Patent Application
20230155864
The present disclosure relates to systems and methods for operating the Radio Access Network (RAN) design for 4G- and 5G-based mobile networks, and relates more particularly to systems and methods to enable efficient uplink mMIMO performance for Open-RAN (O-RAN).
Traditionally, the radio access networks were built as an integrated unit where the entire RAN was processed. The RAN network traditionally uses application-specific hardware for processing, making them difficult to upgrade and evolve. As future networks evolve to have massive densification of networks to support increased capacity requirements, there is a growing need to reduce the CAPEX/OPEX costs of RAN deployment and make the solution scalable and easy to upgrade.
In the field of cloud-based Radio Access Networks (RAN), a significant portion of the RAN layer processing is performed at a central unit (CU) and a distributed unit (DU). Both CUs and DUs are also known as the baseband units (BBUs). CUs are usually located in the cloud on commercial off-the-shelf servers, while DUs can be distributed. Also, the RF and real-time functions can be processed in the remote radio unit (RU).
The context of the present disclosure is related to systems and methods to enable efficient uplink mMIMO performance for O-RAN-based RANs.
3GPP has defined multiple split options across the entire radio access network (RAN). There are various factors affecting the selection of the fronthaul split option, such as bandwidth, latency, implementation cost, virtualization benefits, complexity of the fronthaul interface, expansion flexibility, computing power, and memory requirement.
One of the most common splits that are standardized recently by the O-RAN alliance is split option 7-2x (Intra-PHY split). This split has multiple advantages such as simplicity, transport bandwidth scalability, beamforming support, interoperability, support for advanced receivers and inter-cell coordination, lower O-RU complexity, future proof-ness, and interface and functions symmetry.
One of the technologies that use the O-RAN 7-2x specifications is 4G/5G massive MIMO (mMIMO). In SU-MIMO, the gNB serves a single user on a set of time-frequency resources, whereas in MU-MIMO setup, the gNB serves multiple users on the same time-frequency resources. In UL, UEs send SRS signals over a relatively long period of time, which are sent to the O-DU from the O-RU via the fronthaul interface. Using the SRS signal, the O-DU then obtains the UL channel estimates of the UEs, using which gNB performs user pairing (in the case of MU-MIMO) and calculates UL combining/digital beamforming matrix for the scheduled user(s). The O-DU sends the combining/digital beamforming matrix elements or weights to the O-RU, which in return applies these elements to the frequency-domain PUSCH IQ samples received at a later time and sends the resulting IQ samples to the O-DU for the rest of PHY processing. One aspect of the combining operation (using the combining/digital beamforming matrix) at O-RU is that it enables the O-RU to reduce the number of streams transferred from O-RU to O-DU to a value smaller than the number of RX antennas. In mMIMO systems, the number of RX antennas is large, and transferring all received streams for PUSCH to O-DU incurs significant and prohibitive front-haul bandwidth consumption.
Although split option 7-2x enables multiple advanced features, such as beamforming and UL CoMP, the system performance may degrade in certain scenarios, such as UL mMIMO for high-speed UEs and/or interference-limited scenarios, where there is significant inter-cell interference (ICI) from neighboring cells contaminating the UL signal from the desired UEs.
Performance degradation in high mobility scenarios: The reason for such degradation in high-speed UEs scenarios is channel aging. Specifically, the SRS signals are sent over a relatively long period of time (e.g., 10 s of milliseconds). By the time the O-RU applies the combining/digital beamforming matrix elements to the PUSCH symbols, these matrix elements become inaccurate since the SRS signals are outdated (i.e., SRS to UL-data TTI delay is long). In other words, the channel gains between the UEs and the gNB at the time PUSCH is received are no longer reflected accurately by the SRS signals (since UEs are moving fast in such a scenario), which cause interference and hence throughput degradation during the combining process (using the combining/digital beamforming matrix) in the UL chain.
Performance degradation in the presence of inter-cell interference (ICI): The main reason for performance degradation in the presence of ICI is signal dimension reduction at O-RU. In particular, squeezing desired signal and interference signal into a smaller number of dimensions causes performance degradation of the interference rejection at O-DU.
One way to mitigate the performance degradation in high mobility scenarios is to configure SRS transmissions with shorter periodicities. However, shorter SRS periodicity may decrease the SRS capacity, i.e., the number of UEs that can be configured to transmit SRS, because more resources are occupied by each UE for SRS transmission. Also, in a time-division duplex (TDD) system, there are certain limitations on SRS periodicity due to the TDD frame structure adopted by the system.
Another approach to mitigate the performance degradation in high mobility scenarios is to predict the UL channel of the PUSCH/PUCCH slot at DU based on UL channels of a number of previous SRS transmissions. In addition to the computation complexity of the underlying prediction algorithm, this approach has its own limitations on performance improvement depending on UE speeds.
For scenarios with inter-cell interference, a typical O-DU processing uses a minimum mean-squared error receiver with interference rejection combining (MMSE-IRC) to mitigate the impact of the interference. However, since O-DU has access to the compressed PUSCH signal transferred from O-RU, the performance of MMSE-IRC may be degraded compared to a typical 7-3 receiver wherein MMSE-IRC is computed based on and applied to the full-dimension signal.
There are four planes specified in the O-RAN specs, namely user-plane (U-plane), control plane (C-plane), synchronization plane (S-plane), and management plane (M-plane). The existing O-RAN spec explains the exact method of sending the C-plane messages in the downlink (DL) direction and sending/receiving the U-plane messages between the O-DU and the O-RU.
The M-plane specifies the management plane protocols used over the fronthaul interface linking the O-RU (O-RAN Radio Unit) with other management plane entities, which can include the O-DU (O-RAN Distributed Unit), the O-RAN defined Service Management and Orchestration (SMO) functionality as well as other generic Network Management Systems (NMS).
