Channel State Information Feedback Compression in New Radio Transmission

Abstract

The present disclosure includes a method for managing channel state information feedback compression in a user equipment (UE), the method including: receiving at least one reference signal; estimating a channel state based on the at least one reference signal; reducing, using a first autoencoder in a user equipment (UE), a dimension of the estimated channel state such that a low-dimensional channel state is generated; calculating, using a second autoencoder and an autoencoder decoder in the UE, an error measure based on a difference between a UE channel state input and a UE channel state output; and transmitting at least one of a channel state information bit stream or at least one message, wherein at least one of the channel state information bit stream or the at least one message is based on at least one of a value indicative of the low-dimensional channel state or the error measure.

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 63 / 373,985, filed August 30, 2022, entitled "CHANNEL STATE INFORMATION FEEDBACK COMPRESSION IN NEW RADIO TRANSMISSIONS," the entire contents of which are incorporated herein by reference.

[0002] FIELD OF THE DISCLOSURE Apparatus and methods according to the present disclosure relate generally to communications, and more particularly to methods, systems, and devices for channel state information (CSI) reporting in wireless networks. [Background technology]

[0003] CSI provides information about the condition of the wireless channel between a base station (gNB) and a user equipment (UE). This information is important for optimizing the performance of wireless networks for various tasks such as beamforming, power control, and scheduling. CSI may include one or more parameters such as channel quality, signal strength, interference level, and other relevant metrics for efficient communication between the base station and the UE.

[0004] In 3GPP New Radio (NR), there can be challenges in accurately describing the channel state. For example, communicating the channel state when the channel state has high dimensionality can require a large number of bits. Summary of the Invention

[0005] In view of the above, embodiments of the present disclosure address the shortcomings of existing systems by providing apparatuses, systems and methods for CSI reporting in wireless networks. [Means for solving the problem]

[0006] According to some embodiments of the present disclosure, a method for managing CSI feedback compression in a UE is provided, the method including: receiving at least one reference signal; estimating a channel state based on the at least one reference signal; reducing, using a first autoencoder in the UE, a dimension of the estimated channel state to generate a low-dimensional channel state; calculating, using a second autoencoder and an autoencoder decoder in the UE, an error measure based on a difference between a UE channel state input and a UE channel state output; and transmitting at least one of a CSI bitstream or at least one message, wherein the at least one of the CSI bitstream or the at least one message is based on at least one of the value indicative of the low-dimensional channel state or the error measure.

[0007] According to some embodiments of the present disclosure, an apparatus for CSI feedback compression is provided, the apparatus including: a memory storing instructions; and executing the instructions stored in the memory to receive at least one reference signal, estimate a channel state based on the at least one reference signal, reduce a dimension of the estimated channel state using a first autoencoder in a UE to generate a low-dimensional channel state, calculate an error measure based on a difference between a UE channel state input and a UE channel state output using a second autoencoder and an autoencoder decoder in the UE, and generate a CSI bitstream or a low-dimensional CSI bitstream. and a processor configured to transmit at least one of the at least one message, wherein at least one of the CSI bitstream or the at least one message is based on at least one of a value indicative of a low-dimensional channel state or an error measurement.

[0008] According to some embodiments of the present disclosure, a non-transitory computer-readable medium storing instructions executable by one or more processors of an apparatus for managing CSI feedback compression to implement a method is provided, the method including receiving at least one reference signal, estimating a channel state based on the at least one reference signal, reducing, using a first autoencoder in the UE, a dimension of the estimated channel state to generate a low-dimensional channel state, calculating, using a second autoencoder and an autoencoder decoder in the UE, an error measure based on a difference between a UE channel state input and a UE channel state output, and transmitting a CSI bitstream or at least one message, wherein at least one of the CSI bitstream or the at least one message is based on at least one of a value indicative of the low-dimensional channel state or the error measure.

