The present invention generally relates to wireless telecommunication, and more particularly, is related to a system and method for providing efficient high speed wireless media access via overlay of feedback and data channels within the same spectrum.
The convergence of communications and computing technologies is leading to the next generation of networks where many millions of homes and small offices are connected full-time to the Internet, thereby leading to a truly global information age. These developments have sparked an interest in home networking. Several technologies are presently being considered for home networks, one of which is wireless technology. Wireless systems enable non-restrictive communication and eliminate the need to rewire homes for the new networks. Bringing any new technology into the home adds new challenges such as cost, reliability and ease of management, in addition to the optimal use of resources. Home networks need to support a wide variety of devices ranging from a television set to a light bulb.
Random access protocols are well suited for such networks because they can multiplex a wide range of data rates efficiently. Random access protocols are best suited for an ad-hoc wireless network because they require no central control. An ad-hoc wireless network is a collection of wireless nodes that can communicate with each other without the aid of any pre-existing communication infrastructure.
A medium access control (MAC) protocol enables multiple users, via multiple devices, to transmit and receive data on a shared wireless medium in an efficient and orderly manner. Wireless MAC protocols have been extensively researched and a large group of protocols proposed. With the exception of high performance local area network (HIPERLAN), as addressed below, all high speed MAC protocols require a central node such as a base station to coordinate transmissions. Such a central node is not possible in an ad-hoc network. As a result, random access is the multiple access mechanism in the ad-hoc networks. Most wireless random access protocols are designed for low data rates (<2 Mbps) and are based on collision avoidance principles.
Advancements in radio frequency and wireless communication technology are driving the data rates of wireless local area networks (WLAN) higher. In fact, the total bandwidth required by applications like digital video and audio is in excess of 50 Mbps. Unfortunately, recent standards developed for ad-hoc wireless LANs, including HIPERLAN, perform poorly at high data rates and for small packet sizes. The efficiency of random access protocol is determined by how fast collisions within a communication channel are detected and how soon this information can be conveyed to a source node. In wired protocols the ability of a node to listen to a medium while transmitting, and the fact that a collision on the medium is heard by all nodes listening to the medium, results in high performance.
In a wireless medium both the above assumptions do not exist. First, a wireless transceiver cannot be transmitting and listening to the medium at the same time due to self-interference. Basically, any transmitted signal that leaks into the receiver usually has a much higher energy than the received signal, and hence, transceivers cannot listen and transmit at the same time. In the wireless medium the signal strength falls off as a function of distance. Hence, depending on the position of a node relative to the source node, channel sensing will produce different results. Consider, for example, the scenario illustrated in
Those skilled in the art are perplexed by the problem of determining how to enable the destination node to convey information about its state of transmission, either idle or collision, to the source node in a timely manner. Unfortunately, a feedback channel is not available in current wireless systems since transmitted and received signal strengths differ by orders of magnitude and current technology does not provide enough isolation between transmit and receive paths to allow correct demodulation of the received signal when transmitting data.
Current collision avoidance protocols try to minimize collisions by exchanging handshaking messages to reserve the communication channel for data transmission. This handshaking can be considered as duplexing feedback information in time. Handshaking requires a node to switch between transmit and receive modes. More importantly, intrinsic to the handshaking protocols are turn-around periods called mute-deaf times. These are periods, when the transceiver is switching from transmit mode to receive mode or vice-versa. During these periods the transceiver can neither listen nor transmit. These switching time constraints are a large overhead at high data-rates.
Another alternative is to use a different frequency bands for the data channel and the feedback channel. This is referred to as frequency duplexed feedback, which is used by busy-tone protocols. However, this has a higher hardware cost, and low spectral efficiency, although it results in very efficient rate-scalable wireless MAC protocols. The high cost results from the requirement of two transceivers, one for the data channel and the other for the feedback channel. Even though the feedback channel is a narrow-band channel, a significant amount of bandwidth needs to be allocated for it because it is hard to implement band-pass filters that have small bandwidth and sharp cutoff.
