Channel state information feedback
By transmitting different types of CSI in the FDD system and using high-precision CSI training models, the problem of difficulty in obtaining channel measurement data is solved, and the CSI feedback efficiency is improved and overhead control is achieved.
Patent Information
- Application Number
- CN201980102635.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2019-11-29
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2039-11-29
AI Technical Summary
In Frequency Division Duplex (FDD) systems, obtaining sufficient channel measurement data for machine learning model training is a major challenge, affecting the high accuracy and efficiency of channel state information (CSI) feedback.
By transmitting different types of CSI between devices, using high-precision CSI to train machine learning models, and generating parameters for CSI feedback, controlling feedback overhead, and achieving efficient CSI feedback. The specific method includes transmitting high-precision CSI and low-precision CSI in the FDD system, and using high-precision CSI to train the model to obtain generation and reconstruction parameters, to improve CSI feedback efficiency.
The CSI feedback efficiency is significantly improved, and the CSI training efficiency is improved in LTE 3D UMa scenarios through simulation results, reducing the overall overhead of CSI feedback.
Smart Images

Figure CN114747250B_ABST
Abstract
Description
Technical Field
[0001] Example embodiments of the present disclosure relate generally to the field of communications, and in particular to devices, methods, apparatuses, and computer-readable storage media for feedback of channel state information (CSI). Background Art
[0002] For the 3rd Generation Partnership Project (3GPP) Release 17 (Rel-17), further enhancements to Multiple Input Multiple Output (MIMO) have been agreed. Knowledge of channel characteristics will facilitate enhanced MIMO or massive MIMO. For time division duplex (TDD) systems, base stations such as new radio NodeBs (or gNBs) can acquire downlink (DL) channel characteristics through uplink (UL) sounding signals from different user equipments (UEs) based on the reciprocity of UL and DL channels.
[0003] Frequency division duplex (FDD) systems do not have the reciprocity property of UL and DL channels. In FDD systems, downlink channel state information (CSI) is reported by the UE to the gNB via an uplink feedback channel. In terms of system efficiency, a trade-off needs to be made between high accuracy of CSI feedback and reasonable overhead. Machine learning (ML) techniques can be used to enhance CSI feedback. For example, by leveraging ML techniques, a multi-layer neural network (NN) is first trained using channel measurement data, and then the NN is used for compression and recovery of CSI during CSI feedback. However, obtaining sufficient channel measurement data for NN training is a big challenge. Summary of the invention
[0004] In general, example embodiments of the present disclosure provide an apparatus, method, device, and computer-readable storage medium for feedback of CSI.
[0005] In a first aspect, a first device is provided, the first device comprising at least one processor and at least one memory comprising computer program code. The at least one memory and the computer program code are configured to, together with the at least one processor, cause the first device to: determine a type of channel state information to be fed back. The first device is caused to generate the channel state information of the type, and then send the generated channel state information to a second device.
[0006] In a second aspect, a second device is provided, the second device comprising at least one processor and at least one memory comprising computer program code. The at least one memory and the computer program code are configured to, together with the at least one processor, cause the second device to: receive channel state information from the first device. If the type of the received channel state information includes high-precision channel state information, the second device is caused to train a machine learning model by using at least the received channel state information to obtain at least a first set of parameters for the generation of the channel state information and at least a second set of parameters for the reconstruction of the channel state information. The second device is also caused to send the first set of parameters to the first device.
[0007] In a third aspect, a method is provided, in which a first device determines a type of channel state information to be fed back, and then generates the channel state information of the type and sends the generated channel state information to a second device.
[0008] In a fourth aspect, a method is provided. In the method, a second device receives channel state information from a first device. If the type of the received channel state information includes high-precision channel state information, the second device trains a machine learning model by using at least the received channel state information to obtain at least a first set of parameters for generating the channel state information and at least a second set of parameters for reconstructing the channel state information. The second device then sends the first set of parameters to the first device.
[0009] In a fifth aspect, an apparatus is provided, the apparatus comprising means for performing the method according to the third aspect or the fourth aspect.
[0010] In a sixth aspect, a computer-readable storage medium comprising program instructions stored thereon is provided. The instructions, when executed by a processor of a device, cause the device to perform the method according to the third aspect or the fourth aspect.
[0011] It should be understood that the summary is not intended to identify the key or essential features of the exemplary embodiments of the present disclosure, nor is it intended to be used to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] Some example embodiments will now be described with reference to the accompanying drawings, in which:
[0013] Figure 1 An example architecture for machine learning based CSI feedback is shown;
[0014] Figure 2 An example environment is shown in which example embodiments of the present disclosure may be implemented;
[0015] Figure 3An example process between a first device and a second device according to some example embodiments of the present disclosure is shown;
[0016] Figure 4 shows an example configuration for feedback of two types of CSI according to some example embodiments of the present disclosure;
[0017] Figure 5 An example process between a first device and a second device according to some other example embodiments of the present disclosure is shown;
[0018] Figure 6 A flowchart illustrating an example method according to some example embodiments of the present disclosure;
[0019] Figure 7 A flowchart illustrating an example method according to some other example embodiments of the present disclosure; and
[0020] Figure 8 A simplified block diagram of a device suitable for implementing an example embodiment of the present disclosure is shown.
[0021] Throughout the drawings, the same or similar reference numerals refer to the same or similar elements. DETAILED DESCRIPTION
[0022] The principle of the present disclosure will now be described with reference to some example embodiments. It should be understood that these example embodiments are described only to illustrate and help those skilled in the art understand and implement the present disclosure, and do not represent any limitation on the scope of the present disclosure. The disclosure described herein can be implemented in various other ways except the way described below.
[0023] In the following description and claims, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
[0024] As used herein, the term "terminal device" or "user equipment" (UE) refers to any terminal device capable of wireless communication with each other or with a base station. Communication may involve the transmission and / or reception of wireless signals using electromagnetic signals, radio waves, infrared signals, and / or other types of signals suitable for transmitting information over the air. In some example embodiments, the UE may be configured to transmit and / or receive information without direct human-computer interaction. For example, when triggered by an internal or external event, or in response to a request from the network side, the UE may transmit information to the base station according to a predetermined schedule.