In contrast to C-plane (described in the next section), Management Plane refers to non-real-time management operations between the O-DU and the O-RU. Specifically, The Lower-Layer Split M-plane (LLS-M) facilitates the initialization, configuration, and management of the O-RU to support the stated functional split
A NETCONF/YANG-based M-Plane is used for supporting the management features, including “startup” installation, software management, configuration management, performance management, fault management, and file management towards the O-RU. The M-Plane supports two architectural models:
The M-Plane provides the following major functionalities to the O-RU. These features are implemented using the NETCONF provided functions such as “Startup” installation, SW management, Configuration management, Performance management, Fault Management, and File Management. The data models representing the M-Plane are organized as a set of reusable YANG modules.
The purpose of the C-plane messages is to transmit data-associated control information required for the processing of user data (e.g., scheduling and beamforming commands). These messages are sent separately for DL-related commands and UL-related commands.
A common frame format is used for C-Plane messages, having a transport layer and an application layer. The application layer is within the transport layer payload and consists of a common header for time reference, followed by information and parameters dependent and specific to the Section Type in use. Multiple sets of section data of the same Section Type value can be lined up one after another within the payload. To minimize packet rate over the interface, a transmitter should fill messages with as many subsequent sections (with or without sequential section IDs) as possible.
Section Type 1 is used for most Downlink and Uplink radio channels. In accord with O-RAN specifications, the fields of Section type 1 can be explained as follows:
One drawback of the O-RAN operation with Cat A O-RU is its large fronthaul throughput requirement, especially in massive MIMO scenarios.
One drawback of O-RAN operation with Cat B O-RU is that the combining/digital beamforming matrix elements are calculated using SRS channel estimates (at O-DU) at a slot n-M which is earlier than the current slot n. Since the channel may change during the M slots between slot n-M and slot n, the combining/digital beamforming matrix may not properly match the UL channel of slot n. This can result in some performance degradation, particularly for UEs with higher mobility. Also, since O-DU operates on a combined version of the signal received from O-RU in slot n (not the full-dimension signal), the interference rejection performance in O-DU may be degraded.
Implementations as described herein provide methods to enable an efficient mMIMO system for O-RAN based RANs.
Implementations as described herein address the aforementioned issues of performance degradation of the O-RAN 7-2x based UL MIMO. Aspects of the implementation include the following:
Asymmetrical functional split (using different functional splits in DL and UL): In implementations, the functional split point in the UL is changed so that further signal processing can be done at O-RU.In one embodiment, O-RU performs a first level of DM-RS channel estimation on PUSCH/PUCCH and uses the result of this channel estimation to calculate the combining/digital beamforming matrix, and after performing combining/digital beamforming on the received signal, transfers the resulting signal to O-DU for the rest of processing. In another embodiment, O-RU performs both DM-RS channel estimation and equalization on PUSCH/PUCCH and transfers the equalized signal to O-DU for the rest of the processing. In yet another embodiment, O-RU performs DM-RS channel estimation, equalization, and demodulation on PUSCH/PUCCH and transfers the resulting demodulated signal to O-DU for the rest of the processing. Described are multiple signaling (between O-DU and O-RU) and processing options for functional splits.
Signaling mechanisms to enable asymmetrical functional splits: Described are C-plane message modifications and M-plane enhancements to enable different split architectures. To enable different asymmetrical functional splits mentioned above and allow interoperability across different O-RU and O-DU vendors, several M-plane and C/U-plane enhancements are described. In one embodiment, the O-RU reports, via the M-plane, to the O-RU controller as part of its capability the supported split points and supported techniques (e.g., channel estimation techniques, MIMO detection techniques, etc.). In another embodiment, the O-RU controller statically configures, via the M-plane, the O-RU with the static parameters and static configurations based on the reported capabilities to operate using specific methods. In another embodiment, new sections and section extensions to the C-plane messages are proposed to carry, in real-time, the parameters and configurations at the O-RU to enable operation in an innovative split architecture.
Described is an Open Radio Access Network (0-RAN) system, comprising:
In an embodiment where the O-DU transfer DM-RS information to the O-RU, the DM-RS information including DM-RS configuration parameters of the scheduled or paired UEs/layers are received by the O-RU, and the O-RU generates a DM-RS sequence using the DM-RS configuration parameters received from O-DU. In an implementation, the O-DU transfers DM-RS channel estimation method information to the O-RU to indicate a channel estimation method. In an implementation, the O-DU indicates a combining/digital beamforming matrix calculation method to the O-RU. In an implementation, the O-DU transfers a combining/digital beamforming matrix enhancement information to the O-RU. In an implementation, the O-DU indicates a combining/digital beamforming matrix calculation method to the O-RU. The O-RU estimates the UL channels of the UE(s) and/or layer(s) transmitted at a current slot n using the DM-RS information. In an implementation, the O-DU indicates a combining/digital beamforming matrix calculation method to the O-RU. In an implementation, the O-RU calculates the combining/digital beamforming matrix. In an implementation, the O-RU applies the calculated combining/digital beamforming matrix to a received signal and transfers the calculated combining/digital beamforming matrix to the O-DU. In an implementation, the O-DU performs a L1/L2 processing on the resulting signal.
