data subbands are used for data transmission if a wideband eigenmode is selected for use. For selective channel inversion, all or a subset of the available data subbands may be selected for use for each wideband eigenmode. The selective channel inversion discards poor subbands with received SNR below a particular threshold and performs channel inversion on only the selected subbands. Selective channel inversion for each wideband eigenmode is also described in commonly assigned U.S. patent application Ser. No. 10/229,209, entitled “Coded MIMO Systems with Selective Channel Inversion Applied Per Eigenmode,” filed Aug. 27, 2002. For simplicity, the following description assumes that full channel inversion is performed for each wideband eigenmode selected for use.
The gain to use for each subband of each wideband eigenmode may be determined based on the transmit power Pm(k) allocated to that subband. The gain gm(k) for each data subband may be expressed as:
g
m(k)=√{square root over (Pm(k))}, for kεK and m={1, 2, 3, 4}. Eq. (14)
A diagonal gain matrix G(k) may be defined for each subband. This matrix G(k) includes the gains for the four eigenmodes for subband k along the diagonal, and may be expressed as: G(k)=diag[g1(k), g2(k), g3(k), g4(k)].
For the spatial multiplexing mode, the transmit vector x(k) for each data subband may be expressed as:
x
(k)=V(k)G(k)s(k), for kεK, Eq. (15)
where
s(k)=[s1(k) s2(k) s3(k) s4(k)]T, and
x(k)=[x1(k) x2(k) x3(k) x4(k)]T.
The vector s(k) includes four modulation symbols to be transmitted on the four eigenmodes for subband k, and the vector x(k) includes four transmit symbols to be transmitted from the four antennas for subband k. For simplicity, equation (15) does not include the correction factors used to account for differences between the transmit/receive chains at the access point and the user terminal, which are described in detail below.
Within processor 720b, a demultiplexer 932 receives the four modulation symbol streams (denoted as s1(n) through s4(n)) to be transmitted on the four wideband eigenmodes, demultiplexes each stream into 48 substreams for the 48 data subbands, and provides four modulation symbol substreams for each data subband to a respective TX subband spatial processor 940. Each processor 940 performs the processing shown in equation (15) for one subband.
Within each TX subband spatial processor 940, the four modulation symbol substreams (denoted as s1(k) through s4(k)) are provided to four multipliers 942a through 942d, which also receive the gains g1(k), g2(k), g3(k), and g4(k) for the four eigenmodes of the associated subband. Each gain gm(k) may be determined based on the transmit power Pm(k) allocated to that subband/eigenmode, as shown in equation (14). Each multiplier 942 scales its modulation symbols with its gain gin (k) to provide scaled modulation symbols. Multipliers 942a through 942d provide four scaled modulation symbol substreams to four beam-formers 950a through 950d, respectively.
Each beam-former 950 performs beam-forming to transmit one symbol substream on one eigenmode of one subband. Each beam-former 950 receives one symbol substream sm(k) and one eigenvector vm(k) for the associated eigenmode. In particular, beam-former 950a receives eigenvector v1(k) for the first eigenmode, beam-former 950b receives eigenvector v2(k) for the second eigenmode, and so on. The beam-forming is performed using the eigenvector for the associated eigenmode.
Within each beam-former 950, the scaled modulation symbols are provided to four multipliers 952a through 952d, which also receive four elements, νm,1(k), νm,2(k), νm,3(k), and νm,4(k), of eigenvector vm(k) for the associated eigenmode. Each multiplier 952 then multiplies the scaled modulation symbols with its eigenvector value νm,j(k) to provide “beam-formed” symbols. Multipliers 952a through 952d provide four beam-formed symbol substreams (which are to be transmitted from four antennas) to summers 960a through 960d, respectively.
Each summer 960 receives and sums four beam-formed symbols for the four eigenmodes for each symbol period to provide a preconditioned symbol for an associated transmit antenna. Summers 960a through 960d provide four substreams of preconditioned symbols for four transmit antennas to buffers/multiplexers 970a through 970d, respectively.
Each buffer/multiplexer 970 receives pilot symbols and the preconditioned symbols from TX subband spatial processors 940a through 940k for the 48 data subbands. Each buffer/multiplexer 970 then, for each symbol period, multiplexes 4 pilot symbols, 48 preconditioned symbols, and 12 zeros for 4 pilot subbands, 48 data subbands, and 12 unused subbands, respectively, to form a sequence of 64 transmit symbols for that symbol period. Each buffer/multiplexer 970 provides a stream of transmit symbols xi(n) for one transmit antenna, where the transmit symbol stream comprises concatenated sequences of 64 transmit symbols. The transmit symbols can be scaled with correction factors to account for differences between the transmit/receive chains at the access point and the user terminal, as described below. The subsequent OFDM modulation for each transmit symbol stream is described above.
Parallel symbol streams may also be transmitted from the four transmit antennas without spatial processing at the access point using the non-steered spatial multiplexing mode. For this mode, the channel inversion process and beam-forming by beam-former 950 may be omitted. Each modulation symbol stream is further OFDM processed and transmitted from a respective transmit antenna.
The non-steered spatial multiplexing mode may be used for various situations such as if the transmitter is unable to perform the spatial processing necessary to support beam-steering based on eigenmode decomposition. This may be because the transmitter has not performed calibration procedures, is unable to generate a sufficiently good estimate of the channel, or does not have calibration and eigenmode processing capabilities at all. For the non-steered spatial multiplexing mode, spatial multiplexing is still used to increase the transmission capacity, but the spatial processing to separate out the individual symbol streams is performed by the receiver.
For the non-steered spatial multiplexing mode, the receiver performs the spatial processing to recover the transmitted symbol streams. In particular, a user terminal may implement a channel correlation matrix inversion (CCMI) technique, a minimum mean square error (MMSE) technique, a successive interference cancellation receiver processing technique, or some other receiver spatial processing technique. These techniques are described in detail in commonly assigned U.S. patent application Ser. No. 09/993,087, entitled “Multiple-Access Multiple-Input Multiple-Output (MIMO) Communication System,” filed Nov. 6, 2001. The non-steered spatial multiplexing mode may be used for both downlink and uplink transmissions.
The multi-user spatial multiplexing mode supports data transmission to multiple user terminals simultaneously on the downlink based on the “spatial signatures” of the user terminals. The spatial signature for a user terminal is given by a channel response vector (for each subband) between the access point antennas and each user terminal antenna. The access point may obtain the spatial signatures, for example, based on the steered reference transmitted by the user terminals. The access point may process the spatial signatures for user terminals desiring data transmission to (1) select a set of user terminals for simultaneous data transmission on the downlink and (2) derive steering vectors for each of the independent data streams to be transmitted to the selected user terminals.
The steering vectors for the multi-user spatial multiplexing mode may be derived in various manners. Two exemplary schemes are described below. For simplicity, the following description is for one subband and assumes that each user terminal is equipped with a single antenna.
In a first scheme, the access point obtains the steering vectors using channel inversion. The access point may select Nap single-antenna user terminals for simultaneous transmission on the downlink. The access point obtains an 1×Nap channel response row vector for each selected user terminal and forms an Nap×Nap channel response matrix Hmu with the Nap row vectors for the Nap user terminals. The access point then obtains a matrix Hsteer of Nap steering vectors for the Nap selected user terminals as Hsteer=Hmu−1. The access point can also transmit a steered reference to each selected user terminal Each user terminal processes its steered reference to estimate the channel gain and phase and coherently demodulates received symbols for its single antenna with the channel gain and phase estimates to obtain recovered symbols.
In a second scheme, the access point precodes Nap symbol streams to be sent to Nap user terminals such that these symbol streams experience little cross-talk at the user terminals. The access point can form the channel response matrix Hmu for the Nap selected user terminals and perform QR factorization on Hmu such that Hmu=FtriQmu, where Ttri is a lower left triangular matrix Ttri and Qmu is a unitary matrix. The access point then precodes the Nap data symbol streams with the matrix Ttri to obtain Nap precoded symbol streams a, and further processes the precoded symbol streams with the unitary matrix Qmu to obtain the Nap transmit symbol streams for transmission to the Nap user terminals. Again, the access point can also transmit a steered reference to each user terminal Each user terminal uses the steered reference to coherently demodulate its received symbols to obtain recovered symbols.
For the uplink in the multi-user spatial multiplexing mode, the access point can recover Nap symbol streams transmitted simultaneously by Nap user terminals using MMSE receiver processing, successive interference cancellation, or some other receiver processing technique. The access point can estimate the uplink channel response for each user terminal and use the channel response estimate for receiver spatial processing and for scheduling uplink transmissions. Each single-antenna user terminal can transmit an orthogonal pilot on the uplink. The uplink pilots from the Nap user terminals can be orthogonal in time and/or frequency. Time orthogonality can be achieved by having each user terminal covers its uplink pilot with an orthogonal sequence assigned to the user terminal. Frequency orthogonality can be achieved by having each user terminal transmits its uplink pilot on a different set of subbands. The uplink transmissions from the user terminals should be approximately time-aligned at the access point (e.g., time-aligned to within the cyclic prefix).
Beam-Steering Mode—Transmit Processing
Within a TX data processor 710c, framing unit 808 frames the data for each FCH/RCH packet to generate one or more PHY frames for the packet. Scrambler 810 then scrambles the data for each transport channel. Encoder 812 next codes the framed data in accordance with a selected coding scheme to provide code bits. Puncture unit 814 then punctures the code bits to obtain the desired code rate for the wideband eigenmode used for data transmission. The code bits from puncture unit 818 are interleaved across all data subbands. Symbol mapping unit 820 then maps the interleaved data in accordance with a selected modulation scheme to provide modulation symbols. A TX spatial processor 720c then performs transmit processing on the modulation symbols for the beam-steering mode.
The beam-steering mode may be used to transmit data on one spatial channel or wideband eigenmode—typically the one associated with the largest eigenvalues for all of the data subbands. The beam-steering mode may be selected if the transmit power allocation to the wideband eigenmodes results in only one entry in the set α being non-zero. Whereas the spatial multiplexing mode performs beam-forming for each of the selected eigenmodes of each subband based on its eigenvector, the beam-steering mode performs beam-steering based on a “normalized” eigenvector for the principal eigenmode of each subband to transmit data on that single eigenmode.
The four elements of each eigenvector v1(k), for kεK, for the principal eigenmode may have different magnitudes. The four preconditioned symbols obtained based on the four elements of eigenvector v1(k) for each subband may then have different magnitudes. Consequently, the four per-antenna transmit vectors, each of which includes the preconditioned symbols for all data subbands for a given transmit antenna, may have different magnitudes. If the transmit power for each transmit antenna is limited (e.g., because of limitations of the power amplifiers), then the beam-forming technique may not fully use the total power available for each antenna.
The beam-steering mode uses only the phase information from eigenvectors v1(k), for kεK, for the principal eigenmode and normalizes each eigenvector such that all four elements in the eigenvector have equal magnitudes. The normalized eigenvector {tilde over (v)}(k) for subband k may be expressed as:
{tilde over (v)}(k)=[Aejθ
where A is a constant (e.g., A=1); and
θi(k) is the phase for subband k of transmit antenna i, which is given as:
As shown in equation (17), the phase of each element in the vector {tilde over (v)}(k) is obtained from the corresponding element of eigenvector v1(k) (i.e., θi(k) is obtained from ν1,i(k), where v1(k)=[ν1,1(k) ν1,2(k) ν1,3(k) ν1,4(k)]T).
Channel inversion may also be performed for the beam-steering mode so that a common rate can be used for all data subbands. The transmit power {tilde over (P)}1(k) allocated to each data subband for the beam-steering mode may be expressed as:
where
The transmit power {tilde over (P)}1 may be given as P1=Ptotal/4 (i.e., uniform allocation of the total transmit power across the four transmit antennas). The power gain {tilde over (λ)}1(k) may be expressed as:
{tilde over (λ)}1(k)={tilde over (v)}H(k)HH(k)H(k){tilde over (v)}(k). Eq. (20)
The channel inversion results in power allocation of {tilde over (P)}1(k), for kεK, for the 48 data subbands. The gain for each data subband may then be given as {tilde over (g)}(k)=√{square root over ({tilde over (P)}1(k))}.