[0025] Examples of UE include, but are not limited to, smartphones, wireless-enabled tablets, laptop embedded devices (LEEs), laptop mounted devices (LMEs), wireless client equipment (CPEs), sensors, metering devices, personal wearable devices such as watches, and / or vehicles capable of communication. For the purposes of discussion, some example embodiments will be described with reference to UEs as examples of terminal devices, and the terms "terminal device" and "user equipment" (UE) may be used interchangeably in the context of the present disclosure. A UE may also correspond to the mobile terminating (MT) portion of an integrated access and backhaul (IAB) node (also referred to as a relay node).
[0026] As used herein, the term "network device" refers to a device via which a service can be provided to a terminal device in a communication network. As an example, a network device may include a base station. As used herein, the term "base station" (BS) refers to a network device via which a service can be provided to a terminal device in a communication network. A base station may include any suitable device via which a terminal device or UE can access a communication network. Examples of base stations include relays, access points (APs), transmission points (TRPs), node Bs (NodeBs or NBs), evolved NodeBs (eNodeBs or eNBs), new radio (NR) NodeBs (gNBs), remote radio modules (RRUs), radio headers (RHs), remote radio heads (RRHs), low power nodes such as femto, micro, etc. A relay node may correspond to a distributed unit (DU) portion of an IAB node.
[0027] As used herein, the term "circuitry" may refer to one or more or all of the following:
[0028] (a) a pure hardware circuit implementation (such as an implementation using only analog and / or digital circuitry), and
[0029] (b) a combination of hardware circuitry and software such as (where applicable):
[0030] (i) a combination of analog and / or digital hardware circuits and software / firmware, and
[0031] (ii) any portion of hardware processor(s) (including digital signal processor(s)) with software, software and memory(s) that work together to enable a device (such as a mobile phone or server) to perform various functions, and
[0032] (c) Hardware circuit(s) and / or processor(s), such as microprocessor(s) or portion(s) of microprocessor(s), which requires software (e.g., firmware)
[0033] The software can be saved when no operation is needed.
[0034] This definition of circuitry applies to all uses of the term in this application, including in any claims. As another example, as used in this application, the term circuitry also covers an implementation of only a hardware circuit or processor (or multiple processors) or a portion of a hardware circuit or processor and its (or their) accompanying software and / or firmware. For example, if applicable to a particular claim element, the term circuitry also covers a baseband integrated circuit or processor integrated circuit for a mobile device, or a similar integrated circuit in a server, cellular base station, or other computing or base station.
[0035] As used herein, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The term "including" and its variations should be understood as open terms meaning "including but not limited to." The term "based on" should be understood as "based at least in part on." The terms "one embodiment" and "embodiment" should be understood as "at least one embodiment." The term "another embodiment" should be understood as "at least one other embodiment." Other definitions (explicit and implicit) may be included below.
[0036] As used herein, the terms "first", "second", etc. may be used herein to describe various elements, which should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element without departing from the scope of the exemplary embodiments. As used herein, the term "and / or" includes any one of the listed terms and all combinations of one or more.
[0037] As mentioned above, machine learning (ML) techniques are used to enhance CSI feedback. An example architecture of ML-based CSI feedback is shown in Figure 1 As shown. Figure 1 In the illustrated architecture 100, the CSI encoder 105 and the CSI decoder 110 are jointly designed using a convolutional neural network (CNN). The CSI encoder 105 can learn the conversion from the input channel samples to the compressed CSI version. The compressed CSI version is then quantized and reported to the CSI decoder 110 via a feedback channel 115. The CSI decoder 110 can learn to recover the original channel matrix from the received CSI.
[0038] For example, the CSI encoder 105 can transform a channel matrix H with N coefficients into an M-dimensional vector s as the CSI by using a CNN, where M < N. The CSI compression ratio γ is γ = M / N. The first layer of the CSI encoder 105 can be a convolutional layer that uses a kernel of dimension 3×3 to generate two feature maps. The feature maps can be reshaped and then compressed into the vector s through a fully connected layer.
[0039] On the other hand, the CSI decoder 110 can use a CNN to perform an inverse transformation from the vector s to the original channel matrix H over several layers. The first layer of the CSI decoder 110 can be a fully connected layer that provides an initial estimate of the channel matrix H. Then, the initial estimate of the channel matrix H can be fed into several "RefineNet units" that continuously refine the reconstruction and each unit includes four layers. After being refined by several RefineNet units, the channel matrix H is input into a final convolutional layer to derive the finally reconstructed channel matrix H.
[0040] After an end-to-end training process based on sufficient channel samples, two sets of parameters related to the CSI encoder 105 and the CSI decoder 110 can be jointly determined and used jointly to achieve satisfactory CSI feedback and recovery capabilities. The end-to-end NN training process can be implemented on the gNB side or the UE side based on a large number of channel samples. Due to the reciprocity of the two-way channel in TDD deployment, it is easy to obtain a large amount of NN training data. However, it is a great challenge to obtain a large amount of channel measurement data for NN training in FDD deployment.
[0041] Example embodiments of the present disclosure propose a CSI feedback mechanism to feedback different types of CSI, where one type of CSI has high precision and another type of CSI has low precision. The high-precision CSI can be obtained by any existing or future-developed method for improving the CSI feedback accuracy. The high-precision CSI is used to train a machine learning model for CSI feedback. Various types of machine learning (ML) models can be used herein, including but not limited to logistic regression models, neural networks (NNs) (such as concurrent NNs and recurrent NNs), support vector machines, random forests, etc. The trained ML model can then be used to generate low-precision CSI to improve the CSI feedback efficiency.
[0042] The training of the ML model can be implemented by a terminal device such as a UE or a network device such as a gNB. Some example embodiments of the present disclosure will be discussed in the context of the training being performed by the network device. Since the network device has more powerful NN processing capabilities and easier access to the channel measurement data of multiple UEs, the training at the network device can bring more benefits.
[0043] Figure 2 An example environment 200 is shown in which example embodiments of the present disclosure may be implemented.