One example of such a protocol is the receiver initiated busy tone multiple access protocol (RI-BTMA). Operation of the RI-BTMA protocol is relatively simple. When a node has data to transmit, it senses the busy-tone channel for any ongoing transmissions. If the channel is idle, the node will initiate a data transmission. A node will then send a busy tone in an out of band channel after it knows that it is the destination for this packet. If the node receives a busy tone signal it will continue with the transmission. Else, it would abort the transmission and retry after a random time. The efficiency of this protocol is limited by how fast the destination node can be determined and the busy tone asserted. As a result, such protocols are efficient over a wide range of data rates unlike handshaking protocols. However, the hardware cost of two transceivers and the inefficiency of a separate busy tone channel, which is attributed to the significant amount of bandwidth that has to be allocated to convey one bit of information, prohibits the use of such protocols.
In light of the foregoing, the present invention provides a cost effective system and method for providing high speed multi-media access within a telecommunication network.
Generally, the system can be implemented as follows. A local transceiver is provided, which is capable of transmitting data via a transmit path while receiving a feedback signal via a receive path. The local transmitter is connected to a feedback generator for generating and transmitting a feedback signal in response to receiving data from a wireless transceiver. A feedback detector is connected to the local transceiver for detecting feedback signals received from another wireless transceiver.
The present invention can also be viewed as providing a method for providing high speed multi-media access within a telecommunication network. In this regard, the method can be broadly summarized by the following steps: testing for a specified amount of energy within a data channel of a first wireless transceiver; decoding data associated with the specified amount of energy in response to detecting the specified amount of energy; the series of wireless transceivers transmitting a feedback signal within the series of wireless transceivers in order to stop the transmission of data within the series of wireless transceivers until a destination of the data is determined; identifying a destination address, associated with a destination wireless transceiver, for the data from the decoded data; and ceasing the transmission of the feedback signal by all wireless transceivers within the series of wireless transceivers, except by the destination wireless transceiver.
The invention has numerous advantages, a few of which are delineated hereinafter as examples. Note that the embodiments of the invention, which are described herein, possess one or more, but not necessarily all, of the advantages set out hereafter.
One advantage is that a wireless standard is provided, which has high efficiency at high data rates.
Another advantage is that it provides high efficiency without requiring addition transceivers.
Other features and advantages of the present invention will become apparent to one of reasonable skill in the art upon examination of the following drawings and detailed description. It is intended that all such additional features and advantages be included herein within the scope of the present invention as described by the claims.
The invention can be better understood with reference to the following drawings. The components of the drawings are not necessarily to scale, emphasis instead being placed upon clearly illustrating the principles of the present invention. Moreover, in the drawings, like referenced numerals designate corresponding parts throughout the several views.
Referring now to the drawings, wherein like reference numerals designate corresponding parts throughout the drawings,
In an ad-hoc network 102 all data transmission and reception is in the same frequency band since there are no special terminals (hereinafter, nodes) to translate the transmission from one frequency band to another. Therefore, the ad-hoc network 102 operates in a time division duplex mode.
Referring back to
Referring to
All nodes within the network 102 are in one of the following three modes; transmitting data, receiving data or listening to the channel used for communication between the nodes. Based on the state and location of the node, the signal received by the node is different. The problems associated with each of these modes is described hereinbelow. To best illustrate these problems, a signal that is received by a node in each state is formulated, thereby defining its decoding problem.
Without loss of generality, a node i, the source, is considered to be transmitting data to a node j the destination. Let si(t) be the signal transmitted by source i. Similarly, fj(t) is the feedback signal transmitted by the destination node j. The noise on the wireless channel is denoted by n(t). hdij(t) is the channel response seen by node j for the data signal transmitted by node i. hfj,i(t) is the channel response for the feedback signal transmitted by node j at node i.
Data Transmitter Problem
When a node is transmitting data, a part of the transmitted signal leaks into the receiver. This leakage signal is a delayed (td) and attenuated (α) version of the transmitted signal. Also, the channel shaped feedback signal generated by the destination, and any other data transmission in its range, are also received. The received signal is represented by the equation:
where k is the set of all nodes in the transmit radius of i that are transmitting data. It is assumed that the communication channel does not change over the duration of the packet transmission. The data transmitter should be able to decode the feedback signal, fj(t), in the presence of interference from its own transmission and any other adjacent data transmissions. It is necessary to reduce self-interference to such an extent that it does not swamp-out the feedback signal. As illustrated hereinbelow, one way to achieve this is to use echo cancelation to make α as small as possible. Also, by positioning the feedback channel at a location in the data spectrum where there is little data energy (spectral null) the self interference can be minimized.