In an embodiment where i) a demodulation reference signal (DM-RS)-based channel estimation is performed by the O-DU in the DL and by the O-RU in the UL, and ii) an equalization is performed by the O-DU in the DL and by the O-RU in the UL. In an implementation, where the O-DU transfers DM-RS information to the O-RU, the DM-RS information including DM-RS configuration parameters of the scheduled or paired UEs/layers are received by the O-RU, and the O-RU generates a DM-RS sequence using the DM-RS configuration parameters received from O-DU. In an implementation, the O-RU estimates the UL channels of the UE(s) and/or layer(s) transmitted at a current slot n using the DM-RS information. In an implementation, prior to the estimation the UL channels of the UE(s) and/or layer(s) transmitted at a current slot n using the DM-RS information, the O-DU indicates ta DM-RS channel estimation method information to the O-RU. In an implementation, the O-DU transfers DM-RS channel estimation method information to the O-RU to indicate the channel estimation method. In an implementation, the O-RU equalizes the UL channels of the one or more UEs, the one or more layers, or both transmitted at a current slot n. In an implementation, the O-RU calculates a supplementary demodulation information and transfers the supplementary demodulation information along with an equalized signal stream to the O-DU. In an implementation the O-DU performs a demodulation on a signal of slot n transferred from the O-RU and a L1/L2 processing on the demodulated signal.
In an embodiment, i) a demodulation reference signal (DM-RS)-based channel estimation is performed by the O-DU in the DL and by the O-RU in the UL, ii) an equalization is performed by the O-DU in the DL and by the O-RU in the UL, and iii) demodulation is performed by the O-DU in the DL and by the O-RU in the UL. In an implementation, the O-DU transfer DM-RS information to the O-RU, the DM-RS information including DM-RS configuration parameters of the scheduled or paired UEs/layers are received by the O-RU, and the O-RU generates a DM-RS sequence using the DM-RS configuration parameters received from O-DU. In an implementation, the the O-RU estimates the UL channels of the UE(s) and/or layer(s) transmitted at a current slot n using the DM-RS information. In an implementation, the, prior to the estimation the UL channels of the UE(s) and/or layer(s) transmitted at a current slot n using the DM-RS information, the O-DU indicates ta DM-RS channel estimation method information to the O-RU. In an implementation, the O-DU transfers DM-RS channel estimation method information to the O-RU to indicate the channel estimation method. In an implementation, the O-RU equalizes the UL channels of the one or more UEs, the one or more layers, or both, transmitted at a current slot n. In an implementation, prior to a slot n, the DU transfers modulation information to the RU and the O-RU performs demodulation on an equalized signal. In an implementation, the O-DU performs a descrambling operation on the equalized signal of slot n transferred from the O-RU and a L1/L2 processing on the descrambled signal.
In an embodiment, i) the demodulation reference signal (DM-RS)-based channel estimation is performed by the O-DU in the DL and by the O-RU in the UL, ii) the equalization is performed by the O-DU in the DL and by the O-RU in the UL, and iii) demodulation is performed by the O-DU in the DL and by the O-RU in the UL.
In an embodiment, described is a method of operating an Open Radio Access Network (O-RAN) system, comprising:
In an embodiment, a method of demodulation reference signal (DM-RS)-based channel estimation is performed by the O-RU in an OFDM system with N subcarriers, comprising at least one of:
In an implementation, the method further comprises: the MMSE channel estimation being calculated as follows:
where ki represents a subcarrier index, lj represents a symbol index, rm is a receiver antenna index, and tn is a DM-RS transmission port index.
In an implementation, the method further comprises:
In an implementation, the method further comprises:
Implementations as described herein provide methods to enable an efficient mMIMO system for O-RAN based RANs.
In order to mitigate the issues of UL 7-2x, an asymmetrical functional split between DL and UL is proposed. The DL processing follows the O-RAN split 7-2x (either Cat A or Cat B) to balance the tradeoff between O-RU complexity and fronthaul throughput, where the UL processing follows a new split architecture. Three different UL functional splits corresponding to three different innovative O-RU categories are summarized in Table 1.
The following describes detailed processing steps for UL functional splits with O-RU Cat C1, O-RU Cat C2, O-RU Cat C3.
Embodiment 1 describes the processing steps for UL functional split with O-RU Cat C1, in which O-RU performs a “DM-RS based Channel Estimation” in the UL.
As shown in
As shown in
Option 1-1-A: “DM-RS configuration parameters” of the scheduled or paired UEs/layers. O-RU generates the “DM-RS sequence” using the “DM-RS configuration parameters” received from O-DU. The “DM-RS configuration parameters” are summarized in Table 4. Also, the details of the “DM-RS sequence” generation and mapping using the “DM-RS configuration parameters” are provided below under section Embodiment 6: 5G NR DM-RS.
Option 1-1-B: “DM-RS sequence” of the scheduled or paired UEs/layers.
As shown in
The above channel estimation methods are described in more detail in Embodiment 7: Channel Estimation. The “details of the indicated channel estimation method” depends on the indicated channel estimation method and may include any of the following:
Indication of the “channel estimation method” and/or “details of the indicated channel estimation method” can be done based on one of the following options:
Option 1-2-A: O-DU indicates the “channel estimation method” to the O-RU using a particular bitfield in the signaling from O-DU to O-RU. O-RU determines the channel estimation details, including parameters and processing steps based on its own implementation. An example of a mapping between the bitfield content and corresponding channel estimation methods is provided in Table 5.
Option 1-2-B: O-DU indicates “channel estimation method” and “details of the indicated channel estimation method” to the O-RU using particular bitfields in the signaling from O-DU to O-RU. O-RU uses a table, such as Table 5, to find the channel estimation method and uses other look-up tables, such as Table 6 to Table 9, to obtain the channel estimation details using these indications from O-DU. A summary of channel estimation methods and their corresponding look-up table(s) is provided in Table 10.