[0044] The environment 200, which may be part of a communication network, includes two devices 210 and 220 communicating with each other, which are respectively referred to as a first device 210 and a second device 220. In this example, the first device 210 is implemented by a terminal device such as a UE, and the second device 220 is implemented by a network device such as a base station. The first device 210 and the second device 220 may be implemented by any other suitable device. For example, the first device 210 and the second device 220 may both be implemented by a terminal device communicating via a device-to-device (D2D) link or a side link.
[0045] The communication between the first device 210 and the second device 220 can follow any suitable communication standard or protocol that already exists or will be developed in the future, such as Universal Mobile Telecommunications System (UMTS), Long Term Evolution (LTE), Advanced LTE (LTE-A), Fifth Generation (5G) New Radio (NR), Wireless Fidelity (Wi-Fi) and Worldwide Interoperability for Microwave Access (WiMAX) standards, and adopt any suitable communication technology, including, for example, Multiple Input Multiple Output (MIMO), Orthogonal Frequency Division Multiplexing (OFDM), Time Division Multiplexing (TDM), Frequency Division Multiplexing (FDM), Code Division Multiplexing (CDM), Bluetooth, ZigBee, Machine Type Communication (MTC), Enhanced Machine Type Communication (eMTC), Enhanced Mobile Broadband (eMBB), Massive Machine Type Communication (mMTC), Ultra-Reliable Low Latency Communication (URLLC), Carrier Aggregation (CA), Dual Connectivity (DC) and New Radio Unlicensed (NR-U) technology.
[0046] The first device 210 may feed back the CSI to the second device 220. In various embodiments of the present disclosure, the first device 210 may report different types of CSI, including high-precision CSI and low-precision CSI. The second device 220 uses the high-precision CSI to train the ML model to obtain a set of parameters for CSI generation and a set of parameters for CSI reconstruction. These two sets of parameters will be used for CSI generation and reconstruction to improve CSI feedback efficiency.
[0047] It should be understood that two devices are shown in environment 200 for illustration purposes only and do not imply any limitation. In some example embodiments, more devices may feed back CSI to second device 220. Therefore, second device 220 may obtain more CSI from more devices and thereby improve training efficiency.
[0048] Figure 3 An example process 300 between the first device 210 and the second device 220 according to some example embodiments of the present disclosure is shown.
[0049] In process 300, first device 210 determines (305) the type of CSI to be fed back. The type of CSI may be high-precision CSI or low-precision CSI. In some example embodiments, feedback instances or opportunities for a particular type of CSI may be predefined or preconfigured in the network. In this example, first device 210 may determine high-precision or low-precision CSI to be fed back in a corresponding feedback instance or opportunity.
[0050] In some example embodiments, first device 210 may determine the type of CSI to be fed back based on an indication from second device 220. For example, second device 220 may send an indication for feedback of a certain type of CSI to first device 210 and other devices. The indication may be sent for each feedback instance or opportunity to indicate the type of CSI feedback to be fed back in the feedback opportunity. Figure 4 Example procedures for indicating the type of CSI to be fed back are discussed.
[0051] Figure 4 An example configuration 400 for feedback of two types of CSI according to some example embodiments of the present disclosure is shown.
[0052] The multiple boxes 405-1, ..., 405-K shown represent continuous time sequences, where K represents a positive integer. In configuration 400, boxes 405-1 and 405-9 are configured as feedback instances for high-precision CSI, and boxes 405-3, 405-5, 405-7 and 405-11 are configured as feedback instances for low-precision CSI. In this example, these two types of CSI can be fed back periodically or aperiodically at unequal intervals. Compared with low-precision CSI, the feedback interval of high-precision CSI is longer. Although high-precision CSI will bring higher feedback overhead, it occurs for a longer period. Therefore, the overall feedback overhead can be controlled within a reasonable range.
[0053] Therefore, the second device 220 needs to indicate to one or more devices including the first device 210 the type of CSI to be fed back for each feedback instance. In an example embodiment in which the first device 210 is implemented by a terminal device and the second device 220 is implemented by a network device, a new parameter or field may be defined in the downlink control information (DCI) to carry the indication. The new DCI parameter may have only 1 bit to indicate these two types of CSI, including high-precision and low-precision CSI. For example, "1" indicates feedback of high-precision CSI, and "0" indicates feedback of low-precision CSI. Alternatively or additionally, the indication may be carried in other signaling such as radio resource control (RRC) signaling.
[0054] Still reference Figure 3After the first device 210 determines (305) the type of CSI to be fed back, the first device 210 generates (310) the CSI of that type. If high-precision CSI is to be fed back, the first device 210 generates (310) CSI with higher precision. The higher precision may be higher than a predetermined threshold precision. Any suitable method may be used to improve the CSI feedback precision.
[0055] In some example embodiments, the first device 210 may quantize the channel measurement data with a number of bits higher than a threshold number of bits, and then form high-precision CSI from the quantized channel measurement data. The threshold number may be predefined or preconfigured. In some example embodiments, the threshold number may be predefined or preconfigured for the entire payload of the CSI. In some other example embodiments, where the channel measurement data is represented by one or more channel matrices, the threshold number may be predefined or preconfigured for the channel elements in the channel matrix. In this case, the first device 210 may use a number of bits greater than the threshold number of bits to quantize a larger number of channel elements or even each channel element in the channel matrix. For example, 5 to 6 bits may be used for quantization of amplitude and phase or real and imaginary parts in the channel element. Then, the first device 210 generates high-precision CSI based on the quantized channel elements.
[0056] The first device 210 then sends (315) the generated CSI to the second device 220. Thus, the second device 220 performs (320) an operation based on the type of the received CSI. If the second device 220 has sent an indication of feedback for a certain type of CSI to the first device 210, such as through DCI or RRC signaling, the second device 220 may know the type of the received CSI.