For the beam-steering mode, the transmit vector x(k) for each subband may be expressed as:
x
(k)={tilde over (v)}(k){tilde over (g)}(k)s(k), for kεK. Eq. (21)
Again for simplicity, equation (21) does not include the correction factors used to account for differences between the transmit/receive chains at the access point and the user terminal.
As shown in equation (16), the four elements of the normalized steering vector {tilde over (v)}(k) for each subband have equal magnitude but possibly different phases. The beam-steering thus generates one transmit vector X(k) for each subband, with the four elements of x(k) having the same magnitude but possibly different phases.
Within processor 720c, a demultiplexer 1032 receives and demultiplexes the modulation symbol stream s(n) into 48 substreams for the 48 data subbands (denoted as s(1) through s(k). Each symbol substream is provided to a respective TX subband beam-steering processor 1040. Each processor 1040 performs the processing shown in equation (14) for one subband.
Within each TX subband beam-steering processor 1040, the modulation symbol substream is provided to a multiplier 1042, which also receives the gain {tilde over (g)}(k) for the associated subband. Multiplier 1042 then scales the modulation symbols with the gain {tilde over (g)}(k) to obtain scaled modulation symbols, which are then provided to a beam-steering unit 1050.
Beam-steering unit 1050 also receives the normalized eigenvector {tilde over (v)}(k) for the associated subband. Within beam-steering unit 1050, the scaled modulation symbols are provided to four multipliers 1052a through 1052d, which also respectively receive the four elements, {tilde over (ν)}1(k), {tilde over (ν)}2(k), {tilde over (ν)}3(k), and {tilde over (v)}4(k), of the normalized eigenvector {tilde over (v)}(k). Each multiplier 1052 multiplies its scaled modulation symbols with its normalized eigenvector value {tilde over (ν)}i(k) to provide preconditioned symbols. Multipliers 1052a through 1052d provide four preconditioned symbol substreams to buffers/multiplexers 1070a through 1070d, respectively.
Each buffer/multiplexer 1070 receives pilot symbols and the preconditioned symbols from TX subband beam-steering processors 1040a through 1040k for the 48 data subbands, multiplexes the pilot and preconditioned symbols and zeros for each symbol period, and provides a stream of transmit symbols xi(n) for one transmit antenna. The subsequent OFDM modulation for each transmit symbol stream is described above.
The processing for the beam-steering mode is described in further detail in commonly assigned U.S. patent application Ser. No. 10/228,393, entitled “Beam-Steering and Beam-Forming for Wideband MIMO Systems,” filed Aug. 27, 2002. The system may also be designed to support a beam-forming mode whereby a data stream is transmitted on the principal eigenmode using the eigenvector instead of the normalized eigenvector.
Framing for PHY Frames
For the diversity and beam-steering modes, only one spatial channel or wideband eigenmode is used for data transmission. The rate for this mode is known, and the number of information bits that may be sent in the payload of each PHY frame may be computed. For the spatial multiplexing mode, multiple spatial channels may be used for data transmission. Since the rate of each spatial channel is known, the number of information bits that may be sent in the payload of each PHY frame for all spatial channels may be computed.
As shown in
Scrambling
In an embodiment, the data bits for each transport channel are scrambled prior to coding. The scrambling randomizes the data so that a long sequence of all ones or all zeros is not transmitted. This can reduce the variation in the peak to average power of the OFDM waveform. The scrambling may be omitted for one or more transport channels and may also be selectively enabled and disabled.
G(x)=X7+x4+x. Eq. (22)
Other generator polynomials may also be used, and this is within the scope of the invention.
As shown in
The framed/unframed bits (d1 d2 d3 d4 . . . ) are provided to an adder 1116, which also receives scrambling bits from adder 1114. Adder 1116 performs modulo-2 addition of each bit dn with a corresponding scrambling bit to provide a scrambled bit qn. Scrambler 810 provides a sequence of scrambled bits, which is denoted as q1 q2 q3 q4 . . . .
The initial state of the scrambler (i.e., the content of delay elements 1112a through 1112g) is set to a 7-bit non-zero number at the start of each TDD frame. The three most significant bits (MSBs) (i.e., delay element 1112e through 11120 are always set to one (‘1’) and the four least significant bits (LSBs) are set to the TDD frame counter, as indicated in the BCH message.
Encoding/Puncturing
In an embodiment, a single base code is used to code data prior to transmission. This base code generates code bits for one code rate. All other code rates supported by the system (as listed in Table 25) may be obtained by either repeating or puncturing the code bits.
Within encoder 812, a multiplexer 1120 receives and multiplexes the scrambled bits and tail bits (e.g., zeros). Encoder 812 further includes six delay elements 1122a through 1122f coupled in series. Four adders 1124a through 1124d are also coupled in series and used to implement the first generator (133). Similarly, four adders 1126a through 1126d are coupled in series and used to implement the second generator (171). The adders are further coupled to the delay elements in a manner to implement the two generators of 133 and 171, as shown in
The scrambled bits are provided to the first delay element 1122a and to adders 1124a and 1126a. For each clock cycle, adders 1124a through 1124d perform modulo-2 addition of the incoming bit and four prior bits stored in delay elements 1122b, 1122c, 1122e, and 1122f to provide the first code bit for that clock cycle. Similarly, adders 1126a through 1126d perform modulo-2 addition of the incoming bit and four prior bits stored in delay elements 1122a, 1122b, 1122c, and 1122f to provide the second code bit for that clock cycle. The code bits generated by the first generator are denoted as a1 a2 a3 a4 . . . , and the code bits generated by the second generator are denoted as b1 b2 b3 b4 . . . . A multiplexer 1128 then receives and multiplexes the two streams of code bits from the two generators into a single stream of code bits, which is denoted as a1 b1 a2 b2 a3 b3 a4 b4 . . . . For each scrambled bit qn, two code bits an and bn are generated, which results in a code rate of ½.
Table 30 lists exemplary puncturing patterns that may be used for the various code rates supported by the system. Other puncturing patterns may also be used, and this is within the scope of the invention.
To obtain a code rate of k/n, puncturing unit 1134 provides n code bits for each group of 2 k rate ½ code bits received from encoder 812. Thus, 2 k−n code bits are deleted from each group of 2 k code bits. The bits to be deleted from each group are denoted by zeros in the puncturing pattern. For example, to obtain a code rate of 7/12, two bits are deleted from each group of 14 code bits from encoder 812, with the deleted bits being the 8th and 14th code bits in the group, as denoted by the puncturing pattern of “11111110111110”. No puncturing is performed if the desired code rate is ½.
A multiplexer 1136 receives the stream of code bits from repeating unit 1132 and the stream of code bits from puncturing unit 1134. Multiplexer 1136 then provides the code bits from repeating unit 1132 if the desired code rate is ¼ and the code bits from puncturing unit 1134 if the desired code rate is ½ or higher.
Other codes and puncturing patterns besides those described above may also be used, and this is within the scope of the invention. For example, a Turbo code, a block code, some other codes, or any combination thereof may be used to code data. Also, different coding schemes may be used for different transport channels. For example, convolutional coding may be used for the common transport channels, and Turbo coding may be used for the dedicated transport channels.
Interleaving
In an embodiment, the code bits to be transmitted are interleaved across the 48 data subbands. For the diversity and beam-steering modes, one stream of code bits is transmitted and interleaved across all data subbands. For the spatial multiplexing mode, up to four streams of code bits may be transmitted on up to four spatial channels. The interleaving may be performed separately for each spatial channel such that each stream of code bits is interleaved across all data subbands of the spatial channel used to transmit that stream. Table 29 shows an exemplary code bit-subband assignment that may be used for the interleaving for all transmission modes.
In one embodiment, the interleaving is performed across all 48 data subbands in each interleaving interval. For this embodiment, each group of 48 code bits in a stream is spread over the 48 data subbands to provide frequency diversity. The 48 code bits in each group may be assigned indices of 0 through 47. Each code bit index is associated with a respective subband. All code bits with a particular index are transmitted on the associated subband. For example, the first code bit (with index 0) in each group is transmitted on subband −26, the second code bit (with index 1) is transmitted on subband 1, the third code bit (with index 2) is transmitted on subband −17, and so on. This interleaving scheme may be used for the diversity, beam-steering, and spatial multiplexing modes. An alternative interleaving scheme for the spatial multiplexing mode is described below.
The interleaving may alternatively or additionally be performed over time. For example, after the interleaving across the data subbands, the code bits for each subband may further be interleaved (e.g., over one PHY frame or one PDU) to provide time diversity. For the spatial multiplexing mode, the interleaving may also be performed over multiple spatial channels.
Additionally, interleaving may be employed across the dimensions of the QAM symbols such that code bits forming QAM symbols are mapped to different bit positions of the QAM symbols.
Symbol Mapping
Table 31 shows the symbol mapping for various modulation schemes supported by the system. For each modulation scheme (except for BPSK), half of the bits are mapped to an inphase (I) component and the other half of the bits are mapped to a quadrature (Q) component.
In an embodiment, the signal constellation for each supported modulation scheme may be defined based on Gray mapping. With Gray mapping, neighboring points in the signal constellation (in both the I and Q components) differ by only one bit position. Gray mapping reduces the number of bit errors for more likely error events, which corresponds to a received symbol being mapped to a location near the correct location, in which case only one code bit would be received in error.
The I and Q values for each modulation scheme shown in Table 31 are scaled by a normalization factor Knorm so that the average power of all signal points in the associated signal constellation is equal to unity. The normalization factor for each modulation scheme is shown in Table 31. Quantized values for the normalization factors for the supported modulation schemes may also be used. A modulation symbol s from a particular signal constellation would then have the following form:
s=(I+jQ)·Knorm,
where I and Q are the values in Table 31 for the signal constellation.
For a given PDU, the modulation may be different across the PDU and may be different for multiple spatial channels used for data transmission. For example, for the BCH PDU, different modulation schemes may be used for the beacon pilot, the MIMO pilot, and the BCH message.
Processing for Spatial Multiplexing Mode
For the spatial multiplexing mode, a PDU may be transmitted over multiple spatial channels. Various schemes may be used to process data for transmission over multiple spatial channels. Two specific processing schemes for the spatial multiplexing mode are described below.
In the first processing scheme, coding and puncturing are performed on a per spatial channel basis to achieve the desired code rate for each spatial channel. The NE spatial channels to use for data transmission are ranked from the highest to lowest received SNR. The data for the entire PDU is first coded to obtain a stream of rate ½ code bits. The code bits are then punctured to obtain the desired code rate for each spatial channel.
The puncturing may be performed in sequential order for the NE spatial channels, from the best (i.e., highest SNR) to the worst (i.e., lowest SNR) spatial channel. In particular, the puncture unit first performs puncturing for the best spatial channel with the highest received SNR. When the correct number of code bits have been generated for the best spatial channel, the puncture unit then performs puncturing for the second best spatial channel with the next highest received SNR. This process continues until the code bits for all NE spatial channels are generated. The order for puncturing is from the largest to smallest received SNR, regardless of the specific code rate used for each spatial channel.
For the example shown in Table 28, the 3456 information bits to be transmitted in the overall PHY frame are first coded with the rate ½ base code to obtain 6912 code bits. The first 3168 code bits are punctured using the puncturing pattern for code rate 11/16 to obtain 2304 code bits, which are provided in the PHY frame for the first spatial channel. The next 2592 code bits are then punctured using the puncturing pattern for code rate ¾ to obtain 1728 code bits, which are provided in the PHY frame for the second spatial channel. The next 864 code bits are then punctured using the puncturing pattern for code rate ¾ to obtain 576 code bits, which are provided in the PHY frame for the third spatial channel. The last 288 code bits for the PHY frame are then punctured using the puncturing pattern for code rate ½ to obtain 288 code bits, which are provided in the PHY frame for the last spatial channel. These four individual PHY frames are further processed and transmitted on the four spatial channels. The puncturing for the next overall PHY frame is then performed in similar manner. The first processing scheme may be implemented by TX data processor 710b in
In the second processing scheme, the coding and puncturing are performed for pairs of subbands. Moreover, the coding and puncturing are cycled through all selected spatial channels for each pair of subbands.