Option 1-2-C: O-DU provides “partial channel estimation details” to O-RU, using which O-RU calculates full channel estimation details. In this option, the “DM-RS channel estimation method information” includes the channel estimation method and its partial details as listed in Table 11 to Table 14. O-RU calculates the channel estimation details, such as MMSE matrix, CE granularity in the frequency-domain, and window size, based on the channel estimation method and its received details from O-DU. The details of these channel estimation calculations can be found in Embodiment 7: Channel Estimation.
Option 1-2-D: O-DU provides “full channel estimation details” to O-RU. In this option, the “DM-RS channel estimation method information” includes the channel estimation method and its full details as listed in Table 15. O-DU transfers the channel estimation details by either indicating bits in a bitfield for a particular parameter or directly transferring IQ/integer values of relevant parameters to O-RU.
As shown in
Option 1-3-A: “Combining/digital beamforming matrix enhancement information” includes SRS channel estimates of the scheduled or paired UEs/layers. Later in Step 1-6, O-RU combines SRS and DM-RS channel estimates using weighting coefficients. Options for the weighting coefficients of Option 1-3-A:
Option 1-3-A-1: O-RU either uses pre-configured weighting coefficients or determines the weighting coefficients based on other parameters.
Option 1-3-A-2: O-DU provides the weighting coefficients to 0-RU as part of “combining/digital beamforming matrix enhancement information”.
Option 1-3-B: “Combining/digital beamforming matrix enhancement information” includes preliminary combining/digital beamforming matrix elements corresponding to the scheduled or paired UEs/layers. This preliminary combining/digital beamforming matrix is calculated at O-DU prior to slot n using SRS channel estimates. Later in Step 1-6, O-RU combines its own calculated combining/digital beamforming matrix elements with the preliminary combining/digital beamforming matrix using weighting coefficients. Options for the weighting coefficients of Option 1-3-B:
Option 1-3-B-1: O-RU either uses pre-configured weighting coefficients or determines the weighting coefficients based on other parameters.
Option 1-3-B-2: O-DU provides the weighting coefficients to O-RU as part of “combining/digital beamforming matrix enhancement information”.
As shown in
Option 1-4-A: (Step 1-2 is not executed): O-RU uses a pre-configured channel estimation method and its details.
Option 1-4-B: (Step 1-2 is executed): O-RU uses “DM-RS channel estimation method information” provided by O-DU to obtain the channel estimation method and its details.
As shown in
An example of a mapping between the bitfield contents and corresponding combining/digital beamforming matrix calculation methods is provided in Table 16.
As shown in
Option 1-6-A (Step 1-3 is not executed): O-RU calculates the “combining/digital beamforming matrix” based on the estimated DM-RS channel(s).
Option 1-6-B (Step 1-3 with Option 1-3-A is executed): O-RU combines SRS channel estimates provided by O-DU with its own estimated DM-RS channel estimates using weighting coefficients. O-RU calculates the “combining/digital beamforming matrix” based on the combined DM-RS and SRS channel estimates.
Option 1-6-C (Step 1-3 with Option 1-3-B is executed): O-RU calculates the “combining/digital beamforming matrix” based on the estimated DM-RS channel(s), and using weighting coefficients, combines it with the SRS-based preliminary combining/digital beamforming matrix received from O-DU.
As shown in
As shown in
Depending on the type of information provided by O-DU to O-RU, different implementations are described, as illustrated in
Embodiment 2 describes the processing steps for UL functional split with O-RU Cat C2, in which O-RU performs the “DM-RS based Channel Estimation” and “Equalization” in the UL.
As shown in
As shown in
Option 2-1-A: “DM-RS configuration parameters” of the scheduled or paired UEs/layers. O-RU generates the “DM-RS sequence” using the “DM-RS configuration parameters” received from O-DU. These “DM-RS configuration parameters” are summarized in Table 4. Also, the details of the “DM-RS sequence” generation and mapping using the “DM-RS configuration parameters” are provided in Embodiment 6: 5G NR DM-RS.
Option 2-1-B: “DM-RS sequence” of the scheduled or paired UEs/layers.
As shown in
The details of the above channel estimation methods are described in Embodiment 7: Channel Estimation below. The “details of the indicated channel estimation method” depends on the indicated channel estimation method and may include any of the following:
Indication of the “channel estimation method” and/or “details of the indicated channel estimation method” can be done based on one of the following options:
Option 2-2-A: O-DU indicates the “channel estimation method” to the O-RU using a particular bitfield in the signaling from O-DU to O-RU. O-RU determines the channel estimation details, including parameters and processing steps based on its own implementation. An example of a mapping between the bitfield content and corresponding channel estimation methods is provided in Table 5.
Option 2-2-B: O-DU indicates “channel estimation method” and “details of the indicated channel estimation method” to the O-RU using particular bitfields in the signaling from O-DU to O-RU. O-RU uses a table, such as Table 5, to find the channel estimation method and uses other look-up tables, such as Table 6 to Table 9, to obtain the channel estimation details using these indications from O-DU. A summary of channel estimation methods and their corresponding look-up table(s) is provided in Table 10.
Option 2-2-C: O-DU provides “partial channel estimation details” to O-RU, using which O-RU calculates full channel estimation details. In this option, the “DM-RS channel estimation method information” includes the channel estimation method and its partial details as listed in Table 11 to Table 14. O-RU calculates the channel estimation details, such as MMSE matrix, CE granularity in the frequency-domain, and window size, based on the channel estimation method and its received details from O-DU. The details of these channel estimation calculations can be found in Embodiment 7: Channel Estimation.
Option 2-2-D: O-DU provides “full channel estimation details” to O-RU. In this option, the “DM-RS channel estimation method information” includes the channel estimation method and its full details as listed in Table 15. O-DU transfers the channel estimation details by either indicating bits in a bitfield for a particular parameter or directly transferring IQ/integer values of relevant parameters to O-RU.