[0057] If high-precision CSI is received, the second device 220 uses at least the received CSI to train the ML model. In some example embodiments, the second device 220 uses the newly received CSI and the previously received stored CSI to further improve the training efficiency of the ML model. After training the ML model, the second device 220 may acquire a set of (referred to as the first set) parameters for the generation of CSI and a set of (referred to as the second set) parameters for the reconstruction of CSI. The parameters may be associated with any suitable characteristics of the ML model, including, for example, the architecture of the ML model and the encoding and decoding weight coefficients. Due to end-to-end training requirements, the two sets of parameters may be jointly applied to CSI generation and reconstruction (or recovery).
[0058] The high-precision CSI may also be used for other purposes by the second device 220. For example, the second device 220 may use the high-precision CSI for multi-user MIMO (MU-MIMO) scheduling and precoding to improve communication efficiency.
[0059] In some example embodiments, second device 220 may send a first set of parameters to first device 210 so that first device 210 may generate CSI (such as low-precision CSI) using the first set of parameters. The first set of parameters may be sent by second device 220 to first device 210 via RRC signaling.
[0060] If it is determined (305) that low-precision CSI is to be fed back, the first device 210 may generate the CSI using a first set of parameters. For example, in the case where the ML model includes a NN, the first device 210 may configure the NN encoder using a first set of parameters. The first device 210 may then transform and compress the channel matrix (representing channel measurement data) through the NN encoder and then quantize it into fewer channel elements to generate low-precision CSI. The low-precision CSI may be sent (315) by the first device 210 to the second device 220 with less payload than the high-precision CSI. For example, the payload may contain a number of bits that is less than a predefined or preconfigured threshold number of bits.
[0061] It is also possible to generate low-precision CSI without using an ML model. In this case, the first device 210 may quantize the channel measurement data with a number of bits lower than a threshold number of bits, and form the low-precision CSI from the quantized channel measurement data.
[0062] After receiving the low-precision CSI, the second device 220 may reconstruct the CSI based on the received CSI using a second set of parameters. For example, the second device 220 may configure the NN decoder using the second set of parameters and perform an inverse transform through the NN decoder to reconstruct or restore the original channel matrix from the received CSI. The reconstructed or restored channel matrix may be used for MU-MIMO scheduling and precoding.
[0063] In some example embodiments, the ML model may be trained by the second device for each channel scenario to obtain more than one set of parameters for CSI generation and more than one set of parameters for CSI construction to further improve CSI feedback efficiency. In these example embodiments, the first device 210 may identify channel scenarios, such as urban micro (UMi), urban macro (UMa), indoor, mixed types, etc. The first device 210 may then send an indication of the identified channel scenario to the second device 220. The indication of the channel scenario may be implemented in any suitable form. For example, the indication may be implemented in an index of the channel type. In fact, the indication may also include a specific value (e.g., 0) that is not associated with any channel type but indicates that CSI generation and reconstruction without using the ML model.
[0064] Therefore, if high-precision CSI is received, the second device 220 performs training of the ML model for a specific channel scenario. A first set of parameters and a second set of parameters are also generated for the specific channel scenario. If low-precision CSI is received, the second device 220 can build a decoder for the specific channel scenario.
[0065] In these example embodiments, second device 220 may send multiple sets of parameters for multiple ML models and an indication of an associated channel scenario so that first device 210 may generate CSI using parameters associated with the current channel scenario.
[0066] One ML model may also be trained for similar multiple channel scenarios by the second device 220. Alternatively or additionally, the second device 220 may train a general ML model for various channel scenarios.
[0067] Figure 5 FIG. 5 shows an example process 500 between the first device 210 and the second device 220 according to some other example embodiments of the present disclosure. The process 500 is as follows: Figure 3 An example implementation of the process 300 is shown. In this example, the first device 210 is implemented by a terminal device, and the second device 220 is implemented by a network device. The ML model is implemented by a NN, and the type of CSI to be fed back is indicated by a DCI parameter having 1 bit.
[0068] In process 500, the second device 220 indicates (505) different channel scenarios and related parameter groups for the NN encoder to the first device 210 through upper layer RRC signaling. Different channel scenarios can be indicated by different NN types. The second device 220 indicates (510) the type of CSI to be fed back to the first device 210 and other devices (such as other terminal devices) through a new DCI parameter with 1 bit, where "1" indicates feedback of high-precision CSI and "0" indicates feedback of low-precision CSI.
[0069] The first device 210 identifies ( 515 ) a current channel scenario, such as UMi, UMa, indoor, mixed type, etc. The first device 210 may also quantify the identified channel scenario into an index of a channel type to be indicated to the second device 220 .
[0070] The first device 210 performs (520) CSI quantization. If the DCI parameter is set to 1, the first device 21 may quantize the channel elements with more bits than usual to generate high-precision CSI. If the DCI parameter is set to 0, the first device 210 may configure the NN encoder according to the first set of parameters indicated in the RRC signaling, and then transform and compress the channel matrix (such as the DL channel matrix) into fewer channel elements through the NN encoder. The first device 210 may ultimately quantize the channel elements in the channel matrix with less payload than the high-precision CSI.
[0071] The first device 210 reports (525) the CSI and the index of the channel type as an indication of the current channel scenario to the second device 220. The second device 220 processes (530) the received CSI. When the DCI parameter is set to 1, the second device 220 can receive a high-precision CSI report and a channel type index as an indication of the channel scenario. Then, the second device 220 can perform NN training by utilizing the channel samples contained in the newly received CSI and the previously stored channel samples contained in the previously received CSI with the same or similar channel scenario.
[0072] When the DCI parameter is set to 0, the second device 220 may receive low-precision CSI and an index of the channel type as an indication of the channel scenario. The second device 220 may configure the NN decoder using a second set of parameters according to the index of the channel type, and perform an inverse transform through the NN decoder, and reconstruct or restore the original channel matrix from the received CSI. The second device 220 may also use the restored channel matrix for MU-MIMO scheduling and precoding.
[0073] The CSI feedback mechanism significantly improves the CSI feedback efficiency. For example, from the simulation results, high-precision CSI improves the end-to-end CSI training efficiency. The simulation is implemented in a long-term evolution (LTE) three-dimensional (3D) UMa scenario. In this case, a large number of channel samples are generated to form a training set. All channel samples have 16 Tx antennas and 1 Rx antenna in the 2GHz band. There are 600 active subcarriers in a 10MHz bandwidth.