Encoder 812 provides a stream of rate ½ code bits to demultiplexer 816, which demultiplexes the received code bit stream into four substreams for the four spatial channels. The demultiplexing is such that the first 4b1r1 code bits are sent to buffer 813a for spatial channel 1, the next 4b2r2 code bits are sent to buffer 813b for spatial channel 2, and so on. Each buffer 813 receives 4bmrm code bits each time demultiplexer 816 cycles through all four spatial channels. A total of
rate ½ code bits are provided to the four buffers 813a through 813d for each cycle. Demultiplexer 816 thus cycles through all four positions for the four spatial channels for every btotal code bits, which is the number of code bits that can be sent on a pair of subbands using all four spatial channels.
Once each buffer 813 has been filled with 4bmrm code chips for the associated spatial channel, the code bits in the buffer are punctured to obtain the code rate for that spatial channel. Since 4bmrm rate ½ code bits span an integer number of puncturing periods for each puncturing pattern, exactly 2bm code bits are provided after the puncturing for each spatial channel m. The 2bm code bits for each spatial channel are then distributed (interleaved) over the data subbands.
In an embodiment, the interleaving is performed for each spatial channel in groups of 6 subbands at a time. The code bits after the puncturing for each spatial channel may be numbered sequentially as ci, for i=0, 1, 2, . . . . A counter Cm may be maintained for each spatial channel to count every group of 6bm code bits provided by the puncturing unit for that spatial channel. For example, for QPSK with bm=2, the counter would be set to Cm=0 for code bits c0 through c11 provided by the puncturing unit, Cm=1 after code bits c12 through c23 and so on. The counter value Cm for spatial channel m may be expressed as:
C
m
=└i/(6bm)┘ mod 8. Eq. (23)
To determine the subband to which code bit ci is assigned, the bit index for the code bit is first determined as follows:
bit index=(i mod 6)+6·Cm. Eq. (24)
The bit index is then mapped to the corresponding subband using Table 29.
For the example above, the first group of 6 code bits c0 through c5 is associated with bit indices 0 through 5, respectively, the second group of 6 code bits c6 through c11 is also associated with bit indices 0 through 5, respectively. Code bits c0 and c6 would be mapped to subband −26, code bits c1 and c7 would be mapped to subband 1, and so on, as shown in Table 29. The spatial processing may then commence for this first group of 6 subbands. The third group of 6 code bits c12 through c17 (with Cm=1) s associated with bit indices 6 through 11, respectively, and the fourth group of 6 code bits c18 through c23 is also associated with bit indices 6 through 11, respectively. Code bits c12 and c18 would be mapped to subband −25, code bits c13 and c19 would be mapped to subband 2, and so on. The spatial processing may then commence for this next group of 6 subbands.
The number 6 in equation (24) comes from the fact that the interleaving is performed in groups of six subbands. The (mod 8) operation in equation (23) comes from the fact that there are eight interleaving groups for the 48 data subbands. Since each cycle of demultiplexer 816 shown in
The interleaving in groups of 6 subbands at a time can reduce processing delays. In particular, the spatial processing can commence once each group of 6 subbands is available.
In alternative embodiments, the interleaving may be performed for each spatial channel in groups of NB subbands at a time, where NB may be any integer (e.g., NB may be equal to 48 for interleaving over all 48 data subbands).
For a TDD system, the downlink and uplink share the same frequency band in a time division duplexed manner. In this case, a high degree of correlation typically exists between the downlink and uplink channel responses. This correlation may be exploited to simplify the channel estimation and spatial processing. For a TDD system, each subband of the wireless link may be assumed to be reciprocal. That is, if H(k) represents the channel response matrix from antenna array A to antenna array B for subband k, then a reciprocal channel implies that the coupling from array B to array A is given by the transpose of H(k), which is HT(k).
However, the responses (gain and phase) of the transmit and receive chains at the access point are typically different from the responses of the transmit and receive chains at the user terminal Calibration may be performed to determine the difference in the frequency responses of the transmit/receive chains at the access point and user terminal, and to account for the difference, such that the calibrated downlink and uplink responses can be expressed in terms of each other. Once the transmit/receive chains have been calibrated and accounted for, a measurement for one link (e.g., the downlink) may be used to derive steering vectors for the other link (e.g., the uplink).
The “effective” downlink and uplink channel responses, Hdn(k) and Hup(k), which include the responses of the applicable transmit and receive chains at the access point and user terminal, may be expressed as:
H
dn(k)=Rut(k)H(k)Tap(k), for kεK, and Eq. (25)
H
up(k)=Rap(k)HT(k)Tut(k), for kεK,
where
Combining the two equations in equation set (25), the following relationship may be obtained:
H
up(k)Kut(k)=(Hdn(k)Kap(k))T, for kεK, Eq. (26)
where Kut(k)=Tut−1(k)Rut(k) and Kap(k)=Tap−1(k)Rap(k).
The left-hand side of equation (26) represents the “true” calibrated channel response on the uplink, and the right-hand side represents the transpose of the “true” calibrated channel response on the downlink. The application of the diagonal matrices Kap(k) and Kut(k) to the effective downlink and uplink channel responses, respectively, as shown in equation (26), allows the calibrated channel responses for the downlink and uplink to be expressed as transposes of each other. The (Nap×Nap) diagonal matrix Kap(k) for the access point is the ratio of the receive chain response Rap(k) to the transmit chain response Tap(k) (i.e.,
where the ratio is taken element-by-element. Similarly, the (Nut×Nut) diagonal matrix Kut(k) for the user terminal is the ratio of the receive chain response Rut(k) to the transmit chain response Tut(k).
The matrices Kap(k) and Kut(k) include values that can account for differences in the transmit/receive chains at the access point and the user terminal. This would then allow the channel response for one link to be expressed by the channel response for the other link, as shown in equation (26).
Calibration may be performed to determine the matrices Kap(k) and Kut(k). Typically, the true channel response H(k) and the transmit/receive chain responses are not known nor can they be exactly or easily ascertained. Instead, the effective downlink and uplink channel responses, Hdn(k) and Hup(k), may be estimated based on pilots sent on the downlink and uplink, respectively, as described below. Estimates of the matrices Kap(k) and Kut(k), which are referred to as correction matrices Ĥap(k) and {circumflex over (K)}ut(k), may then be derived based on the downlink and uplink channel response estimates, Ĥdn(k) and Ĥup(k), as described below. The matrices {circumflex over (K)}ap(k) and {circumflex over (K)}ut(k) include correction factors that can account for differences in the transmit/receive chains at the access point and the user terminal.
The “calibrated” downlink and uplink channel responses observed by the user terminal and the access point, respectively, may then be expressed as:
H
cdn(k)=Hdn(k){circumflex over (K)}ap(k), for kεK, and Eq. (27)
H
cup(k)=Hup(k){circumflex over (K)}ut(k), for kεK,
The calibration may be performed using various schemes. For clarity, a specific calibration scheme is described below. To perform the calibration, the user terminal initially acquires the timing and frequency of the access point based on the beacon pilot transmitted on the BCH. The user terminal then sends a message on the RACH to initiate a calibration procedure with the access point. The calibration may be performed in parallel with registration/authentication.
Since the frequency responses of the transmit/receive chains at the access point and user terminal are typically flat over most of the band of interest, the phase/gain differences of the transmit/receive chains may be characterized with a small number of subbands. The calibration may be performed for 4, 8, 16, 48 or some other number of subbands, which may be specified in the message sent to initiate the calibration. Calibration may also be performed for the pilot subbands. Calibration constants for subbands on which calibration is not explicitly performed may be computed by interpolation on calibrated subbands. For clarity, the following assumes that calibration is performed for all data subbands.
For the calibration, the access point allocates to the user terminal a sufficient amount of time on the RCH to send an uplink MIMO pilot of sufficient duration plus a message. The duration of the uplink MIMO pilot may be dependent on the number of subbands over which calibration is performed. For example, 8 OFDM symbols may be sufficient if calibration is performed for four subbands, and more (e.g., 20) OFDM symbols may be needed for more subbands. The total transmit power is typically fixed, so if the MIMO pilot is transmitted on a small number of subbands, then higher amounts of transmit power may be used for each of these subbands and the SNR for each subband is high. Conversely, if the MIMO pilot is transmitted on a large number of subbands then smaller amounts of transmit power may be used for each subband and the SNR for each subband is worse. If the SNR of each subband is not sufficiently high, then more OFDM symbols may be sent for the MIMO pilot and integrated at the receiver to obtain a higher overall SNR for the subband.
The user terminal then transmits a MIMO pilot on the RCH, which is used by the access point to derive an estimate of the effective uplink channel response, Ĥup(k), for each of the data subbands. The uplink channel response estimates are quantized (e.g., to 12-bit complex values, with inphase (I) and quadrature (Q) components) and sent to the user terminal
The user terminal also derives an estimate of the effective downlink channel response, Ĥdn(k), for each of the data subbands based on the downlink MIMO pilot sent on the BCH. Upon obtaining the effective uplink and downlink channel response estimates, Ĥup(k) and Ĥdn(k), for all data subbands, the user terminal determines correction factors, {circumflex over (K)}ap(k) and {circumflex over (K)}ut(k), for each of the data subbands, which are to be used by the access point and user terminal, respectively. A correction vector {circumflex over (k)}ap(k) may be defined to include only the diagonal elements of {circumflex over (K)}ap(k), and a correction vector {circumflex over (k)}ut (k) may be defined to include only the diagonal elements of {circumflex over (K)}ut(k).
The correction factors may be derived in various manners, including by a matrix-ratio computation and an MMSE computation. Both of these computation methods are described in further detail below. Other computation methods may also be used, and this is within the scope of the invention.
Matrix-Ratio Computation
To determine the correction vectors {circumflex over (k)}ap(k) and {circumflex over (k)}ut(k) given the effective downlink and uplink channel response estimates, Ĥdn(k) and Ĥup(k), an (Nut×Nap) matrix C(k) is first computed for each data subband, as follows:
where the ratio is taken element-by-element.
Each element of C(k) may thus be computed as:
The correction vector {circumflex over (k)}ap(k) for the access point is then equal to the mean of the normalized rows of C(k). Each row of C(k) is first normalized by scaling each of the Nap elements in the row with the first element in the row. Thus, if ci(k)=[ci,1(k) . . . ci,N
{circumflex over (c)}
i(k)=[ci,1(k)/ci,1(k) . . . ci,j(k)/ci,1(k) . . . ci,N
The mean of the normalized rows is then the sum of the Nut normalized rows divided by Nut, which may be expressed as:
Because of the normalization, the first element of {circumflex over (k)}ap(k) is unity.
The correction vector {circumflex over (k)}ut(k) for the user terminal is equal to the mean of the inverses of the normalized columns of C(k). The j-th column of C(k) is first normalized by scaling each element in the column with the j-th element of the vector {circumflex over (k)}ap(k), which is denoted as Kap,j,j(k). Thus, if cj(k)=[c1,j(k) . . . cN
{hacek over (c)}
j(k)=[c1,j(k)/Kap,j,j(k) . . . ci,j(k)/Kap,j,j(k) . . . cN
The mean of the inverses of the normalized columns is then the sum of the inverses of the Nap normalized columns divided by Nap, which may be expressed as:
where the inversion of the normalized columns, {hacek over (c)}j(k), is performed element-wise.
MMSE Computation
For the MMSE computation, the correction factors {circumflex over (K)}ap(k) and {circumflex over (K)}ut(k) are derived from the effective downlink and uplink channel response estimates, Ĥdn(k) and Ĥup(k), such that the mean square error (MSE) between the calibrated downlink channel response and the calibrated uplink channel response is minimized. This condition may be expressed as:
min|({circumflex over (H)}dn(k){circumflex over (K)}ap(k))T−({circumflex over (H)}up(k){circumflex over (K)}ut(k))|2, for kεK, Eq. (34)
which may also be written as:
min|{circumflex over (K)}ap(k)ĤdnT(k)−Ĥup(k){circumflex over (K)}ut(k)|2, for kεK,
where {circumflex over (K)}ap(k)={circumflex over (K)}ap(k) since {circumflex over (K)}ap(k) is a diagonal matrix.
Equation (34) is subject to the constraint that the lead element of {circumflex over (K)}ap(k) is set equal to unity (i.e. {circumflex over (K)}ap,0,0(k)=1). Without this constraint, the trivial solution would be obtained with all elements of the matrices {circumflex over (K)}ap(k) and {right arrow over (K)}ut(k) set equal to zero. In equation (34), a matrix Y(k) is first obtained as Y(k)={circumflex over (K)}ap(k)ĤdnT(k)−Ĥup(k){circumflex over (K)}ut(k). The square of the absolute value is next obtained for each of the Nap·Nut entries of the matrix Y(k). The mean square error (or the square error, since a divide by Nap·Nut is omitted) is then equal to the sum of all Nap·Nut squared values.