As shown in
Option 2-3-A: (Step 2-2 is not executed): O-RU uses a pre-configured channel estimation method and its details.
Option 2-3-B: (Step 1-2 is executed): O-RU uses “DM-RS channel estimation method information” provided by O-DU to obtain the channel estimation method and its details.
As shown in
The details of the channel equalization are described in Embodiment 8: Channel Equalization below. The “additional channel equalization information” of the demodulation procedure during the log-likelihood ratio (LLR) calculation can be provided in several ways. Based on the indicated “additional channel equalization information”, which may be a part of the “channel equalization information”, the supplementary demodulation information is calculated. Exemplary “additional channel equalization information” options are listed below:
Indication of the “channel equalization method” and “additional channel equalization information” can be done based on selecting one or both of the following options below:
Option 2-4-A: O-DU indicates the “channel equalization method” to the O-RU using a particular bitfield in the signaling from O-DU to O-RU. An example of a mapping between the bitfield content and corresponding channel equalization methods is provided in Table 17.
Option 2-4-B: O-DU indicates “additional channel equalization information” to the O-RU using particular bitfields in the signaling from O-DU to O-RU. Examples of mapping between the bitfield content and corresponding “additional channel equalization information”, which determines the “supplementary demodulation information” calculation method, are provided in Table 18 and Table 19.
As shown in
Option 2-5-A: (Step 2-4-A is not executed): O-RU uses a pre-configured channel equalization method and its details.
Option 2-5-B: (Step 2-4-A is executed): O-RU uses “channel equalization method information” provided by O-DU to obtain the channel equalization method and its details.
As shown in
Option 2-6-A: (Step 2-4-B is not executed): O-RU uses a pre-configured “supplementary demodulation information” calculation method.
Option 2-6-B: (Step 2-4-B is executed): O-RU uses the “supplementary demodulation information” calculation method provided by O-DU.
As shown in
Embodiment 3 describes the processing steps for UL functional split with O-RU Cat C3, in which O-RU performs the “DM-RS based Channel Estimation”, “Equalization”, and “Demodulation” in the UL.
In this embodiment, O-RU performs the following steps to process the received signal:
As shown in
Option 3-1-A: “DM-RS configuration parameters” of the scheduled or paired UEs/layers. O-RU generates the “DM-RS sequence” using the “DM-RS configuration parameters” received from O-DU. These “DM-RS configuration parameters” are summarized in Table 4. Also, the details of the “DM-RS sequence” generation and mapping using the “DM-RS configuration parameters” are provided in Embodiment 6: 5G NR DM-RS.
Option 3-1-B: “DM-RS sequence” of the scheduled or paired UEs/layers.
As shown in
The details of the above channel estimation methods are described in Embodiment 7: Channel Estimation. The “details of the indicated channel estimation method” depends on the indicated channel estimation method and may include any of the following:
Indication of the “channel estimation method” and/or “details of the indicated channel estimation method” can be done based on one of the following options:
Option 3-2-A: O-DU indicates the “channel estimation method” to the O-RU using a particular bitfield in the signaling from O-DU to O-RU. O-RU determines the channel estimation details, including parameters and processing steps based on its own implementation. An example of a mapping between the bitfield content and corresponding channel estimation methods is provided in Table 5.
Option 3-2-B: O-DU indicates “channel estimation method” and “details of the indicated channel estimation method” to the O-RU using particular bitfields in the signaling from O-DU to O-RU. O-RU uses a table, such as Table 5, to find the channel estimation method and uses other look-up tables, such as Table 6 to Table 9, to obtain the channel estimation details using these indications from O-DU. A summary of channel estimation methods and their corresponding look-up table(s) is provided in Table 10.
Option 3-2-C: O-DU provides “partial channel estimation details” to O-RU, using which O-RU calculates full channel estimation details. In this option, the “DM-RS channel estimation method information” includes the channel estimation method and its partial details as listed in Table 11 to Table 14. O-RU calculates the channel estimation details, such as MMSE matrix, CE granularity in the frequency-domain, and window size, based on the channel estimation method and its received details from O-DU. The details of these channel estimation calculations can be found in Embodiment 7: Channel Estimation.
Option 3-2-D: O-DU provides “full channel estimation details” to O-RU. In this option, the “DM-RS channel estimation method information” includes the channel estimation method and its full details as listed in Table 15. O-DU transfers the channel estimation details by either indicating bits in a bitfield for a particular parameter or directly transferring IQ/integer values of relevant parameters to O-RU.
As shown in
Option 3-3-A: (Step 3-2 is not executed): O-RU uses a pre-configured channel estimation method and its details.
Option 3-3-B: (Step 3-2 is executed): O-RU uses “DM-RS channel estimation method information” provided by O-DU to obtain the channel estimation method and its details.
As shown in
As shown in
Option 3-5-A: (Step 3-4 is not executed): O-RU uses a pre-configured channel equalization method and its details.
Option 3-5-B: (Step 3-4 is executed): O-RU uses “channel equalization method information” provided by O-DU to obtain the channel equalization method and its details.
As shown in
As shown in
In one embodiment, the O-RU exposes its ability to support category C1, category C2, or category C3 to the SMO or the O-RU controller in the O-RAN module-cap.yang. One example of the RU-supported-category leaf under the RU-capabilities structure in the O-RAN-module-cap.yang can be shown as follows:
The O-RU exposes its ability to support DM-RS-based channel estimation locally and the dependent features by supporting of the feature DMRS-BASED-CHANNEL-ESTIMATION-SUPPORTED in o-ran-module-cap.yang module.
If feature DMRS-BASED-CHANNEL-ESTIMATION-SUPPORTED is true, the O-RU reports its support for the following dependent features:
One example implementation in the o-ran-module-cap.yang can be described as follows.