[0074] Frequency domain compression is performed based on discrete Fourier transform (DFT) operations, and 32 main taps are selected for each channel sample representing the channel measurement results. All channel samples with dimensions of 32×16 in the training and validation sets are used to train the complete parameter set in the encoder and decoder of the NN architecture. The channel samples in the test set are used to evaluate the performance or ability of channel reconstruction using the trained parameter set.
[0075] The channel resilience is evaluated based on two metrics, including the normalized mean square error (NMSE) and the channel correlation coefficient ρ. NMSE can be defined as follows:
[0076]
[0077] in represents the reconstructed channel matrix, and H represents the original channel samples in the test set. The channel correlation coefficient ρ can be defined by equation (2):
[0078]
[0079] in represents the reconstructed channel vector of subcarrier n, h n represents the original channel vector of subcarrier n in the test set.
[0080] Table 1 shows the comparison of different quantization methods for channel samples in the training and validation sets used as input for NN training.
[0081] Table 1
[0082] Quantization of input samples NMSE(dB) Correlation ρ No quantification -11.46 0.96 6-bit quantization -10.08 0.95 5-bit quantization -7.82 0.92
[0083] The comparison uses 5 to 6 bits to quantize the real and imaginary parts of each channel element respectively. The baseline uses the original channel samples in the training and validation sets without quantization. The compression ratio γ for NN training is set to As shown in Table 1, the NMSE loss of 6-bit quantization of input channel samples is only 1.38 dB and the channel correlation capability is almost the same compared to no quantization of input samples. Further reduction of CSI feedback accuracy using 5-bit quantization may affect the performance of NN training.
[0084] Figure 6 600 according to some example embodiments of the present disclosure. The method 600 may be performed by Figure 2 The first device 210 shown is implemented. For ease of discussion, reference will be made to Figure 2 Method 600 is described.
[0085] In block 605, the first device 210 determines the type of CSI to be fed back. The type of CSI may include high-precision CSI and low-precision CSI. In block 610, the first device 210 generates the type of CSI. In block 615, the first device 210 sends the generated CSI to the second device 220.
[0086] In some example embodiments, first device 210 may receive an indication of feedback for the type of CSI from second device 220. Based on the indication, first device 210 may determine the type of CSI to be fed back. The type of CSI to be fed back may also be determined according to a predefined configuration for CSI feedback.
[0087] In some example embodiments, if high-precision CSI is to be fed back, first device 210 may quantize the channel measurement data with a number of bits higher than a predefined or preconfigured threshold number of bits and then form the high-precision CSI from the quantized channel measurement data.
[0088] In some example embodiments, the first device 210 may receive at least a first set of parameters for generating CSI from the second device 220. The first device 210 may also receive an indication of a channel scenario associated with the first set of parameters from the second device 220. In the case of feeding back low-precision CSI, the first device 210 may generate low-precision CSI having a number of bits lower than a threshold number of bits based on the channel measurement data using the first set of parameters. For example, the first device 210 may transform the channel measurement data into low-precision CSI using the first set of parameters. The channel measurement data may be represented in any suitable form. As an example, the channel measurement data may be represented by a plurality of channel elements in one or more channel matrices.
[0089] In some example embodiments, first device 210 may identify a channel scenario associated with channel measurement data used to generate CSI and then send an indication of the identified channel scenario to second device 220 .
[0090] Figure 7 1 shows a flow chart of an example method 700 according to some example embodiments of the present disclosure. The method 700 may be performed by Figure 2 The second device 220 shown is implemented. For the purpose of discussion, reference will be made to Figure 2 Method 700 is described.
[0091] At block 705, second device 220 receives CSI from first device 210. In some example embodiments, second device 220 may send an indication of feedback of the type of CSI to first device 210 to indicate that first device 210 feeds back the type of CSI. At block 710, if the type of received CSI includes high-precision CSI, second device 220 uses at least the received CSI to train a machine learning model to obtain at least a first set of parameters for generation of CSI and at least a second set of parameters for reconstruction of CSI.
[0092] In some example embodiments, the second device 220 may train a machine learning model using the received CSI and the previously stored CSI. In some example embodiments, the second device 220 may receive an indication of a channel scenario associated with the received CSI from the first device 210. Thus, the received CSI may be used to train a ML model associated with the channel scenario.
[0093] At block 715, second device 220 sends the first set of parameters to first device 210 so that first device 210 may generate CSI using the first set of parameters. In some example embodiments, second device 220 may send to first device 210 an indication of a channel scenario associated with the first set of parameters.
[0094] In some example embodiments, if the type of received CSI includes low-precision CSI, second device 220 may reconstruct the CSI based on the received CSI using the second set of parameters. In an example embodiment where the indication of the channel scenario is received from first device 210, second device 220 may use the set of parameters associated with the channel scenario for CSI reconstruction or recovery.
[0095] As referenced above Figures 3 to 5 All operations and features described in processes 300 and 500 are equally applicable to methods 600 and 700 and have similar effects. For simplicity, details will be omitted.
[0096] Figure 8 800 is a simplified block diagram of a device 800 suitable for implementing an example embodiment of the present disclosure. The device 800 may be configured as follows: Figure 2 The illustrated first device 210 or second device 220 is implemented at or as a part thereof.
[0097] As shown, the device 800 includes a processor 810, a memory 820 coupled to the processor 810, a communication module 830 coupled to the processor 810, and a communication interface (not shown) coupled to the communication module 830. The memory 820 stores at least a program 840. The communication module 830 is used for two-way communication, for example, via multiple antennas. The communication interface can represent any interface required for communication.
[0098] Assume that program 840 includes program instructions that, when executed by associated processor 810, enable device 800 to operate in accordance with example embodiments of the present disclosure, as referred to herein. Figures 2 to 7 The example embodiments herein may be implemented by computer software executable by the processor 810 of the device 800, or by hardware, or by a combination of software and hardware. The processor 810 may be configured to implement various example embodiments of the present disclosure.