The MMSE computation is performed for each designated subband to obtain the correction factors {circumflex over (K)}ap(k) and {circumflex over (K)}ut(k) for that subband. The MMSE computation for one subband is described below. For simplicity, the subband index, k, is omitted in the following description. Also for simplicity, the elements of the downlink channel response estimate ĤdnT are denoted as {aij}, the elements of the uplink channel response estimate Ĥup are denoted as {bij}, the diagonal elements of the matrix {circumflex over (K)}ap are denoted as {ui}, and the diagonal elements of the matrix {circumflex over (K)}ut are denoted as {νj}, where i={1 . . . Nap} and j={1 . . . Nut}.
The mean square error may be rewritten from equation (34), as follows:
again subject to the constraint u1=1. The minimum mean square error may be obtained by taking the partial derivatives of equation (35) with respect to u and v and setting the partial derivatives to zero. The results of these operations are the following equation sets:
In equation (36a), u1=1 so there is no partial derivative for this case, and the index i runs from 2 through Nap.
The set of (Nap+Nut−1) equations in equation sets (36a) and (36b) may be more conveniently expressed in matrix form, as follows:
Ay=z, Eq. (37)
where
The matrix A includes (Nap+Nut−1) rows, with the first Nap−1 rows corresponding to the Nap−1 equations from equation set (36a) and the last Nut rows corresponding to the Nut equations from equation set (36b). In particular, the first row of the matrix A is generated from equation set (36a) with i=2, the second row is generated with i=3, and so on. The Nap-th row of the matrix A is generated from equation set (36b) with j=1, and so on, and the last row is generated with j=Nut. As shown above, the entries of the matrix A and the entries of the vector z may be obtained based on the entries in the matrices ĤdnT and Ĥup.
The correction factors are included in the vector y, which may be obtained as:
y=A
−1
z. Eq. (38)
The results of the MMSE computation are correction matrices {circumflex over (K)}ap and {circumflex over (K)}ut that minimize the mean square error in the calibrated downlink and uplink channel responses, as shown in equation (34). Since the matrices {circumflex over (K)}ap and {circumflex over (K)}ut are obtained based on the downlink and uplink channel response estimates, Ĥdn and Ĥup, the quality of the correction matrices {circumflex over (K)}ap and {circumflex over (K)}ut are are thus dependent on the quality of the channel estimates Ĥdn and Ĥup. The MIMO pilot may be averaged at the receiver to obtain more accurate estimates for Ĥdn and Ĥup.
The correction matrices, {circumflex over (K)}ap and {circumflex over (K)}ut, obtained based on the MMSE computation are generally better than the correction matrices obtained based on the matrix-ratio computation. This is especially true when some of the channel gains are small and measurement noise can greatly degrade the channel gains.
Post Computation
A pair of correction vectors, {circumflex over (k)}ap(k) and {circumflex over (k)}ut(k), may be determined for each of the data subbands. Since adjacent subbands are likely to be correlated, the computation may be simplified. For example, the computation may be performed for every n-th subband instead of each subband, where n may be determined by the expected response of the transmit/receive chains. If the calibration is performed for fewer than all of the data and pilot subbands, then the correction factors for the “uncalibrated” subbands may be obtained by interpolating the correction factors obtained for the “calibrated” subbands.
Various other calibration schemes may also be used to derive the correction vectors, {circumflex over (k)}ap(k) and {circumflex over (k)}ut(k), for the access point and the user terminal, respectively. However, the scheme described above allows “compatible” correction vectors to be derived for the access point when the calibration is performed by different user terminals.
After the derivation, the user terminal sends the correction vectors {circumflex over (k)}ap(k) for all data subbands back to the access point. If the access point has already been calibrated (e.g., by other user terminals), then the current correction vectors are updated with the newly received correction vectors. Thus, if the access point uses correction vectors {umlaut over (k)}ap1(k) to transmit the MIMO pilot from which the user terminal determines new correction vectors {circumflex over (k)}ap2(k), then the updated correction vectors are the product of the current and new correction vectors, i.e., {circumflex over (k)}ap3(k)={circumflex over (k)}ap1(k)·{circumflex over (k)}ap2(k), where the multiplication is element-by-element. The updated correction vectors {circumflex over (k)}ap3(k) may then be used by the access point until they are updated again.
The correction vectors {circumflex over (k)}ap1(k) and {circumflex over (k)}ap2(k) may be derived by the same or different user terminals. In one embodiment, the updated correction vectors are defined as {circumflex over (k)}ap3(k)={circumflex over (k)}ap1(k)·{circumflex over (k)}ap2(k), where the multiplication is element-by-element. In another embodiment, the updated correction vectors may be redefined as {circumflex over (k)}ap3(k)={circumflex over (k)}ap1(k)·{circumflex over (k)}ap2α(k), where α is a factor used to provide weighted averaging (e.g., 0<α<1). If the calibration updates are infrequent, then close to one might perform best. If the calibration updates are frequent but noisy, then a smaller value for is better. The updated correction vectors {circumflex over (k)}ap3(k) may then be used by the access point until they are updated again.
The access point and user terminal use their respective correction vectors {circumflex over (k)}ap(k) and {circumflex over (k)}ut(k), or the corresponding correction matrices {circumflex over (K)}ap(k) and {circumflex over (K)}ut(k), for kεK, to scale the modulation symbols prior to transmission, as described below. The calibrated downlink and uplink channels that the user terminal and access point observe are shown in equation (27).
The spatial processing at the access point and user terminal may be simplified for a TDD system, after calibration has been performed to account for the difference in the transmit/receive chains. As noted above, the calibrated downlink channel response is Hcdn(k)=Hdn(k){circumflex over (K)}ap(k). The calibrated uplink channel response is Hcup(k)=Hup(k){circumflex over (K)}ut(k)≈(Hdn(k){circumflex over (K)}ap(k))T.
Uplink Spatial Processing
Singular value decomposition of the calibrated uplink channel response matrix, Hcup(k), may be expressed as:
H
cup(k)=Uap(k)Σ(k)VutH(k), for kε, Eq. (39)
where
Correspondingly, singular value decomposition of the calibrated downlink channel response matrix, Hcdn(k), may be expressed as:
H
cdn(k)=Vut*(k)Σ(k)UapT(k), for kεK. Eq. (40)
The matrices Vut*(k) and Uap*(k) are also matrices of left and right eigenvectors, respectively, of Hcdn(k). As shown in equations (39) and (40) and based on the above description, the matrices of left and right eigenvectors for one link are the complex conjugate of the matrices of right and left eigenvectors, respectively, for the other link. The matrices Vut(k), Vut*(k), VutT (k), and VutH(k) are different forms of the matrix Vut(k), and the matrices Uap(k), Uap*(k), UapT(k), and UapH(k) are also different forms of the matrix Uap(k). For simplicity, reference to the matrices Uap(k) and Vut(k) in the following description may also refer to their various other forms. The matrices Uap(k) and Vut(k) are used by the access point and user terminal, respectively, for spatial processing and are denoted as such by their subscripts. The eigenvectors are also often referred to as “steering” vectors.
The user terminal can estimate the calibrated downlink channel response based on the MIMO pilot sent by the access point. The user terminal can then perform the singular value decomposition of the calibrated downlink channel response estimate Ĥcdn(k), for kεK, to obtain the diagonal matrix {circumflex over (Σ)}(k) and the matrix Vut*(k) of left eigenvectors of Hcdn(k). This singular value decomposition may be given as Ĥcdn(k)={circumflex over (V)}ut*(k){circumflex over (Σ)}(k)UapT(k), where the hat (“̂”) above each matrix indicates that it is an estimate of the actual matrix.
Similarly, the access point can estimate the calibrated uplink channel response based on a MIMO pilot sent by the user terminal. The access point may then perform singular value decomposition for the calibrated uplink channel response estimate Ĥcup(k), for kεK, to obtain the diagonal matrix {circumflex over (Σ)}(k) and the matrix Ûap(k) of left eigenvectors of Ĥcup(k). This singular value decomposition may be given as Ĥcup(k)Ûap(k){circumflex over (Σ)}(k){circumflex over (V)}utH(k).
An (Nut×Nut) matrix Fut(k) may be defined as:
F
ut(k)={circumflex over (K)}ut(k){circumflex over (V)}ut(k), for kεK. Eq. (41)
While it is active, the user terminal keeps a running estimate of the calibrated downlink channel Ĥcdn(k) and the matrices {circumflex over (V)}ut(k) of left eigenvectors of Ĥcdn(k), which are used to update the matrix Fut(k).
The user terminal uses the matrix Fut(k) for the spatial processing for the beam-steering and spatial multiplexing modes. For the spatial multiplexing mode, the transmit vector xup(k) for each subband may be expressed as:
x
up(k)=Fut(k)sup(k), for kεK, Eq. (42)
where
At the access point, the received vector rup(k) for the uplink transmission may be expressed as:
where
r
up(k) is the received vector for the uplink subband k; and
n
up(k) is additive white Gaussian noise (AWGN) for subband k.
Equation (43) uses the following relationships: Hup(k){circumflex over (K)}ut(k)=Hcup(k)≈Ĥcup(k) and Ĥcup(k)=Ûap(k){circumflex over (Σ)}(k){circumflex over (V)}utH(k). As shown in equation (43), at the access point, the received uplink transmission is transformed by Ûap(k){circumflex over (Σ)}(k), which is the matrix Ûap(k) of left eigenvectors of Ĥcup(k) scaled by the diagonal matrix Σ(k) of singular values.
The user terminal transmits a steered reference on the uplink using the matrix Fut(k). The steered reference is a pilot transmission on one wideband eigenmode using either beam-steering or beam-forming, and is described in detail below. At the access point, the received uplink steered reference (in the absence of noise) is approximately Ûap(k){circumflex over (Σ)}(k). The access point can thus obtain an estimate of the unitary matrix Ûap(k) and diagonal matrix {circumflex over (Σ)}(k) based on the steered reference sent by the user terminal. Various estimation techniques may be used to obtain the estimate of the unitary and diagonal matrices.
In one embodiment, to obtain an estimate of Ûap(k), the received vector rm(k) for the steered reference for subband k of wideband eigenmode m is first multiplied with the complex conjugate of a pilot OFDM symbol, p*(k), sent for the steered reference. The generation of the steered reference and the pilot OFDM symbol are described in detail below. The result is then integrated over multiple received steered reference symbols for each wideband eigenmode to obtain an estimate of ûm(k)σm(k), which is a scaled left eigenvector of Ĥcup(k) for wideband eigenmode m. Since eigenvectors have unit power, the singular values (or σm(k)) in {circumflex over (Σ)}(k) may be estimated based on the received power of the steered reference, which can be measured for each subband of each wideband eigenmode.
In another embodiment, an MMSE technique is used to obtain an estimate of ûm(k) based on the received vector rm(k) for the steered reference.
The steered reference may be sent for one wideband eigenmode in any given symbol period, and may in turn be used to obtain an estimate of one eigenvector for each subband of that wideband eigenmode. Thus, the receiver is able to obtain an estimate of one eigenvector in a unitary matrix for any given symbol period. Since estimates of multiple eigenvectors for the unitary matrix are obtained over different symbol periods, and due to noise and other sources of degradation in the transmission path, the estimated eigenvectors for the unitary matrix are not likely be orthogonal. If the estimated eigenvectors are thereafter used for spatial processing of data transmission on the other link, then any errors in orthogonality in these estimated eigenvectors would result in cross-talk among the eigenmodes, which may degrade performance.
In an embodiment, the estimated eigenvectors for each unitary matrix are forced to be orthogonal to each other. The orthogonalization of the eigenvectors may be achieved using various techniques such as QR factorization, minimum square error computation, polar decomposition, and so on. QR factorization decomposes a matrix MT (with non-orthogonal columns) into an orthogonal matrix QF and an upper triangle matrix RF. The matrix QF forms an orthogonal basis for the columns of MT. The diagonal elements of RF give the length of the components of the columns of MT in the directions of the respective columns of QF. The matrix QF may be used for spatial processing on the downlink. The matrices QF and RF may be used to derive an enhanced matched filter matrix for the uplink. The QR factorization may be performed by various methods including a Gram-Schmidt procedure, a householder transformation, and so on.