Another example implementation in the o-ran-uplane-conf.yang can be described as follows, where dependent features can be marked as read-write for capabilities reporting and configurations
C-plane Embodiment: If dmrs-local-generation and dynamic-dmrs-based-channel-estimation-configuration-type features are supported by the O-RU and configured by the SMO/O-RU controller, the O-DU appends the UL C-plane message sent in the DL to the O-RU with the following section extension and includes the DM-RS configuration parameters to enable the O-RU to locally generate the DM-RS signal.
ExtType=21: Section description for DMRS sequence generatio
This section extension applies to DMRS signal generation. This enables the O-DU to provide the DMRS configuration parameters at the O-RU to generate the DMRS signal locally. This section extension applies to section types 1 and 3.
Description of the remaining fields in the above section extension follows the same definitions in Table 2 above.
M-plane Embodiment for static DMRS sequence generation: If dmrs-local-generation and static-dmrs-based-channel-estimation-configuration-type features are supported by the O-RU and configured by the SMO/O-RU controller, the SMO/O-RU controller statically configures the DM-RS parameters to enable the O-RU to locally generate the DM-RS signal and execute DMRS-based channel estimation.
C-plane Embodiment: If dmrs-sequence-reception and dynamic-dmrs-based-channel-estimation-configuration-type features are supported by the O-RU and configured by the SMO/O-RU controller, the O-DU appends the UL C-plane message sent in the DL to the O-RU with the following section extension and includes the DM-RS sequence to enable the O-RU to do DMRS-based channel estimation.
ExtType=22: Section description for DMRS sequence transmission
This section extension applies to DMRS signal transmission. This enables the O-DU to provide the DMRS sequence needed at the O-RU to execute DMRS-based channel estimation. This section extension applies to section types 1 and 3.
If dynamic-dmrs-based-channel-estimation-configuration-type and real-time-channel-estimation-method features are supported by the O-RU and configured by the SMO/O-RU controller, the O-DU configures the O-RU in real time with the channel estimation method and needed parameters to enable the O-RU to do channel estimation locally.
ExtType=23: Section description for channel estimation
This section extension applies to channel estimation at the O-RU. This enables the O-DU to provide the channel estimation technique and relevant configuration parameters needed at the O-RU to locally perform channel estimation. This section extension applies to section types 1 and 3.
In 5G NR, DM-RS is used for channel estimation as part of coherent demodulation of the physical channels (i.e., PUSCH/PDSCH). It is designed to support various use cases, channel conditions, and deployment scenarios. DM-RS configuration depends on other PHY layer parameters such as the resource allocation type and utilized waveform (i.e., CP-OFDM or DFT-s-OFDM), and these are detailed in the following parts. In order for O-RU to be able to perform DM-RS channel estimation, either DM-RS sequence or DM-RS configuration parameters, which are required to generate DM-RS sequence at O-RU, need to be provided by O-DU to O-RU. This part explains the latter option and lists the parameters with detailed instructions for DM-RS sequence generation, precoding, and mapping.
A length-31 Gold sequence is used as the DM-RS sequence for CP-OFDM (i.e., when transform precoding is disabled). This pseudo-random sequence, c(n), with a length of MPN is defined as follows:
where n = 0,1, ..., MPN - 1 and Nc is 1600. The first m-sequence, x1(n), can be initialized with x1(0) = 1 and x1(n) = 0 for n = 1,2, ...,30. Also, the second m-sequence, x2(n) is initialized according to
Using the pseudo-random sequence, the DM-RS sequence for CP-OFDM, r(n), can be generated as follows:
The initialization of the random sequence is performed as follows:
where I is the OFDM symbol number within the slot and
is the slot number within a frame. Furthermore,
denotes the DM-RS scrambling identity and nSCID represents DM-RS scrambling initialization parameter.
The Zadoff-Chu sequence is used as the DM-RS sequence for DFT-s-OFDM (i.e., when transform precoding is enabled). This low-PAPR sequence,
(n), is defined by a cyclic shift a of a base sequence r̅u,v (n) as follows:
where
denotes the length of the sequence. Also,
represents the number of subcarriers in a resource block. Multiple orthogonal sequences can be generated from a base sequence through different values of a and δ. The base sequences, r̅u,v (n), are divided into groups, where u ∈ {0,1,2, ..., 29} specifies the group number and ν ∈ {0,1} represents the base sequence number within the group.
The base sequence definition depends on the sequence length as follows:
The length NZC is defined as the largest prime number, which is less than the base sequence length MZC.
Using the Zadoff-Chu sequence, the DM-RS sequence for DFT-s-OFDM, r(n), can be generated as follows:
where
The low-PAPR sequence parameters δ is given as 1 and the cyclic shift, α, is set to 0 for a PUSCH transmission. The sequence group number, u, equals to
where
denotes the DM-RS scrambling identity. Furthermore, the values of fgh and the base sequence number, ν, are given as follows:
The orthogonality of the DM-RS sequences that are allocated to the different antenna ports is achieved by multiplexing in time, frequency, and code domains. To realize this, the DM-RS sequence is multiplied with the orthogonal cover codes in time and frequency domains (i.e., wt(k) and wf(k) respectively). Also, shifts in the frequency domain are included to support more antenna ports. The DM-RS sequence, r(m) is mapped to an intermediate quantity
as follows:
where wf(k′), wt(l′), and Δ are defined by Tables 6.4.1.1.3-1 and Tables 6.4.1.1.3-2 in TS .. Also, p̃j represents the antenna port for transmission layer j. However, DFT-s-OFDM supports only single-layer transmission, and j is equal to 0 when transform-precoding is enabled.