Data Receiver Problem
Considering the case when a node is receiving data, the node must send the feedback signal concurrently. Unfortunately, the feedback signal leaks into the data reception path and interferes with data demodulation. The leakage signal in the receive path is a delayed and attenuated version of the feedback signal. The received signal may be represented by the equation:
where k is the set of all nodes in the transmit radius of j that are transmitting the feedback signal. The data transmitter should be able to decode the data signal si(t) in the presence of multiple interfering feedback signals. The position of the feedback channel is crucial. As shown hereinbelow, placing the feedback channel at the spectral null of the receiver matched filter minimizes the influence of the feedback signal on the demodulation of data, as the matched filter rejects most of the energy around its null frequency.
Listener Problem
A node in listen state is neither transmitting nor receiving data. Instead, it is waiting for its turn to send or receive data. This node should not initiate a transmission if it will collide with any ongoing transmissions. To satisfy this condition, the node should be able to detect the feedback transmission from any one of the nodes that are within its transmission range. The received signal may be represented by the equation:
where l and k are the set of the transmitting and the receiving nodes respectively, within the transmit range of the listening node i. For correct operation of the protocol, the listener should not initiate a transmission if any one of the feedback transmissions is active. The listener node should reliably detect any feedback transmission in the presence of interfering data transmissions and noise. False detection of the feedback signal that is really not present decreases the efficiency of the protocol, however, non detection of a feedback signal that exists results in destroying an ongoing transmission.
In addressing the above-mentioned problems, it should be noted that theoretically, full duplex communication should be possible between a source node and a destination node since each node knows its own transmitted signal and therefore can cancel self-interference. Unfortunately, such a receiver is not yet possible with current-day technology. However, a narrow-band feedback channel can be enabled within the data channel of a communication channel by taking advantage of the fact that the power spectrum of a modulation scheme contains nulls, and any energy at a null frequency is orthogonal to the data transmission. Transmission in the feedback channel results in minimal interference because a receiver's front-end matched filter also has a null at that frequency and removes much of the energy at the null frequency. Thus, the cross interference between the data and feedback channels can be minimized.
Typical circulators can achieve only 20 dB isolation between undesired ports. As a result, an attenuated version of the transmitted signal leaks into the receive path. Since the received signal has a much lower signal strength than the transmitted signal, the RF echo canceler is needed in the feedback channel band to minimize self-interference. This modified RF front-end achieves reasonable isolation between the transmit and receive paths in the frequency band of the feedback channel. Reasonable isolation is accordance with the preferred embodiment of the invention is further described hereinbelow.
The feedback detector 252 further comprises a bandpass filter 254 to select a feedback channel. The bandpass filter 254 is followed by an energy detector 256 which estimates the signal energy in the feedback channel. The energy detector 256 is connected to a threshold device 258 which makes decisions on whether the energy in the feedback channel is noise or a valid busy signal.
The feedback generator 232 further comprises an oscillator 234 at the feedback channel's center frequency which injects energy into the feedback channel.
As shown by block 302, when a node is in listening mode, a receive phase is asserted. The transceiver 202 then determines if the node is receiving data (block 304). If the node is already receiving data, a data detect signal is asserted and a feedback signal is received by every node within the transmit radius of the transceiver 202 (block 306). As shown by block 308, if the node is not receiving data, the demodulator 274 is enabled and data can be received by the transceiver 202.
Design Choice and Challenge
Two main challenges exist in making the wireless transceiver 202 feasible, namely the isolation required between the transmit and receive paths, and the spectral position of the feedback channel. The following describes these challenges in detail, thereby further defining functionality of the wireless transceiver 202.
Addressing the issue of isolation, it is well known that the traditional RF transceiver based on time division duplex comprises a transmitter and a receiver which time-share the same RF front-end. At any given time only one of the paths is connected to an antenna. With multi-path fading and path loss, the dynamic range of a received signal in a local area network (LAN) environment is approximately 60–80 dB. As the minimum detectable signal is in the microvolt range, the signal that leaks into a LNA through a duplexer from a transmit power amplifier may significantly desensitize the LNA. This is the main reason why the transmit and the receive time slots are offset in TDD systems. Therefore isolation is a major concern for having full duplex communication on wireless systems.