Other techniques to estimate the unitary and diagonal matrices based on the steered reference may also be used, and this is within the scope of the invention.
The access point can thus estimate both Ûap(k) and {circumflex over (Σ)}(k) based on the steered reference sent by the user terminal, without having to perform singular value decomposition on Ĥcup(k).
A normalized matched filter matrix Map(k) for the uplink transmission from the user terminal may be expressed as:
M
ap(k)={circumflex over (Σ)}−1(k){circumflex over (U)}apH(k), for kεK. Eq. (44)
The matched filtering at the access point for the uplink transmission may then be expressed as:
Downlink Spatial Processing
For the downlink, the access point uses an (Nap×Nap) matrix Fap(k) for spatial processing. This matrix may be expressed as:
F
ap(k)={circumflex over (K)}ap(k){circumflex over (U)}ap(k), for kεK. Eq. (46)
The correction matrix {circumflex over (K)}ap(k) is derived by the user terminal and sent back to the access point during calibration. The matrix Ûap(k) may be obtained based on the steered reference sent on the uplink by the user terminal.
For the spatial multiplexing mode, the transmit vector xdn(k) for the downlink for each data subband may be expressed as:
x
dn(k)=Fap(k)sdn(k), for kεK, Eq. (47)
At the user terminal, the received vector rdn(k) for the downlink transmission may be expressed as:
As shown in equation (48), at the user terminal, the received downlink transmission is transformed by {circumflex over (V)}ut*(k){circumflex over (Σ)}(k), which is the matrix {circumflex over (V)}ut(k) of left eigenvectors of Ĥcdn(k) scaled by the diagonal matrix {circumflex over (Σ)}(k) of singular values.
A normalized matched filter matrix Mut(k) for the downlink transmission from the access point may be expressed as:
M
ut(k)={circumflex over (Σ)}−1(k){circumflex over (V)}utT(k), for kεK. Eq. (49)
The diagonal matrix {circumflex over (Σ)}(k) and matrix {circumflex over (V)}ut(k) of left eigenvectors can be derived by the user terminal by performing singular value decomposition on the calibrated downlink channel response estimate Ĥcdn(k), as described above.
The matched filtering at the user terminal for the downlink transmission may then be expressed as:
Access Point and User Terminal Spatial Processing
Because of the reciprocal channel for the TDD system and the calibration, the spatial processing at both the access point and the user terminal may be simplified. Table 32 summarizes the spatial processing at the access point and user terminal for data transmission and reception.
x
up (k) = {circumflex over (K)}ut (k){circumflex over (V)}ut (k)sup (k)
ŝ
dn (k) = {circumflex over (Σ)}−1 (k){circumflex over (V)}utT (k)rdn (k)
ŝ
up (k) = {circumflex over (Σ)}−1 (k)ÛapH (k)rup (k)
x
dn (k) = {circumflex over (K)}ap (k)Ûap* (k)sdn (k)
The spatial processing for data reception is also commonly referred to as matched filtering.
Because of the reciprocal channel, {circumflex over (V)}ut(k) is both the matrix of right eigenvectors of Ĥcup(k) (to transmit) and left eigenvectors of Ĥcdn(k) (to receive) for the user terminal. Similarly, Ûap(k) is both the matrix of right eigenvectors of Ĥcdn(k) (to transmit) and left eigenvectors of Ûap(k) (to receive) for the access point. The singular value decomposition only needs to be performed by the user terminal for the calibrated downlink channel response estimate Ĥcdn(k) to obtain {circumflex over (V)}ut*(k) and {circumflex over (Σ)}(k). The access point can derive Ûap(k) and {circumflex over (Σ)}(k) based on the steered reference sent by the user terminal and does not need to perform the singular value decomposition on the uplink channel response estimate Ĥcup(k). The access point and user terminal may have different versions of the matrix {circumflex over (Σ)}(k) due to the different means used by the access point and user terminal to derive {circumflex over (Σ)}(k). Moreover, the matrix Ûap(k) derived by the access point based on the steered reference is typically different from the matrix Ûap(k) derived by the user terminal using singular value decomposition. For simplicity, these differences are not shown in the above derivation.
Beam-Steering
For certain channel conditions, it is better to transmit data on only one wideband eigenmode—typically the best or principal wideband eigenmode. This may be the case if the received SNRs for all other wideband eigenmodes are sufficiently poor so that improved performance is achieved by using all of the available transmit power on the principal wideband eigenmode.
Data transmission on one wideband eigenmode may be achieved using either beam-forming or beam-steering. For beam-forming, the modulation symbols are spatially processed with the eigenvectors {circumflex over (v)}ut,1(k) or ûap,1(k), for kεK, for the principal wideband eigenmode (i.e., the first column of {circumflex over (V)}ut(k) or Ûap(k), after the ordering). For beam-steering, the modulation symbols are spatially processed with a set of “normalized” (or “saturated”) eigenvectors {tilde over (v)}ut(k) or ũap(k), for kεK, for the principal wideband eigenmode. For clarity, beam-steering is described below for the uplink.
For the uplink, the elements of each eigenvector {circumflex over (v)}ut,1(k), for kεK, for the principal wideband eigenmode may have different magnitudes. Thus, the preconditioned symbols for each subband, which are obtained by multiplying the modulation symbol for subband k with the elements of the eigenvector {circumflex over (v)}ut,1(k) for subband k, may then have different magnitudes. Consequently, the per-antenna transmit vectors, each of which includes the preconditioned symbols for all data subbands for a given transmit antenna, may have different magnitudes. If the transmit power for each transmit antenna is limited (e.g., because of limitations of power amplifiers), then beam-forming may not fully use the total power available for each antenna.
Beam-steering uses only the phase information from the eigenvectors {circumflex over (v)}ut,1(k), for kεK, for the principal wideband eigenmode and normalizes each eigenvector such that all elements in the eigenvector have equal magnitudes. The normalized eigenvector {tilde over (v)}ut(k) for subband k may be expressed as:
where A is a constant (e.g., A=1); and
θi(k) is the phase for subband k of transmit antenna i, which is given as:
As shown in equation (52), the phase of each element in the vector {tilde over (v)}ut(k) is obtained from the corresponding element of the eigenvector {circumflex over (v)}ut,1(k) (i.e., θi(k) is obtained from {circumflex over (ν)}ut,1,i(k) where {circumflex over (v)}ut,1(k)=[{circumflex over (ν)}ut,1,1(k) {circumflex over (ν)}ut,1,N
Uplink Beam-Steering
The spatial processing by the user terminal for beam-steering on the uplink may be expressed as:
{tilde over (x)}
up(k)={circumflex over (K)}ut{tilde over (v)}ut(k)sup(k), for kεK, Eq. (53)
where
sup(k) is the modulation symbol to be transmitted on subband k; and
{tilde over (x)}
up(k) is the transmit vector for subband k for beam-steering.
As shown in equation (53), the Nut elements of the normalized steering vector {tilde over (v)}ut(k) for each subband have equal magnitude but possibly different phases. The beam-steering thus generates one transmit vector {tilde over (x)}up(k) for each subband, with the Nut elements of {tilde over (x)}up(k) having the same magnitude but possibly different phases.
The received uplink transmission at the access point for beam-steering may be expressed as:
where {tilde over (r)}up(k) is the received vector for the uplink for subband k for beam-steering.
A matched filter row vector {tilde over (m)}ap(k) for the uplink transmission using beam-steering may be expressed as:
{tilde over (m)}
ap(k)=(Hcup(k) {tilde over (v)}ut(k))H, for kεK. Eq. (55)
The matched filter vector {tilde over (m)}ap(k) may be obtained as described below. The spatial processing (or matched filtering) at the access point for the received uplink transmission with beam-steering may be expressed as:
where
Downlink Beam-Steering
The spatial processing by the access point for beam-steering on the downlink may be expressed as:
{tilde over (x)}
dn(k)={circumflex over (K)}ap{tilde over (u)}ap(k)sdn(k), for kεK, Eq. (57)
A matched filter row vector {tilde over (m)}ut(k) for the downlink transmission using beam-steering may be expressed as:
{tilde over (m)}
ut(k)=(Hcdn(k) {tilde over (u)}ap(k))H, for kεK. Eq. (58)
The spatial processing (or matched filtering) at the user terminal for the received downlink transmission may be expressed as:
Spatial Processing with Channel Inversion
For the uplink, the transmit vector xup(k) for the spatial multiplexing mode may be derived by the user terminal as:
x
up(k)={circumflex over (K)}ut(k) {circumflex over (V)}ut(k)G(k)sup(k), for kεK, Eq. (60)
where G(k) is a diagonal matrix of gains for the channel inversion described above.
Equation (60) is similar to equation (15), except that {circumflex over (K)}ut(k){circumflex over (V)}ut(k) is used in place of V(k). The elements of {circumflex over (K)}ut(k){circumflex over (V)}ut(k) are provided to multipliers 952 within beam-formers 950 in
For the uplink, the transmit vector {tilde over (x)}up(k) for the beam-steering mode may be derived by the user terminal as:
{tilde over (x)}
up(k)={circumflex over (K)}ut(k) {tilde over (v)}ut(k){tilde over (g)}(k)sup(k), for kεK, Eq. (61)
where {tilde over (v)}ut(k) is a vector with four elements having equal magnitude but phases obtained based on eigenvector {circumflex over (v)}ut,1(k) for the principal eigenmode.
The vector {circumflex over (v)}ut(k) may be derived similar to that shown above in equations (16) and (17). The gain {tilde over (g)}(k) achieves channel inversion and may be derived similar to that shown above in equations (18) through (20), except that {tilde over (λ)}1(k)={tilde over (v)}utH(k) ĤcupH(k) Ĥcup(k) {tilde over (v)}ut(k) is used for equation (20). The elements of {tilde over (v)}ut(k) are provided to multipliers 1052 within beam-steering unit 1050 in
For the downlink, the transmit vector xdn(k) for the spatial multiplexing mode may be derived by the access point as:
x
dn(k)={circumflex over (K)}ap(k){circumflex over (U)}ap(k)G(k)sdn(k), for kεK. Eq. (62)
Equation (62) is similar to equation (15), except that {circumflex over (K)}ap(k) Ûap(k) is used in place of V(k). The elements of {circumflex over (K)}ap(k) {circumflex over (K)}ap(k) may be provided to multipliers 952 within beam-formers 950 in
For the downlink, the transmit vector {tilde over (x)}dn(k) for the beam-steering mode may be derived by the access point as:
{tilde over (x)}
n(k)={circumflex over (K)}ap(k){tilde over (u)}ap(k){tilde over (g)}(k)sdn(k), for kεK, Eq. (63)
where ũap(k) is a vector with four elements having equal magnitude but phases obtained based on eigenvector ûap,1(k) for the principal eigenmode.
The gain {tilde over (g)}(k) achieves channel inversion and may be derived in a similar manner to that shown above in equations (18) through (20), except that {tilde over (λ)}1(k)=ũapH(k)ĤcdnH(k) Ĥcdn(k) ũap(k) is used for equation (20). The elements of ũap(k) are provided to multipliers 1052 within beam-steering unit 1050 in
A pilot structure is provided for the MIMO WLAN system to allow the access points and user terminals to perform timing and frequency acquisition, channel estimation, and other functions needed for proper system operation. Table 33 lists four types of pilot and their short description for an exemplary pilot structure. Fewer, different, and/or additional pilot types may also be used for the pilot structure.
In an embodiment, the pilot structure includes (1) for the downlink—a beacon pilot, a MIMO pilot, a steered reference, and a carrier pilot transmitted by the access point, and (2) for the uplink—a MIMO pilot, a steered reference, and a carrier pilot transmitted by the user terminals.
The downlink beacon pilot and MIMO pilot are sent on the BCH (as shown in
An uplink steered reference is transmitted by each active user terminal that supports the beam-steering or spatial multiplexing mode and may be used by the access point to (1) derive the steering vectors for the downlink transmission and (2) derive a matched filter for the uplink transmission. In general, the steered reference is only sent for/by user terminals that support the beam-steering and/or spatial multiplexing modes. The reference sent works regardless of whether or not it is steered properly (e.g., due to a poor channel estimate). That is, the reference becomes orthogonal on a per transmit antenna basis since the steering matrix is diagonal.