[0099] The memory 820 may be of any type suitable for the local technology network and may be implemented using any suitable data storage technology, such as, as non-limiting examples, non-transitory computer-readable storage media, semiconductor-based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory, and removable memory. Although only one memory 820 is shown in the device 800, there may be several physically different memory modules in the device 800. The processor 810 may be of any type suitable for the local technology network and, as non-limiting examples, may include one or more of a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), and a processor based on a multi-core processor architecture. The device 800 may have multiple processors, such as a dedicated integrated circuit chip that is time-slave to a clock synchronized with a main processor.
[0100] When the device 800 acts as the first device 210 or a part of the first device 210, the processor 810 and the communication module 830 may cooperate to implement the above reference Figure 6 When the device 800 acts as the second device 220 or a part of the second device 220, the processor 810 and the communication module 830 can cooperate to implement the method 600 described above. Figure 7 The method 700 described above is Figures 2 to 7 All operations and features described are equally applicable to the device 800 and have similar effects. For simplicity, the details will be omitted.
[0101] In general, various example embodiments of the present disclosure may be implemented using hardware or dedicated circuits, software, logic, or any combination thereof. Some aspects may be implemented using hardware, while other aspects may be implemented using firmware or software that may be executed by a controller, microprocessor, or other computing device. Although various aspects of the example embodiments of the present disclosure are illustrated and described as block diagrams, flow charts, or using some other graphical representations, it should be understood that, as non-limiting examples, the blocks, devices, systems, techniques, or methods described herein may be implemented using hardware, software, firmware, dedicated circuits or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.
[0102] The present disclosure also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer executable instructions, such as instructions included in a program module, which are executed in a device on a target real or virtual processor to perform the above referenced Figures 2 to 7The processes 300 and 500 and methods 600 and 700 described herein are described herein. Typically, a program module includes a routine, program, library, object, class, component, data structure, etc. that performs a specific task or implements a specific abstract data type. In various embodiments, the functions of the program modules can be combined or split between program modules as needed. The machine executable instructions of the program modules can be executed in a local or distributed device. In a distributed device, the program modules can be located in both local and remote storage media.
[0103] The program code for executing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device so that the program code is implemented when executed by the processor or controller in the flow chart and / or the block diagram. The program code can be executed completely on the machine, partially on the machine, as an independent software package, partially on the machine and partially on a remote machine, or completely on a remote machine or server.
[0104] In the context of the present disclosure, computer program codes or related data may be carried by any suitable carrier to enable a device, apparatus or processor to perform various processes and operations as described above. Examples of carriers include signals and computer-readable media.
[0105] The computer readable medium can be a computer readable signal medium or a computer readable storage medium. The computer readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. More specific examples of computer readable storage media will include electrical connections with one or more wires, portable computer floppy disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), digital versatile disks (DVD), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0106] In addition, although operations are described in a particular order, this should not be understood as requiring such operations to be performed in the particular order shown or in sequence or performing all of the operations shown to obtain the desired result. In some cases, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be interpreted as limitations on the scope of the present disclosure, but rather as descriptions of features that may be specific to a particular example embodiment. Certain features described in the context of a separate example embodiment may also be implemented in combination in a single embodiment. On the contrary, the various features described in the context of a single embodiment may also be implemented in multiple example embodiments individually or in any suitable sub-combination.
[0107] Although the present disclosure has been described in language specific to structural features and / or methodological acts, it should be understood that the present disclosure defined in the appended claims is not necessarily limited to the specific features or acts described above. Instead, the specific features and acts described above are disclosed as example forms of implementing the claims.
[0108] Various example embodiments of the technology have been described. In addition to or in lieu of the above, the following embodiments are described. Features described in any of the following examples may be used with any other examples described herein.
[0109] In some aspects, a first device includes: at least one processor; and at least one memory, including computer program code; the at least one memory and the computer program code are configured to, together with the at least one processor, enable the first device to: determine the type of channel state information to be fed back; generate channel state information of the type; and send the generated channel state information to a second device.
[0110] In some example embodiments, the first device is caused to determine the type of channel state information to be fed back by: receiving an indication of the type of channel state information to be fed back from the second device; and determining the type of channel state information to be fed back based on the indication.
[0111] In some example embodiments, the type of channel state information includes high-precision channel state information, and the first device is caused to generate the type of channel state information by: quantizing channel measurement data using a number of bits greater than a threshold number of bits; and forming the high-precision channel state information from the quantized channel measurement data.
[0112] In some example embodiments, the first device is further caused to: receive at least a first set of parameters for generation of the channel state information from the second device.
[0113] In some example embodiments, the first device is further caused to: receive, from the second device, an indication of a channel scenario associated with the first set of parameters.
[0114] In some exemplary embodiments, the type of channel state information includes low-precision channel state information, and the first device is further caused to generate the type of channel state information by transforming the channel measurement data into low-precision channel state information having a bit number lower than a threshold bit number by using a first set of parameters.
[0115] In some example embodiments, the first device is further caused to: identify a channel scenario associated with the channel measurement data used to generate the channel state information; and send an indication of the identified channel scenario to the second device.
[0116] In some aspects, a second device includes: at least one processor; and at least one memory including computer program code; the at least one memory and the computer program code are configured to, together with the at least one processor, cause the second device to: receive channel state information from a first device; based on the type of the received channel state information including high-precision channel state information, train a machine learning model by using at least the received channel state information to obtain at least a first set of parameters for generating channel state information and at least a second set of parameters for reconstructing the channel state information; and send the first set of parameters to the first device.
[0117] In some example embodiments, the second device is further caused to send an indication of feedback of the type of channel state information to the first device.
[0118] In some example embodiments, the second device is caused to train the machine learning model by using the received channel state information and previously stored channel state information to train the machine learning model.
[0119] In some example embodiments, the second device is further caused to send an indication of a channel scenario associated with the first set of parameters to the first device.
[0120] In some example embodiments, the second device is further caused to reconstruct the channel state information based on the received channel state information using a second set of parameters according to a type of the received channel state information comprising low-precision channel state information.