It is feasible to isolate the transmit and receive paths by 72 dB over a 200 kHz radio channel by using a dual-antenna and an adaptive RF echo canceler. Such self-interference rejection in the narrow-band feedback channel suffices for transceiver operation in accordance with the preferred embodiment of the invention. Alternatively, isolation can be further improved by using a baseband digital echo canceler. The preferred transceiver requires such isolation only in the frequency band used for the feedback channel, which is a much smaller band compared to the data channel.
Addressing the issue of the spectral position of the feedback channel, it should be noted that if self interference could be completely canceled, the position of the feedback channel within the databand would not be a factor at all and full duplex communication would be feasible. In practical cases, however, the feedback channel should be located at a position where it would have minimum interference on data demodulation. This reduces the design requirements on the RF echo canceler. For example, if the feedback channel and data channel have the same frequency, then the feedback signal should be completely canceled or it would result in a DC offset at the demodulator output.
A coherent receiver is a commonly used receiver structure in digital communications. The received signal is cross-correlated with each candidate pulse shape and the pulse shape that correlates best with the received signal is assumed to be the symbol transmitted. Letting R(t) be the received signal, R(t) comprises the transmitted data signal (S(t)), additive white gaussian noise (n(t)) and an interfering signal (I(t)). The transmitted signal S(t) is a pulse pn(t) where the transmitted pulses are orthogonal and of equal energy.
The received waveform (R(t)=pn(t)+n(t)+I(t)) is correlated with each one of possible N pulse shapes. The correlation with the ith pulse is given by the equation:
Ki=∫∞∞R(t)pi (t)dt=∫∞∞(pn(t)+n(t)+I(t))pi(t)dt (Eq. 4)
As illustrated, the equation consists of three terms. The third term is introduced by the feedback signal and other interfering signals that are transmitted during data transmission. The condition for this interference to not affect the data reception is
∫∞∞I(t)pi(t)=0.
In the simplest case, if it is assumed that the feedback signal is a sinusoid and there are no other interferers, then correlation of the interference term becomes
∫∞∞A sin(2πfTt)pi(t)=A sin(2πfTt)*pi(−t)|t=0=Pi(fT)
where Pi(f) is the Fourier transform of the pulse pi(t).
Hence, if the feedback is introduced at the spectral null of the data power spectrum, correlation receiver can demodulate the data signal with little interference. Some modulation pulses have spectral nulls at the symbol frequency. Spectral nulls can be introduced at any frequency or a finite set of frequencies using a filter at the transmitter at the cost of increasing the peak transmitter power. In wide-band modulation techniques like code division multiple access (CDMA), a narrow-band signal in the data channel is like a narrow-band interferer. Spread-spectrum systems are designed to work well in the presence of such narrow-band interference. Thus, in spread spectrum channels the presence of a feedback signal causes very little interference to the demodulation of data.
Wireless Collision Detect Protocol
In accordance with the preferred embodiment of the invention, a wireless collision detect protocol (WCD) is used for the current ad-hoc network 102, and all ad-hoc networks confined within a short radius (≦50 m). It should be noted that the WCD protocol can operate correctly in an unslotted environment, however for simplicity, it is assumed that the system is slotted and the unslotted version is not discussed further herein.
The following defines a few terms used in the WCD protocol description. First, the slot should be long enough such that the carrier can be detected and the feedback signal asserted by the most distant node in the system. This consists of three components; a round trip time, time to detect the carrier, and time to assert the feedback signal. This amount of time is called the idle detection interval (IDI) since it is the minimum period of time a node needs to sample the feedback channel to ascertain if the data channel is idle.
The time period taken to determine the destination of the current transmission and assert on the feedback channel is defined as the receiver detection interval (RDI). The RDI comprises the time for a node to decode the transmission, identify itself as the destination, and assert the feedback signal.