If a user terminal is calibrated, then it can transmit a steered reference on the principal eigenmode on the RACH using the vector {circumflex over (K)}utk(k){circumflex over (v)}ut,0(k), for kεK, where {circumflex over (v)}ut,0(k) is the column of {circumflex over (V)}ut(k) for the principal eigenmode. If the user terminal is not calibrated, then it can transmit a pilot on the RACH using a vector
for kεK. The vector vut,p(k) for each subband includes Nut random steering coefficients whose phases θi(k), for iε{1, 2, . . . Nut} may be selected in accordance with a pseudo-random procedure. Since only the relative phases among the Nut steering coefficients matter, the phase of the first steering coefficient may be set to zero (i.e., θ1(k)=0). The phases of the other Nut−1 steering coefficients may change for each access attempt, so that all 360° are covered by each steering coefficient in intervals of 360°/Nθ
The access point does not have knowledge of the channel for any user terminal until the user terminal communicates directly with the access point. When a user terminal desires to transmit data, it first estimates the channel based on the MIMO pilot transmitted by the access point. The user terminal then sends steered reference in the preamble of the RACH when it attempts to access the system. The access point uses the reference on the RACH preamble for signal detection and channel estimation.
Once the user terminal has been granted access to the system and assigned FCH/RCH resources by the access point, the user terminal sends a reference (e.g., a MIMO pilot) at the beginning of each RCH PDU it transmits. If the user terminal is using the diversity mode, then the reference is sent on the RCH without steering. If the user terminal is using the beam-steering or spatial multiplexing mode, then a steered reference is sent on the RCH to allow the access point to determine the eigenvector for the principal eigenmode (for the beam-steering mode) or the set of four eigenvectors (for the spatial multiplexing mode) for each of the 48 data subbands. The steered reference allows the access point to improve its estimate of the channel and to track the channel.
Beacon Pilot—Downlink
The downlink beacon pilot is included in the first portion of the BCH (as shown in
In a specific embodiment, the B OFDM symbol comprises a set of 12 BPSK modulation symbols, b(k), for 12 specific subbands, which is shown in Table 34.
For the beacon pilot embodiment shown in Table 34, the B OFDM symbol comprises (1) BPSK modulation symbol (1+j) for subbands −24, −16, −4, 12, 16, 20, and 24, (2) BPSK modulation symbol −(1+j) for subbands −20, −12, −8, 4, and 8, and (3) signal values of zero for the remaining 52 subbands. The B OFDM symbol is specifically designed to facilitate timing and frequency acquisition by the user terminals. However, other OFDM symbols may also be used for the beacon pilot, and this is within the scope of the invention.
MIMO Pilot—Downlink
The downlink MIMO pilot is included in the second portion of the BCH (as shown in
In a specific embodiment, the P OFDM symbol comprises a set of 52 QPSK modulation symbols, p(k), for the 48 data subbands and 4 pilot subbands, which is shown in Table 34. Signal values of zero are transmitted on the remaining 12 subbands. The P OFDM symbol comprises a unique “word” of 52 QPSK modulation symbols that is designed to facilitate channel estimation by the user terminals. This unique word is also selected to minimize the peak-to-average variation in the transmitted MIMO pilot. This may then reduce the amount of distortion and non-linearity generated by the receiver circuitry at the user terminals, which can result in improved accuracy for the channel estimation. However, other OFDM symbols may also be used for the MIMO pilot, and this is within the scope of the invention.
In an embodiment, the four antennas at the access point are assigned 4-chip Walsh sequences of W1=1111, W2=1010, W3=1100, and W4=1001 for the MIMO pilot. For a given Walsh sequence, a value of “1” indicates that a P OFDM symbol is transmitted and a value of “0” indicates that a −P OFDM symbol is transmitted (i.e., each of the 52 modulation symbols in P is inverted).
Table 35 lists the OFDM symbols to be transmitted from each of the four antennas at the access point for the beacon pilot and MIMO pilot. The B and P OFDM symbols are as described above.
The MIMO pilot may be used by the user terminal to estimate the channel response of the downlink. In particular, to recover the pilot sent from access point antenna i and received by user terminal antenna j, the pilot received by terminal antenna j is first multiplied with the Walsh sequence assigned to access point antenna i. The eight decovered OFDM symbols for all eight symbol periods for the MIMO pilot are then accumulated, where the accumulation is performed individually for each of the 52 subbands used to carry the MIMO pilot. The results of the accumulation is ĥcdni,j(k), for k=±{1, . . . , 26}, which is an estimate of the calibrated downlink channel response from access point antenna i to user terminal antenna j for the 52 data and pilot subbands.
The same pilot processing may be performed by the access point to recover the pilot from each access point antenna at each user terminal antenna. The pilot transmitted from each access point antenna may be recovered by decovering with the Walsh sequence assigned to that antenna. The pilot processing provides Nap·Nut values for each of the 52 subbands, where Nap denotes the number of antennas at the access point and Nut denotes the number of antennas at the user terminal. The Nap·Nut values for each subband are the elements of the calibrated downlink channel response estimate {circumflex over (H)}cdn(k) for that subband.
The MIMO pilot may also be transmitted on the uplink by the user terminal for calibration and in the diversity mode. The same processing described above for the user terminal to recover the MIMO pilot sent by the access point may also be performed by the access point to recover the MIMO pilot sent by the user terminal
Steered Reference
A steered reference may be transmitted in the preamble portion of an RACH PDU (as shown in
Steered Reference for Spatial Multiplexing
The steered reference comprises a specific OFDM symbol (e.g., the same P OFDM symbol used for the MIMO pilot) that is transmitted from all of the transmit antennas at the user terminal (for the uplink) or the access point (for the downlink). However, the P OFDM symbol for each symbol period is spatially processed (i.e., beam-formed) with a steering vector for one eigenmode.
The first symbol of steered reference transmitted by the user terminal in the preamble of the RACH may be expressed as:
x
(k)={circumflex over (K)}ut(k)·{circumflex over (v)}ut,1(k)·p(k), for kεK′, Eq. (64)
where
The second symbol of steered reference transmitted by the user terminal in the preamble of the RACH includes the data rate indicator (DRI) for the RACH PDU. The DRI indicates the rate used for the RACH message sent in the RACH PDU. The DRI is embedded in the second steered reference symbol by mapping the DRI to a specific QPSK symbol sdri, as shown in Table 15. The sdri symbol is then multiplied with the pilot symbol p(k) before performing the spatial processing. The second symbol of steered reference for the RACH may be expressed as:
x
(k)={circumflex over (K)}ut(k)·{circumflex over (v)}ut,1(k)·sdri·p(k), for kεK′. Eq. (65)
As shown in equations (64) and (65), only eigenvector {circumflex over (v)}ut,1(k) for the principal eigenmode is used for the steered reference for the RACH.
A symbol of steered reference transmitted by the user terminal in the preamble of the RCH may be expressed as:
xup,sr,m(k)={circumflex over (K)}ut(k)·{circumflex over (v)}ut,m(k)·p(k), for kεK′, Eq. (66)
where
A symbol of steered reference transmitted by the access point in the preamble of the FCH may be expressed as:
x
dn,sr,m(k)={circumflex over (K)}ap(k)·{circumflex over (u)}ap,m(k)p(k)m, for kεK′, Eq. (67)
where
x
ch,sr,m(k) is the transmit vector for subband k of wideband eigenmode m;
The steered reference may be transmitted in various manners. In one embodiment, one or more eigenvectors are used for the steered reference for each TDD frame and are dependent on the duration of the steered reference, which is indicated by the FCH/RCH Preamble Type fields in the FCCH information element. Table 36 lists the eigenmodes used for the preamble for the FCH and RCH for various preamble sizes, for an exemplary design.
As shown in Table 36, the steered reference is transmitted for all four eigenmodes within a single TDD frame when the preamble size is four or eight OFDM symbols. The steered reference transmitted by the user terminal for the n-th OFDM symbol in the preamble for the RCH may be expressed as:
x
up,sr,n(k)={circumflex over (K)}ut(k)·{circumflex over (v)}ut,n mod 4(k)·p(k), for kεK′ and n={1, . . . , L}, Eq. (68)
where L is the preamble size, i.e., L=4 for Type 2 and L=8 for Type 3.
Similarly, the steered reference transmitted by the access point for the n-th OFDM symbol in the preamble for the FCH may be expressed as:
x
dn,sr,n(k)={circumflex over (K)}ap(k)·{circumflex over (u)}ap,n mod 4(k)p(k), for kεK′ and n={1, . . . , L}, Eq. (69)
As shown in equations (68) and (69), the four eigenmodes are cycled through in each 4-symbol period by the (n mod 4) operation for the steering vector. This scheme may be used if the channel changes more rapidly and/or during the early part of a connection when a good channel estimate needs to be obtained quickly for proper system operation.
In another embodiment, the steered reference is transmitted for one wideband eigenmode for each TDD frame. The steered reference for four wideband eigenmodes may be cycled through in four TDD frames. For example, the steering vectors {circumflex over (v)}ut,1(k), {circumflex over (v)}ut,2(k), {circumflex over (v)}ut,3(k), and {circumflex over (v)}ut,4(k) may be used for the first, second, third, and fourth TDD frames, respectively, by the user terminal. The particular steering vector to use may be specified by the 2 LSBs of the Frame Counter value in the BCH message. This scheme allows a shorter preamble portion to be used in the PDU but may require a longer time period to obtain a good estimate of the channel.
For both embodiments described above, the steered reference may be transmitted on all four eigenmodes that may be used for data transmission, even though fewer than four eigenmodes are currently used (e.g., because the unused eigenmodes are poor and discarded by the water-filling). The transmission of the steered reference on a currently unused eigenmode allows the receiver to determine when the eigenmode improves enough to be selected for use.
Steered Reference for Beam-Steering
For the beam-steering mode, the spatial processing on the transmit side is performed using a set of normalized eigenvectors for the principal wideband eigenmode. The overall transfer function with a normalized eigenvector is different from the overall transfer function with an unnormalized eigenvector (i.e., Hcup(k) {circumflex over (v)}ut,1(k)≠Hcup(k) {tilde over (v)}ut(k)). A steered reference generated using the set of normalized eigenvectors for all subbands may then be sent by the transmitter and used by the receiver to derive the matched filter vectors for these subbands for the beam-steering mode.
For the uplink, the steered reference for the beam-steering mode may be expressed as:
{tilde over (x)}
up,sr(k)={circumflex over (K)}ut(k){tilde over (v)}ut(k)p(k), for kεK. Eq. (70)
At the access point, the receive uplink steered reference for the beam-steering mode may be expressed as:
To obtain the matched filter row vector {tilde over (m)}ap(k) for the uplink transmission with beam-steering, the received vector {tilde over (r)}up,sr(k) for the steered reference is first multiplied with p*(k). The result is then integrated over multiple received steered reference symbols to form an estimate of Hcup(k) {tilde over (v)}ut(k). The vector {tilde over (m)}ap(k) is then the conjugate transpose of this estimate.
While operating in the beam-steering mode, the user terminal may transmit multiple symbols of steered reference, for example, one or more symbols using the normalized eigenvector {tilde over (v)}ut(k), one or more symbols using the eigenvector {circumflex over (v)}ut,1(k) for the principal wideband eigenmode, and possibly one or more symbols using the eigenvectors for the other wideband eigenmodes. The steered reference symbols generated with {tilde over (v)}ut(k) may be used by the access point to derive the matched filter vector {tilde over (m)}ap(k). The (k). The steered reference symbols generated with {circumflex over (v)}ut,1(k) may be used to obtain ûap,1(k), which may then be used to derive the normalized eigenvector ũap(k) that is used for beam-steering on the downlink. The steered reference symbols generated with the eigenvectors {circumflex over (v)}ut,2(k) through {circumflex over (v)}ut,N
For the downlink the user terminal may derive the matched filter vector {tilde over (m)}ut(k) for the beam-steering mode based on the calibrated downlink channel response estimate Ĥcdn(k). In particular, the user terminal has from the singular value decomposition of Ĥcdn(k) and can derive the normalized eigenvector ûap(k). The user terminal can then multiply ũap(k) with Ĥcdn(k) to obtain Ĥcdn(k)ũap(k), and may then derive {tilde over (m)}ut(k) based on {umlaut over (H)}cdn(k)ûap(k). Alternatively, a steered reference may be sent by the access point using the normalized eigenvector ûap(k), and this steered reference may be processed by the user terminal in the manner described above to obtain {tilde over (m)}ut(k).