DM-RS configuration type determines the DM-RS density in the frequency domain, which also impacts the available number of orthogonal sequences. DM-RS configuration Type 1 has a higher density in the frequency domain and provides a more accurate channel estimation. On the other hand, DM-RS configuration Type 2 has a lower density in the frequency domain, and it allows supporting more antenna ports. Type 1 can support up to 8 antenna ports, whereas Type 2 can support up to 12 antenna ports. These are illustrated in
The reference point for k depends on the waveform type. For CP-OFDM (i.e., when transform precoding is disabled), k is subcarrier 0 in the common resource block. On the other hand, for DFT-s-OFDM (i.e., when transform precoding is enabled), k is subcarrier 0 of the lowest-numbered RB of the scheduled PUSCH allocation.
The reference point of l and the position of the first DM-RS symbol (lo) depends on the PUSCH mapping type and activation of the frequency hopping scheme. For PUSCH Mapping Type A, l is defined relative to the start of the slot if frequency hopping is disabled, and it is defined relative to the start of each hop if frequency hopping is enabled. l0 is indicated by dmrs — TypeA — Position parameter, and it refers to either the third or fourth OFDM symbol in a slot. On the other hand, for PUSCH Mapping Type B, l is defined relative to the start of the scheduled PUSCH resources if frequency hopping is disabled, and it is defined relative to the start of each hop if frequency hopping is enabled. In this case, l0 equals to 0; in other words, it refers to the first symbol in the scheduled data symbol duration. Exemplary DM-RS mappings are illustrated in
The position(s) of the DM-RS symbols is provided by l̃ parameter and it depends on the PUSCH mapping type, activation of the frequency hopping scheme, and the number of additional DM-RS symbols. Up to 4 DM-RS symbols (i.e., up to 3 additional positions) in a slot are possible for high mobility scenarios. Exemplary DM-RS allocations with additional positions are presented in
In the following step, the intermediate quantity,
, is precoded with W, multiplied with the amplitude scaling factor
and mapped to the physical resources as follows:
where and βDMRS is given by Table 6.2.2-1 in TS 38.214. The required βDMRS value is signaled with the number of DM-RS CDM groups without data parameter. Also, W is given by Table 6.3.1.5-1 to Table 6.3.1.5.-7 in TS 38.211. The required precoding matrix is signaled through the TPMI index parameter.
The DM-RS sequence can either be sent from O-DU to O-RU or the DM-RS configuration parameters, which are required to re-generate the DM-RS sequence in the O-RU, can be transmitted from O-DU to O-RU.
In an OFDM system with N subcarriers, the received signal in the frequency domain can be expressed as follows:
where Y is received signal vector, X is a diagonal matrix where each diagonal element represents a reference signal (such as DM-RS) in subcarrier k, H is the channel vector, and Z is the noise vector.
where ki represents the subcarrier index, lj represents the symbol index, rm is the receiver antenna index, and tn is the DM-RS transmission port index. If the DM-RS belongs to code-division multiplexing (CDM) group of size L with an orthogonal cover code (OCC), the channel estimate can be de-spreaded as follows:
where cij(tn) represents the code in symbol i, subcarrier j, and DM-RS transmission port n. Also, L is the size of the RE group where the 2-length OCC is applied. When single DM-RS is used, L is equal to 2, and when double DM-RS is used, L is equal to 4.
The channel estimates, Ĥ′LS(DMRSg,rm,tn), in a given RB can be interpolated in the time and frequency domains to obtain channel estimates over data symbol. Also, these channel estimates can be averaged for an initial channel estimate as follows:
where nRB represents the RB index, ns shows the slot index, DMRSg denotes the DM-RS RE group index, and G is the total number of DM-RS RE groups in a given RB. Depending on the complexity and channel conditions, the channel estimates belong to multiple RBs can be averaged as well.
where
denotes the
and R is the cross-correlation matrix. The expression inside the parenthesis (i.e., RHH +
I) is called as the MMSE matrix and can be denoted with Ø. The elements of RHĤ
Assuming an exponentially decaying multipath power delay profile, the frequency-domain correlation can be stated as follows:
where τRMS denotes the RMS delay spread of the channel, and Δf is the subcarrier spacing. Also, assuming a fading channel with Jake’s spectrum, the time-domain correlation can be expressed as follows:
where fD
After calculating W, the MMSE channel estimation can be performed as below:
Considering the DM-RS structure, the MMSE channel estimation in the vicinity of a group of REs, where a 2-length OCC is applied, can be calculated as follows:
where ki, and lj, show the subcarrier and symbol indices of the DM-RS RE group. Also, these channel estimates can be averaged for an initial channel estimate as follows:
where nRB represents the RB index, ns shows the slot index, I and J represents the total number of subcarriers and symbols that are used in this averaging operation. Depending on the complexity and channel conditions, the channel estimates belong to multiple RBs can be averaged as well.
Accordingly, the noise power can be estimated as follows:
where the mean operation is defined over all receiver antennas and DM-RS REs.
A DFT-based channel estimation is illustrated in
Afterward, a window (e.g., rectangular) can be applied to filter the noise outside the maximum delay of the channel as follows:
where w denotes the window length. To improve the performance further, the channel taps that are below a certain threshold value are set to 0. This threshold value can be calculated by the average energy of the channel taps outside the window. The thresholding operation is performed as follows:
where Pn
by taking the average power of all taps that are set to 0. Following this operation, the time-domain estimate is transformed to the frequency domain by the DFT operation.
These channel estimates can be averaged for an initial channel estimate as follows:
Depending on the complexity and channel conditions, the channel estimates belong to multiple RBs can be averaged as well.