[0121] In some example embodiments, the second device is further caused to: receive, from the first device, an indication of a channel scenario associated with the received channel state information.
[0122] In some aspects, a method implemented at a first device includes: determining a type of channel state information to be fed back; generating channel state information of the type; and sending the generated channel state information to a second device.
[0123] In some example embodiments, determining the type of channel state information to be fed back includes: receiving an indication of feedback of the type of channel state information from the second device; and determining the type of channel state information to be fed back based on the indication.
[0124] In some example embodiments, the type of channel state information includes high-precision channel state information, and generating the type of channel state information includes: quantizing channel measurement data with a number of bits above a threshold number of bits; and forming the high-precision channel state information from the quantized channel measurement data.
[0125] In some example embodiments, the method further comprises receiving, from the second device, at least a first set of parameters for generation of the channel state information.
[0126] In some example embodiments, the method further comprises receiving, from the second device, an indication of a channel scenario associated with the first set of parameters.
[0127] In some example embodiments, the type of channel state information includes low-precision channel state information, and generating the type of channel state information includes: transforming the channel measurement data into the low-precision channel state information having a number of bits below a threshold number of bits by using a first set of parameters.
[0128] In some example embodiments, the method further comprises: identifying a channel scenario associated with the channel measurement data used to generate the channel state information; and sending an indication of the identified channel scenario to the second device.
[0129] In some aspects, a method implemented at a second device includes: receiving channel state information from a first device; based on a type of the received channel state information including high-precision channel state information, training a machine learning model by using at least the received channel state information to obtain at least a first set of parameters for generating channel state information and at least a second set of parameters for reconstructing the channel state information; and sending the first set of parameters to the first device.
[0130] In some example embodiments, the method further comprises sending an indication of feedback of the type of channel state information to the first device.
[0131] In some example embodiments, training the machine learning model includes: using the received channel state information and previously stored channel state information to train the machine learning model.
[0132] In some example embodiments, the method further comprises sending an indication of a channel scenario associated with the first set of parameters to the first device.
[0133] In some example embodiments, the method further comprises reconstructing the channel state information based on the received channel state information using a second set of parameters according to a type of the received channel state information comprising low-precision channel state information.
[0134] In some example embodiments, the method further comprises receiving, from the first device, an indication of a channel scenario associated with the received channel state information.
[0135] In some aspects, an apparatus includes means for determining a type of channel state information to feed back; means for generating channel state information of the type; and means for sending the generated channel state information to a second device.
[0136] In some example embodiments, means for determining the type of channel state information to be fed back includes: means for receiving an indication of feedback of the type of channel state information from the second device; and means for determining the type of channel state information to be fed back based on the indication.
[0137] In some example embodiments, the type of channel state information includes high-precision channel state information, and the components for generating the type of channel state information include: components for quantizing channel measurement data using a number of bits greater than a threshold number of bits; and components for forming the high-precision channel state information from the quantized channel measurement data.
[0138] In some example embodiments, the apparatus further comprises means for receiving at least a first set of parameters for generation of the channel state information from a second device.
[0139] In some example embodiments, the apparatus further comprises means for receiving, from the second device, an indication of a channel scenario associated with the first set of parameters.
[0140] In some example embodiments, the type of channel state information includes low-precision channel state information, and the means for generating the type of channel state information includes means for transforming the channel measurement data into the low-precision channel state information having a number of bits below a threshold number of bits by using a first set of parameters.
[0141] In some example embodiments, the apparatus further comprises: means for identifying a channel scenario associated with channel measurement data used to generate the channel state information; and means for sending an indication of the identified channel scenario to the second device.
[0142] In some aspects, an apparatus includes: a component for receiving channel state information from a first device; a component for training a machine learning model to obtain at least a first set of parameters for generating channel state information and at least a second set of parameters for reconstructing the channel state information based on a type of the received channel state information including high-precision channel state information by using at least the received channel state information; and a component for sending the first set of parameters to the first device.
[0143] In some example embodiments, the apparatus further comprises means for sending an indication of feedback of the type of channel state information to the first device.
[0144] In some example embodiments, the components for training the machine learning model include: components for training the machine learning model using the received channel state information and previously stored channel state information.
[0145] In some example embodiments, the apparatus further comprises means for sending an indication of a channel scenario associated with the first set of parameters to the first device.
[0146] In some example embodiments, the apparatus further comprises means for reconstructing the channel state information based on the received channel state information using a second set of parameters in accordance with a type of the received channel state information comprising low-precision channel state information.
[0147] In some example embodiments, the apparatus further comprises means for receiving, from the first device, an indication of a channel scenario associated with the received channel state information.
[0148] In some aspects, a computer-readable storage medium includes program instructions stored thereon, which, when executed by a processor of a device, cause the device to perform methods according to some example embodiments of the present disclosure.
Claims
1. A first device for channel state information feedback, comprising: at least one processor; as well as at least one memory including computer program code; The at least one memory and the computer program code are configured to, with the at least one processor, cause the first device to: Determining the type of channel state information to be fed back; generating channel state information of the type described; as well as Sending the generated channel state information to the second device; The type of channel state information includes high-precision channel state information, and the first device is configured to generate the type of channel state information in the following manner: quantizing the channel measurement data with a number of bits above a threshold number of bits; and forming the high-precision channel state information from the quantized channel measurement data; The first device sends the high-precision channel state information to the second device, so that the second device uses at least the received channel state information to train a machine learning model to obtain at least a first set of parameters for generating channel state information and at least a second set of parameters for reconstructing the channel state information; The first device is further configured to: receiving, from the second device, the at least a first set of parameters for generation of channel state information; The type of channel state information further includes low-precision channel state information, and the first device is further configured to generate the channel state information of the type in the following manner: The first device configures a multi-layer neural network (NN) encoder using the first set of parameters; and The first device transforms, compresses and quantizes the channel matrix into fewer channel elements through the NN encoder to generate low-precision channel state information; The first device sends the low-precision channel state information to the second device, so that the second device reconstructs channel state information based on the received channel state information using the second set of parameters.