Carrier Pilot—Uplink
The OFDM subband structure described herein includes four pilot subbands with indices of −21, −7, 7, and 21. In an embodiment, a carrier pilot is transmitted on the four pilot subbands in all OFDM symbols that are not part of a preamble. The carrier pilot may be used by the receiver to track phase changes due to drifts in the oscillators at both the transmitter and receiver. This may provide improved data demodulation performance.
The carrier pilot comprises four pilot sequences, Pc1(n), Pc2(n), Pc3(n), and Pc4(n) which are transmitted on the four pilot subbands. The pilot sequences may be defined as:
P
c1(n)=Pc2(n)=Pc3(n)=−Pc4(n), for n={1, 2, . . . 127}, Eq. (72)
where n is an index for OFDM symbol period.
The pilot sequences may be defined based on various data sequences. In an embodiment, the pilot sequence Pc1(n) is generated based on a polynomial G(x)=x7+x4+x, d where the initial state is set to all ones and the output bits are mapped to signal values as follows: 1−1 and 0
1. The pilot sequence Pc1(n), for n={1, 2, . . . 127}, may then be expressed as:
pc1(n)={1,1,1,1,−1,−1,−1,1,−1,−1,−1,−1,1,1,−1,1,−1,−1,1,1,−1,1,1,−1,1,1,1,1,1,1,−1,1, 1,1,−1,1,1,−1,−1,1,1,1,−1,1,−1,−1,−1,1,−1,1,−1,−1,1,−1,−1,1,1,1,1,1,−1,−1,1,1, −1,−1,1,−1,1,−1,1,1,−1,−1,−1,1,1,−1,−1,−1,−1,1,−1,−1,1,−1,1,1,1,1,−1,1,−1,1,−1,1, −1,−1,−1,−1,−1,1,−1,1,1,−1,1,−1,1,1,1,−1,−1,1,−1,−1,−1,1,1,1,−1,−1,−1,−1,−1,−1,−1}.
The values of “1” and “−1” in the pilot sequence Pc1(n) may be mapped to pilot symbols using a particular modulation scheme. For example, using BPSK, a “1” may be mapped to 1+j and a “−1” may be mapped to (1+j). If there are more than 127 OFDM symbols, then the pilot sequence may be repeated so that Pc1(n)=Pc1(n mod 127) for n>127.
In one embodiment, the four pilot sequences are reset for each transport channel. Thus, on the downlink, the pilot sequences are reset for the first OFDM symbol of the BCH message, reset again for the first OFDM symbol of the FCCH message, and reset for the first OFDM symbol sent on the FCH. In another embodiment, the pilot sequences are reset at the start of each TDD frame and repeat as often as needed. For this embodiment, the pilot sequences may be stalled during the preamble portions of the BCH and FCH.
In the diversity mode, the four pilot sequences are mapped to four subband/antenna pairing as shown in Table 29. In particular, Pc1(n) is used for subband −21 of antenna 1, Pc2(n) is used for subband −7 of antenna 2, Pc3(n) is used for subband 7 of antenna 3, and Pc4(n) is used for subband 21 of antenna 4. Each pilot sequence is then transmitted on the associated subband and antenna.
In the spatial multiplexing mode, the four pilot sequences are transmitted on the principal eigenmode of their respective subbands. The spatial processing for the carrier pilot symbols is similar to that performed for the modulation symbols, as described above. In the beam-steering mode, the four pilot sequences are transmitted on their respective subbands using beam-steering. The beam-steering for the carrier pilot symbols is also similar to that performed for the modulation symbols.
A specific pilot structure has been described above for the MIMO WLAN system. Other pilot structures may also be used for the system, and this is within the scope of the invention.
In the Init state, the user terminal acquires the system frequency and timing and obtains system parameters sent on the BCH. In the Init state, the user terminal may perform the following functions:
In the Dormant state, the user terminal periodically monitors the BCH for updated system parameters, indications of pages and broadcast messages being sent on the downlink, and so on. No radio resources are allocated to the user terminal in this state. In the Dormant state, the user terminal may perform the following functions:
In the Access state, the user terminal is in the process of accessing the system. The user terminal may send short messages and/or requests for FCH/RCH resources using the RACH. The operation on the RACH is described in further detail below. If the user terminal is released by the access point, then it transitions back to the Dormant state. If the user terminal is assigned resources for the downlink and/or uplink, then it transitions to the Connected state.
In the Connected state, the user terminal is assigned the FCH/RCH resources, although not necessarily for every TDD frame. The user terminal may actively use the allocated resources or may be idle (while still maintaining the connection) in the Connected state. The user terminal remains in the Connected state until it is released by the access point or if it times out after no activity for a particular timeout period, in which case it transitions back to the Dormant state.
While in the Dormant, Access, or Connected state, the user terminal transitions back to the Init state if it is powered down or if the connection is dropped.
In the Setup substate, the user terminal is in the process of setting up the connection and is not yet exchanging data. The connection setup may include channel estimation for the access point, rate determination, service negotiation, and so on. Upon entering the Setup substate, the user terminal sets a timer for a specified amount of time. If the timer expires before the user terminal leaves this substate, then it transitions back to the Dormant state. The user terminal transitions to the Open substate upon completion of the connection setup.
In the Open substate, the user terminal and access point exchange data on the downlink and/or uplink. While in the Open substate, the user terminal monitors the BCH for system parameters and indication of page/broadcast messages. If a BCH message cannot be decoded correctly within a specified number of TDD frames, then the user terminal transitions back to the Init state.
The user terminal also monitors the FCCH for channel assignment, rate control, RCH timing control, and power control information. The user terminal estimates the received SNR using the BCH beacon pilot and the FCH preamble and determines the maximum rate that can be sustained reliably on the FCH.
The FCH and RCH assignments for the user terminal for each TDD frame are given by the information elements in the FCCH PDU transmitted in the current (or possibly prior) TDD frame. The user terminal may not be assigned for data transmission on the FCH and/or RCH for any given TDD frame. For each TDD frame in which the user terminal is not scheduled for data transmission, it does not receive an FCH PDU on the downlink and does not transmit on the uplink.
For each TDD frame in which the user terminal is scheduled, the data transmissions on the downlink and/or uplink are performed using the rate, transmission mode, and RCH timing offset (for the uplink) indicated in the FCCH assignments (i.e., the FCCH information elements addressed to the user terminal). The user terminal receives, demodulates, and decodes FCH PDUs sent to it. The user terminal also transmits RCH PDUs, which include the preamble and FCH data rate indicator. The user terminal adjusts the rate used on the RCH according to the rate control information contained in the FCCH assignment. If power control is being applied for the uplink transmission, then the user terminal adjusts its transmit power based on the power control commands included in the FCCH assignment. The data exchange may be bursty, in which case the user terminal may enter into the Idle substate whenever no data is being exchanged. The user terminal enter the Idle substate when directed by the access point. If the access point does not assign the FCH or RCH to the user terminal within a specified number of TDD frames, then the user terminal transitions back to the Dormant state and retains its MAC ID.
In the Idle substate, both the uplink and downlink are idling. Data is not being sent in either direction. However, the links are maintained with the steered reference and control messages. In this substate, the access point periodically assigns Idle PDUs to the user terminal on the RCH and possibly the FCH (not necessarily simultaneously). The user terminal may be able to remain in the Connected state indefinitely, provided that the access point periodically assigns Idle PDUs on the FCH and RCH to maintain the link.
While in the Idle substate, the user terminal monitors the BCH. If a BCH message is not decoded correctly within a specified number of TDD frames, then the user terminal transitions back to the Init state. The user terminal also monitors the FCCH for channel assignment, rate control, RCH timing control, and power control information. The user terminal may also estimate the receive SNR and determine the maximum rate supported by the FCH. The user terminal transmits an Idle PDU on the RCH, when assigned, and sets the RCH Request bit in the Idle PDU if it has data to send. If the access point does not assign an FCH or RCH to the user terminal within a specified number of TDD frames, then the user terminal transitions back to the Dormant state and retains its MAC ID.
A time-out timer may be set to a particular value upon entering any of the three substates. This timer would then count down if there is no activity while in the substate. While in the Setup, Active, or Idle substate, the terminal would transition back to the Dormant state if the time-out timer expires and to the Init state if the connection is dropped. While in the Active or Idle substate, the terminal would also transition back to the Dormant state if the connection is released.
In an embodiment, a random access scheme is employed to allow the user terminals to access the MIMO WLAN system. In an embodiment, the random access scheme is based on a slotted Aloha scheme whereby a user terminal transmits in a randomly selected RACH slot to attempt to gain access to the system. The user terminal may send multiple transmissions on the RACH until access is gained or the maximum number of access attempts has been reached. Various parameters for each RACH transmission may be changed to improve the likelihood of success, as described below.
When a user terminal desires to access the system, it first processes the BCH to obtain pertinent system parameters. The user terminal then sends a RACH PDU on the RACH. This RACH PDU includes a RACH message that contains information needed by the access point to process the access request from the user terminal. For example, the RACH message includes the user terminal's assigned MAC ID that allows the access point to identify the user terminal A registration MAC ID (i.e., a specific MAC ID value) may be reserved for unregistered user terminals. In this case, the user terminal's long ID may be included in the Payload field of the RACH message along with the registration MAC ID.
As described above, the RACH PDU may be transmitted at one of four data rates, which are listed in Table 15. The selected rate is embedded in the preamble of the RACH PDU (as shown in
To transmit the RACH PDU, the user terminal first determines the number of RACH slots that may be used for transmission (i.e., the number of “usable” RACH slots). This determination is made based on (1) the number of RACH slots available in the current TDD frame, (2) the duration of each RACH slot, (3) the guard interval, and (4) the length of the RACH PDU to be transmitted. The RACH PDU cannot extend beyond the end of the RACH segment of the TDD frame. Thus, if the RACH PDU is longer than one RACH slot plus the guard interval, then this PDU may not be transmitted in one or more later available RACH slots. The number of RACH slots that may be used to transmit the RACH PDU may be fewer than the number of available RACH slots, based on the factors enumerated above. The RACH segment includes a guard interval, which is provided to prevent the uplink transmission from the user terminals from interfering with the next BCH segment, which is a possibility for user terminals that do not compensate for their round trip delay.
The user terminal then randomly selects one of the usable RACH slots to transmit the RACH PDU. The user terminal then transmits the RACH PDU starting in the selected RACH slot. If the user terminal knows the round trip delay to the access point, then it can account for this delay by adjusting its timing accordingly.
When the access point receives a RACH PDU, it checks the received RACH message using the CRC included in the message. The access point discards the RACH message if the CRC fails. If the CRC passes, the access point sets the RACH Acknowledgment bit on the BCH in the subsequent TDD frame and transmits an RACH acknowledgement on the FCCH within 2 TDD frames. There may be a delay between the setting of the Acknowledgment bit on the BCH and the sending of the acknowledgment on the FCCH, which may be used to account for scheduling delay and so on. For example, if the access point receives the message on the RACH, it can set the Acknowledgment bit on the BCH and have a delay response on the FCCH. The Acknowledgment bit prevents user terminals from retrying and allows unsuccessful user terminals to retry quickly, except during busy RACH periods.
If the user terminal is performing a registration, then it uses the registration MAC ID (e.g., 0x0001). The access point responds by sending a MAC ID Assignment Message on the FCH. All other RACH transmission types include the user terminal's MAC ID assigned by the system. The access point explicitly acknowledges all correctly received RACH messages by sending acknowledgments on the FCCH using the MAC ID assigned to the user terminal
After the user terminal sends the RACH PDU, it monitors the BCH and FCCH to determine whether or not its RACH PDU has been received and processed by the access point. The user terminal monitors the BCH to determine whether or not the RACH Acknowledgment Bit in the BCH message is set. If this bit is set, which indicates that an acknowledgment for this and/or some other user terminals is being sent on the FCCH, then the user terminal further processes the FCCH to obtain IE Type 3 information elements containing acknowledgements. Otherwise, if the RACH Acknowledgment Bit is not set, then the user terminal continues to monitor the BCH or resumes its access procedure on the RACH.