A DCT based channel estimation is illustrated in
The performance of channel estimation (LS or MMSE) can be improved with a DCT-based channel estimation technique by filtering the noise outside the maximum delay of the channel, similar to the DFT-based estimation. However, the DCT-based estimation provides better power concentration and lower aliasing error than the DFT-based interpolation. Considering the complexity of the system, the LS channel estimate (ĤLS) can be used as an initial estimate, and it can be converted to the transform domain with a DCT operation as follows:
Afterward, a window (e.g., rectangular) can be applied to filter the noise outside the maximum delay of the channel as follows:
where w denotes the window length. To improve the performance further, the channel taps that are below a certain threshold value are set to 0. This threshold value can be calculated by the average energy of the channel taps outside the window. The thresholding operation is performed as follows:
where Pn
These channel estimates can be averaged for an initial channel estimate as follows:
Depending on the complexity and channel conditions, the channel estimates belong to multiple RBs can be averaged as well, similar to the other channel estimation techniques that are described previously.
Examples of channel equalization techniques that O-RU can use include the following:
In a MIMO OFDM system, the received data symbols in the frequency domain can be expressed as follows:
where YDS represents the frequency domain data symbols, X shows the transmitted signal, H is the estimated channel response, Z is the noise, k denotes the subcarrier index, r shows the receiver antenna index, t represents the DM-RS transmission port index, Ntx is the number of DM-RS transmission ports and Nrx is the number of receiver antennas.
The MIMO ZF equalization matrix W is calculated as follows:
Correspondingly, the equalized data symbol of a user (ŶDS) can be expressed as follows:
Considering the same system model in part a, the MIMO MMSE equalization matrix W is calculated as follows:
where
represents the estimated noise power. Correspondingly, the equalized data symbol of a user (ŶDS) can be expressed as follows:
Considering an interference component in the system model that is described in part a, the received data symbols in the frequency domain can be expressed as follows:
where Q denotes the sum of interference from all interference users. The MIMO MMSE-IRC equalization matrix W is calculated as follows:
where RIpN represents the covariance matrix of interference plus noise and NDMRS denotes the number of DM-RS subcarriers in an RB. An interpolation operation in the time/frequency domains is required to calculate the covariance matrix in the scheduled subcarriers. Correspondingly, the equalized data symbol of a user (ŶDS) can be expressed as follows:
The demodulation procedure requires “additional channel equalization information” during the log-likelihood ratio (LLR) calculation, and it can be provided in several ways. A universal solution, which is independent of the equalization method, can be providing one of the following options:
The equalization gain can be calculated as follows:
where y is the equalization gain, W denotes the equalization matrix, and H represents the estimated channel response. The post-processed noise plus interference power refers to the noise plus interference power that is obtained after the equalization procedure. Also, it should be noted that in the case of options (b) and (c), the equalized signal should be normalized by either equalization gain or post-processed noise plus interference power, respectively.
When the MMSE equalizer is used, only “equalization gain” can be sufficient as the “additional channel equalization information”, whereas only “post-processed noise plus interference power” is required when ZF equalizer is used. Furthermore, these values, which are used during the LLR calculation operation, can be calculated from the channel estimates and estimated DM-RS noise plus interference power as well.
In the following, we provide a brief description of example combining/digital beamforming matrix calculation methods which can be used in any of the embodiments in this disclosure. In all of these me, the UL channel between the UEs and BS is denoted as H, which is a matrix of size NRX × NUE where NRX is the number of RX antennas at BS and NUE is the number of UEs or layers being simultaneously transmitted over the same resources on the PUSCH slot. It is noted that H is an estimation of the UL channel, for example, based on SRS or DM-RS or a combination thereof or a linearly predicted UL channel based on SRS channel estimation.
Channel: the contiguous frequency range between lower and upper frequency limits.
C-plane: Control Plane: refers specifically to real-time control between O-DU and O-RU, and should not be confused with the UE’s control plane
DL: DownLink: data flow towards the radiating antenna (generally on the LLS interface)
LLS: Lower Layer Split: logical interface between O-DU and O-RU when using a lower layer (intra-PHY based) functional split.
M-Plane: Management Plane: refers to non-real-time management operations between the O-DU and the O-RU
O-CU: O-RAN Control Unit - a logical node hosting PDCP, RRC, SDAP and other control functions
O-DU: O-RAN Distributed Unit: a logical node hosting RLC/MAC/High-PHY layers based on a lower layer functional split.
O-RU: O-RAN Radio Unit: a logical node hosting Low-PHY layer and RF processing based on a lower layer functional split. This is similar to 3GPP’s “TRP” or “RRH” but more specific in including the Low-PHY layer (FFT/iFFT, PRACH extraction).
OTA: Over the Air
S-Plane: Synchronization Plane: refers to traffic between the O-RU or O-DU to a synchronization controller which is generally an IEEE 1588 Grand Master (however, Grand Master functionality may be embedded in the O-DU).
U-Plane: User Plane: refers to IQ sample data transferred between O-DU and O-RU
UL: UpLink: data flow away from the radiating antenna (generally on the LLS interface)
ORAN-WG4.MP.0-v06.00: O-RAN Fronthaul Working Group Management Plane Specification, Release 06.00.
ORAN-WG4.CUS-plane.0-v06.00: O-RAN Fronthaul Working Group Control, User, and Synchronization Plane Specification, Release 06.00.
The present application claims priority to U.S. Provisional Pat. Application No. 63/244,774, filed on Sep. 16, 2021, U.S. Provisional Pat. Application No. 63/257,811, filed on Oct. 20, 2021, and U.S. Provisional Pat. Application No. 63/300,108, filed on Jan. 17, 2022, the entirety of each of which is incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| 63244774 | Sep 2021 | US | |
| 63257811 | Oct 2021 | US | |
| 63300108 | Jan 2022 | US |