The frequency band is split into a data channel and a feedback channel (FC). Two logically different signals are transmitted in the feedback channel. The first is the carrier detect (CD) signal which is asserted when the start of a burst transmission is detected. The second is an address detect (AD) signal which indicates that an address has been matched with a station identification and data is being received. Station Identification is a unique number assigned to a terminal. Basically, it can be something similar to the Ethernet address attached to every hardware device. This signal is similar to the receiver initiated feedback in RI-BTMA. The additional CD signal does not add extra complexity or hardware requirements.
The channel is said to be idle when there is no signal in the feedback channel. Any node in the network is either in data receive mode or data transmit mode. If the node does not have any packets to transmit or is waiting to transmit a packet, it is in data receive mode. When a new packet arrives, the node picks a random number in the range [0, W] and stores it in a backoff counter. This counter is decremented for each idle slot. When the counter goes to zero, the node switches from data receive mode to data transmit mode. The protocol is described hereinbelow and illustrated with reference to
Further describing the example illustration of the protocol depicted in
Rules For a Node in Data Receive Mode
Rules for a Node in Data Transmit Mode
As a brief summary of the properties of the WCD protocol, the following is offered. The time for which the packet is vulnerable for collision is very small because carrier detection usually takes about 10–20 symbol periods while collision detection takes more than a hundred symbols. Under the assumption of reciprocal wireless links this protocol eliminates the hidden node problem and the exposed node problem.
Addressing the elimination of the hidden node problem, consider the same four nodes as illustrated by
Considering the same scenario as that described hereinabove, except now the second node 106 is transmitting to the first node 104. In accordance with the wireless collision detect protocol, the second node 106 can send data to the first node 104, and the third node 108 can simultaneously send data to the fourth node 112. For example, the second node 106 sends data to the first node 104, and the first node 104 asserts on the feedback channel. The third node 108 does not hear this feedback signal and hence can start a transmission. Since the fourth node 112 is out of range of the second node, it can receive the third node's 108 transmission and it asserts on the feedback channel.
The third node 108, on detecting the feedback signal completes its data transmission. All nodes in transmitting range of the sender will hear the data transmission and only those nodes in the transmit range of the destination will sense the feedback signal. Any node out of range of the destination node can have a parallel conversation with any node out of range of the sender. Although the third node 108 can send to the fourth node 112, the fourth node 112 cannot transmit to the third node 108 because the third node is in range of two senders (the second node 106 and the fourth node 112, in this case).
Appendix A provides further disclosure of the analysis and simulation of the protocol used within the wireless transceiver 202, which is herein incorporated by reference.
It should be emphasized that the above-described embodiments of the present invention, particularly, any “preferred” embodiments, are merely possible examples of implementations, merely set forth for a clear understanding of the principles of the invention. Many variations and modifications may be made to the above-described embodiment(s) of the invention without departing substantially from the spirit and principles of the invention. All such modifications and variations are intended to be included herein within the scope of this disclosure and the present invention and protected by the following claims.
This application claims priority to and the benefit of the filing dates of U.S. Provisional Application Ser. No. 60/153,292, filed on Sep. 10, 1999, and entitled “High Speed Wireless Media Access,” and U.S. Provisional Application Ser. No. 60/153,351, filed on Sep. 10, 1999, and entitled “Wireless Transceiver for High Speed Media Access,” both of which are incorporated by reference herein in their entirety.
Number | Name | Date | Kind |
---|---|---|---|
4755990 | Bohler et al. | Jul 1988 | A |
5022046 | Morrow, Jr. | Jun 1991 | A |
5319641 | Fridrich et al. | Jun 1994 | A |
5394391 | Chen et al. | Feb 1995 | A |
5418785 | Olshansky et al. | May 1995 | A |
5438683 | Durtler et al. | Aug 1995 | A |
5502724 | Chen et al. | Mar 1996 | A |
5608729 | Orsic | Mar 1997 | A |
5648961 | Ebihara | Jul 1997 | A |
5661727 | Kermani et al. | Aug 1997 | A |
6088343 | King | Jul 2000 | A |
6295285 | Whitehead | Sep 2001 | B1 |
6611495 | Meyer et al. | Aug 2003 | B1 |
6850617 | Weigand | Feb 2005 | B1 |
20020001291 | Willars et al. | Jan 2002 | A1 |
Number | Date | Country | |
---|---|---|---|
60153292 | Sep 1999 | US | |
60153351 | Sep 1999 | US |