The FCCH IE Type 3 is used to carry quick acknowledgements for successful access attempts. Each acknowledgement information element contains the MAC ID associated with the user terminal for which the acknowledgment is sent. A quick acknowledgement is used to inform the user terminal that its access request has been received but is not associated with an assignment of FCH/RCH resources. Conversely, an assignment-based acknowledgement is associated with an FCH/RCH assignment. If the user terminal receives a quick acknowledgement on the FCCH, it transitions to the Dormant state. If the user terminal receives an assignment-based acknowledgement, it obtains scheduling information sent along with the acknowledgment and begins using the FCH/RCH as assigned in the current TDD frame.
The user terminal resumes the access procedure on the RACH if it does not receive an acknowledgement on the FCCH within a specified number of TDD frames after transmitting the RACH PDU. In this case, the user terminal can assume that the access point did not receive the RACH PDU correctly. A counter is maintained by the user terminal to count the number of access attempts. This counter may be initialized to zero for the first access attempt and is incremented by one for each subsequent access attempt. The user terminal would terminate the access procedure if the counter value reaches the maximum number of attempts.
For each subsequent access attempt, the user terminal first determines various parameters for this access attempt including (1) the amount of time to wait before transmitting the RACH PDU, (2) the RACH slot to use for the RACH PDU transmission, and (3) the rate for the RACH PDU. To determine the amount of time to wait, the user terminal first determines the maximum amount of time to wait for the next access attempt, which is referred to as the contention window (CW). In an embodiment, the contention window (which is given in units of TDD frames) exponentially increases for each access attempt (i.e., CW=2access
For the next access attempt, the user terminal reduces the rate for the RACH PDU, if the lowest rate was not used for the last access attempt. The initial rate used for the first access attempt may be selected based on the received SNR of the pilot sent on the BCH. The failure to receive an acknowledgment may be caused by the access point's failure to correctly receive the RACH PDU. Thus, the rate for the RACH PDU in the next access attempt is reduced to improve the likelihood of correct reception by the access point.
After waiting the randomly selected wait time, the user terminal again randomly selects an RACH slot for transmission of the RACH PDU. The selection of the RACH slot for this access attempt may be performed in similar manner as that described above for the first access attempt, except that the RACH parameters (i.e., number of RACH slots, slot duration, and guard interval) for the current TDD frame, as conveyed in the BCH message, are used along with the current RACH PDU length. The RACH PDU is then transmitted in the randomly selected RACH slot.
The access procedure described above continues until either (1) the user terminal receives an acknowledgment from the access point or (2) the maximum number of permitted access attempts has been reached. For each access attempt, the amount of time to wait before transmitting the RACH PDU, the RACH slot to use for the RACH PDU transmission, and the rate for the RACH PDU may be selected as described above. If the acknowledgment is received, then the user terminal operates as indicated by the acknowledgment (i.e., it waits in the Dormant state if a quick acknowledgment is received or starts using the FCH/RCH if an assignment-based acknowledgment is received). If the maximum number of permitted access attempts has been reached, then the user terminal transitions back to the Init state.
The access point schedules downlink and uplink transmissions on the FCH and RCH and further controls the rates for all active user terminals. Moreover, the access point adjusts the transmit power of certain active user terminals on the uplink. Various control loops may be maintained to adjust the rate, transmit power, and timing for each active user terminal
Fixed and Variable Rate Services
The access point can support both fixed and variable rate services on the FCH and RCH. Fixed-rate services may be used for voice, video, and so on. Variable rate services may be used for packet data (e.g., Web browsing).
For fixed rate services on the FCH/RCH, a fixed rate is used for the entire connection. Best effort delivery is used for the FCH and RCH (i.e., no retransmission). The access point schedules a constant number of FCH/RCH PDUs per specified time interval to satisfy the QoS requirements of the service. Depending on the delay requirements, the access point may not need to schedule an FCH/RCH PDU for every TDD frame. Power control is exercised on the RCH but not the FCH for fixed rate services.
For variable rate services on the FCH/RCH, the rate used for the FCH/RCH is allowed to change with channel conditions. For some isochronous services (e.g., video, audio), the QoS requirements may impose a minimum rate constraint. For these services, the scheduler at the access point adjusts the FCH/RCH allocation so that a constant rate is provided. For asynchronous data services (e.g., web browsing, file transfer, and so on), a best effort delivery is provided with the option of retransmissions. For these services, the rate is the maximum that can be reliably sustained by the channel conditions. The scheduling of the FCH/RCH PDUs for the user terminals is typically a function of their QoS requirements. Whenever there's no data to send on the downlink/uplink, an Idle PDU is sent on the FCH/RCH to maintain the link. Closed loop power control is not exercised on the FCH or RCH for variable rate services.
Rate Control
Rate control may be used for variable rate services operating on the FCH and RCH to adapt the rate of the FCH/RCH to changing channel conditions. The rates to use for the FCH and RCH may be independently controlled. Moreover, in the spatial multiplexing mode, the rate for each wideband eigenmode of each dedicated transport channel may be independently controlled. The rate control is performed by the access point based on feedback provided by each active user terminal. The scheduler within the access point schedules data transmission and determines rate assignments for the active user terminals.
The maximum rate that can be supported on either link is a function of (1) the channel response matrix for all of the data subbands, (2) the noise level observed by the receiver, (3) the quality of the channel estimate, and possibly other factors. For a TDD system, the channel may be considered to be reciprocal for the downlink and uplink (after calibration has been performed to account for any differences at the access point and user terminal). However, this reciprocal channel does not imply that the noise floors are the same at the access point and user terminal Thus, for a given user terminal, the rates on the FCH and RCH may be independently controlled.
Closed-loop rate control may be used for data transmission on one or more spatial channels. Closed-loop rate control may be achieved with one or multiple loops. An inner loop estimates the channel conditions and selects a suitable rate for each spatial channel used for data transmission. The channel estimation and rate selection may be performed as described above. An outer loop may be used to estimate the quality of the data transmission received on each spatial channel and to adjust the operation of the inner loop. The data transmission quality may be quantified by packet error rate (PER), decoder metrics, and so on, or a combination thereof. For example, the outer loop may adjust the SNR offset for each spatial channel to achieve the target PER for that spatial channel. The outer loop may also direct the inner loop to select a lower rate for a spatial channel if excessive packet errors are detected for the spatial channel.
Downlink Rate Control
Each active user terminal can estimate the downlink channel based on the MIMO pilot transmitted on the BCH in each TDD frame. The access point may also transmit a steered reference in an FCH PDU sent to a specific user terminal. Using the MIMO pilot on the BCH and/or the steered reference on the FCH, the user terminal can estimate the received SNR and determine the maximum rate that can be supported on the FCH. If the user terminal is operating in the spatial multiplexing mode, then the maximum rate may be determined for each wideband eigenmode. Each user terminal can send back to the access point the maximum rate supported by each wideband eigenmode (for the spatial multiplexing mode), the maximum rate supported by the principal wideband eigenmode (for the beam-steering mode), or the maximum rate supported by the MIMO channel (for the diversity mode) in the FCH Rate Indicator field of the RCH PDU. These rates may be mapped to received SNRs, which may then be used to perform the water-filling described above. Alternatively, the user terminal may send back sufficient information (e.g., the received SNRs) to allow the access point to determine the maximum rate supported by the downlink.
The determination of whether to use the diversity, beam-steering, or spatial multiplexing mode may be made based on the feedback from the user terminal. The number of wideband eigenmodes selected for use may increase as isolation between the steering vectors improves.
The scheduler uses the maximum rates that the downlink can support for each active user terminal to schedule downlink data transmission in subsequent TDD frames. The rates and other channel assignment information for the user terminal are reflected in an information element sent on the FCCH. The rate assigned to one user terminal can impact the scheduling for other user terminals. The minimum delay between the rate determination by the user terminal and its use is approximately a single TDD frame.
Using the Gram-Schmidt ordered procedure, the access point can accurately determine the maximum rates supported on the FCH directly from the RCH preamble. This can then greatly simplify rate control.
Uplink Rate Control
Each user terminal transmits a steered reference on the RACH during system access and on the RCH upon being assigned FCH/RCH resources. The access point can estimate the received SNR for each of the wideband eigenmodes based on the steered reference on the RCH and determine the maximum rate supported by each wideband eigenmode. Initially, the access point may not have a good estimate of the channel to permit reliable operation at or near the maximum rate supported by each wideband eigenmode. To improve reliability, the initial rate used on the FCH/RCH may be much lower than the maximum supported rate. The access point can integrate the steered reference over a number of TDD frames to obtain improved estimate of the channel. As the estimate of the channel improves, the rate may be increased.
Power Control
Power control may be used for uplink transmissions on the RCH (instead of rate control) for fixed-rate services. For fixed-rate services, the rate is negotiated at call setup and remains fixed for the duration of the connection. Some fixed-rate services may be associated with limited mobility requirement. In an embodiment, power control is implemented for the uplink to guard against interference among the user terminals but is not used for the downlink.
A power control mechanism is used to control the uplink transmit power of each active user terminal such that the received SNR at the access point is maintained at a level that achieves the desired service quality. This level is often referred to as the target received SNR, the operating point, or the setpoint. For a mobile user terminal, the propagation loss will likely change as the user terminal moves about. The power control mechanism tracks changes in the channel so that the received SNR is maintained near the setpoint.
The power control mechanism may be implemented with two power control loops—an inner loop and an outer loop. The inner loop adjusts the transmit power of the user terminal such that the received SNR at the access point is maintained near the setpoint. The outer loop adjusts the setpoint to achieve a particular level of performance, which may be quantified by a particular frame error rate (FER) (e.g., 1% FER), packet error rate (PER), block error rate (BLER), message error rate (MER), or some other measure.
The initial setpoint used at the access point is set to achieve a particular level of performance. This setpoint is adjusted by the outer loop based on the FER or PER for the RCH. For example, if no frame/packet errors occur over a specified time period, then the setpoint may be reduced by a first amount (e.g., 0.1 dB). If the average FER is exceeded by the occurrence of one or more frame/packet errors, then the setpoint may be increased by a second amount (e.g., 1 dB). The setpoint, hysteresis margin, and outer loop operation are specific to the power control design used for the system.
Timing Control
Timing control may be advantageously used in a TDD-based frame structure where the downlink and uplink share the same frequency band in a time division duplexed manner. The user terminals may be located throughout in the system and may thus be associated with different propagation delays to the access point. In order to maximize efficiency on the uplink, the timing of the uplink transmission on the RCH and RACH from each user terminal can be adjusted to account for its propagation delay. This would then ensure that the uplink transmissions from different user terminals arrive within a particular time window at the access point and do not interfere with one another on the uplink, or with the downlink transmission.
The user terminal receives the initial timing advance from the access point and thereafter uses this timing advance on all subsequent uplink transmissions on both the RCH and RACH. If the user terminal is assigned FCH/RCH resources, then its timing advance can be adjusted by commands sent by the access point in the RCH Timing Adjustment field of an FCCH information element. The user terminal would thereafter adjust its uplink transmissions on the RCH based on the current timing advance, which is equal to the initial timing advance plus all timing adjustments sent by the access point to the user terminal
Various parts of the MIMO WLAN system and various techniques described herein may be implemented by various means. For example, the processing at the access point and user terminal may be implemented in hardware, software, or a combination thereof. For a hardware implementation, the processing may be implemented within one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, micro-controllers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof
For a software implementation, the processing may be implemented with modules (e.g., procedures, functions, and so on) that perform the functions described herein. The software codes may be stored in a memory unit (e.g., memory 732 or 782 in
Headings are included herein for reference and to aid in locating certain sections. These headings are not intended to limit the scope of the concepts described therein under, and these concepts may have applicability in other sections throughout the entire specification.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
This application is a continuation of U.S. application Ser. No. 12/112,728, filed Apr. 30, 2008, which is a divisional application of, and claims the benefit of priority from, U.S. patent application Ser. No. 10/693,419, filed Oct. 23, 2003 and entitled “MIMO WLAN System,” which claims the benefit of priority from U.S. Provisional Patent Application Ser. No. 60/421,309, filed Oct. 25, 2002 and entitled “MIMO WLAN System.” All of said applications are assigned to the assignee of this application and are fully incorporated herein by reference for all purposes.
| Number | Date | Country | |
|---|---|---|---|
| 60421309 | Oct 2002 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10693419 | Oct 2003 | US |
| Child | 12112728 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 12112728 | Apr 2008 | US |
| Child | 14458933 | US |