2. The first device according to claim 1, wherein the first device is caused to determine the type of channel state information to be fed back by: receiving an indication of feedback of the type of channel state information from the second device; and The type of channel state information to be fed back is determined based on the indication.
3. The first device according to claim 1, wherein the first device is further configured to: An indication of a channel scenario associated with the first set of parameters is received from the second device.
4. The first device according to claim 1, wherein the first device is further configured to: identifying a channel scenario associated with channel measurement data used to generate the channel state information; and An indication of the identified channel scenario is sent to the second device.
5. A second device for channel state information feedback, comprising: at least one processor; as well as at least one memory including computer program code; The at least one memory and the computer program code are configured to, with the at least one processor, cause the second device to: receiving channel state information from a first device; According to the type of the received channel state information including high-precision channel state information, training a machine learning model by using at least the received channel state information to obtain at least a first set of parameters for generation of channel state information and at least a second set of parameters for reconstruction of channel state information; as well as sending the first set of parameters to the first device; The second device is further configured to: sending an indication of a channel scenario associated with the first set of parameters to the first device; The second device is further configured to: reconstructing channel state information based on the received channel state information using the second set of parameters according to a type of the received channel state information including low-precision channel state information; The second device is configured to train the machine learning model by: The machine learning model is trained using the received channel state information and previously stored channel state information.
6. The second device according to claim 5, wherein the second device is further configured to: An indication of feedback of the type of channel state information is sent to the first device.
7. The second device according to claim 5, wherein the second device is further configured to: An indication of a channel scenario associated with the received channel state information is received from the first device.
8. A method for channel state information feedback implemented at a first device, comprising: Determining the type of channel state information to be fed back; generating channel state information of the type described; as well as Sending the generated channel state information to the second device; The type of the channel state information includes high-precision channel state information, and generating the type of channel state information includes: quantizing the channel measurement data with a number of bits above a threshold number of bits; and forming the high-precision channel state information from the quantized channel measurement data; Sending the high-precision channel state information to the second device, so that the second device uses at least the received channel state information to train a machine learning model to obtain at least a first set of parameters for generating channel state information and at least a second set of parameters for reconstructing the channel state information; The method further comprises: receiving, from the second device, the at least a first set of parameters for generation of channel state information; The type of channel state information further includes low-precision channel state information, and generating the type of channel state information includes: Using the first set of parameters to configure a multi-layer neural network (NN) encoder; and transforming and compressing the channel matrix and quantizing it into fewer channel elements by the NN encoder to generate low-precision channel state information; The low-precision channel state information is sent to the second device, so that the second device reconstructs channel state information based on the received channel state information using the second set of parameters.
9. The method of claim 8, wherein determining the type of channel state information to be fed back comprises: receiving an indication of feedback of the type of channel state information from the second device; as well as The type of channel state information to be fed back is determined based on the indication.
10. The method according to claim 8, further comprising: An indication of a channel scenario associated with the first set of parameters is received from the second device.
11. The method according to claim 8, further comprising: identifying a channel scenario associated with channel measurement data used to generate the channel state information; as well as An indication of the identified channel scenario is sent to the second device.
12. A method for channel state information feedback implemented at a second device, comprising: receiving channel state information from a first device; According to the type of the received channel state information including high-precision channel state information, training a machine learning model by using at least the received channel state information to obtain at least a first set of parameters for generation of channel state information and at least a second set of parameters for reconstruction of channel state information; as well as sending the first set of parameters to the first device; The method further comprises: sending an indication of a channel scenario associated with the first set of parameters to the first device; reconstructing channel state information based on the received channel state information using the second set of parameters according to a type of the received channel state information including low-precision channel state information; Wherein training the machine learning model comprises: The machine learning model is trained using the received channel state information and previously stored channel state information.
13. The method according to claim 12, further comprising: An indication of feedback of the type of channel state information is sent to the first device.
14. The method according to claim 12, further comprising: An indication of a channel scenario associated with the received channel state information is received from the first device.
15. A device for channel state information feedback, comprising: means for determining the type of channel state information to be fed back; means for generating channel state information of said type; as well as A component for sending the generated channel state information to a second device; wherein the type of the channel state information comprises high-precision channel state information, and the means for generating the channel state information of the type is further configured to: quantize channel measurement data with a number of bits higher than a threshold number of bits, and form the high-precision channel state information from the quantized channel measurement data; The component for sending the generated channel state information to the second device is further used to: send the high-precision channel state information to the second device, so that the second device at least uses the received channel state information to train a machine learning model to obtain at least a first set of parameters for generating channel state information and at least a second set of parameters for reconstructing channel state information; The device also includes: means for receiving, from said second device, said at least a first set of parameters for generation of channel state information; The type of channel state information further comprises low-precision channel state information, and the means for generating the channel state information of the type is further used to: configure a multi-layer neural network (NN) encoder using the first set of parameters; and transform and compress and quantize a channel matrix into fewer channel elements by the NN encoder to generate low-precision channel state information; The component for sending the generated channel state information to the second device is further used for sending the low-precision channel state information to the second device, so that the second device reconstructs the channel state information based on the received channel state information using a second set of parameters.
16. A device for channel state information feedback, comprising: means for receiving channel state information from a first device; means for training a machine learning model by using at least the received channel state information to obtain at least a first set of parameters for generation of channel state information and at least a second set of parameters for reconstruction of channel state information according to the type of the received channel state information including high-precision channel state information; as well as means for sending the first set of parameters to the first device; The device is also used for: sending an indication of a channel scenario associated with the first set of parameters to the first device; reconstructing channel state information based on the received channel state information using the second set of parameters according to a type of the received channel state information including low-precision channel state information; Wherein training the machine learning model comprises: The machine learning model is trained using the received channel state information and previously stored channel state information.
17. A computer-readable storage medium comprising program instructions stored thereon, which instructions, when executed by a processor of a device, cause the device to perform the method according to any one of claims 8 to 11.
18. A computer-readable storage medium comprising program instructions stored thereon, which instructions, when executed by a processor of a device, cause the device to perform the method according to any one of claims 12 to 14.