[0009] According to some embodiments of the present disclosure, a method for managing CSI feedback compression is provided. The method includes transmitting at least one reference signal, receiving at least one of a CSI bit stream or at least one message in response to the transmitted at least one reference signal, determining at least one of a value indicative of a low-dimensional channel condition or an error measure based on at least one of the CSI bit stream or the at least one message, calculating at least one of updated weights or updated parameters for an autoencoder decoder in a base station based on the at least one of the estimated channel condition or the value indicative of the error measure, where the estimated channel condition is based on the low-dimensional channel condition, and updating the calculated at least one of the weights or parameters of the autoencoder decoder in the base station based on the at least one of the updated weights or updated parameters.

[0010] According to some embodiments of the present disclosure, an apparatus for managing CSI feedback is provided, including: a memory storing instructions; and a processor configured to execute the instructions stored in the memory to transmit at least one reference signal, receive at least one of a CSI bitstream or at least one message in response to the transmitted at least one reference signal, determine at least one of a value indicative of a low-dimensional channel condition or an error measure based on at least one of the CSI bitstream or the at least one message, calculate at least one updated weight or updated parameter for an autoencoder decoder in a base station based on the at least one value indicative of the estimated channel or error measure, where the estimated channel condition is based on the low-dimensional channel condition, and update at least one weight or parameter of the autoencoder decoder in the base station based on the calculated at least one of the updated weight or updated parameter.

[0011] According to some embodiments of the present disclosure, a non-transitory computer-readable medium storing instructions executable by one or more processors of an apparatus for managing CSI feedback compression to implement a method includes transmitting at least one reference signal, receiving at least one of a CSI bitstream or at least one message in response to the transmitted at least one reference signal, and determining at least one of a value indicative of a low-dimensional channel state or an error measure based on at least one of the CSI bitstream or the at least one message. and calculating at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on at least one of an estimated channel condition or a value indicative of an error measure, wherein the estimated channel condition is based on a low-dimensional channel condition; and updating at least one of the weights or parameters of the autoencoder decoder in the base station based on the calculated at least one of the updated weights or updated parameters.

[0012] According to some embodiments of the present disclosure, a method for managing CSI feedback compression in a UE is provided. The method includes receiving at least one reference signal, estimating a channel state based on the at least one reference signal, reducing, using a first autoencoder in the UE, a dimension of the estimated channel state to generate a low-dimensional channel state, calculating, using a second autoencoder and an autoencoder decoder in the UE, an error measure based on a difference between a UE channel state input and a UE channel state output, calculating at least one of updated weights or updated parameters for the first autoencoder in the UE based on a value indicative of the error measure, updating at least one of weights or parameters of the first autoencoder in the UE based on the calculated at least one of the updated weights or updated parameters, generating at least one of a CSI bitstream or at least one message based on at least one of the low-dimensional channel state or the at least one of the updated weights or updated parameters for the first autoencoder in the UE, and transmitting the at least one of the CSI bitstream or the at least one message.

[0013] According to some embodiments of the present disclosure, a method for managing CSI feedback compression in a base station is provided, the method including: transmitting at least one reference signal; receiving at least one of a CSI bitstream or at least one message in response to the transmitted at least one reference signal; determining at least one of updated weights, updated parameters, or a structure for at least one of a first autoencoder in the UE, a second autoencoder in the UE, or an autoencoder decoder in the UE; calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station based on the at least one of the updated weights or updated parameters for the autoencoder encoder in the UE; and updating at least one of weights, parameters, or a structure of the autoencoder decoder in the base station based on the calculated at least one of the updated weights, updated parameters, or structure.

[0014] Some embodiments of the present disclosure include receiving, from a base station, at least weights, parameters, or structures of at least a first autoencoder encoder, a second autoencoder encoder, and / or an autoencoder decoder in a UE, and updating at least weights, parameters, or structures of at least the first autoencoder encoder, the second autoencoder encoder, and / or an autoencoder decoder in the UE using the weights, parameters, or structures received from the base station.

[0015] Some embodiments of the present disclosure include calculating and / or updating, at the base station, at least one of weights, parameters, or structures of the first autoencoder decoder, the second autoencoder decoder, and / or at least one of the autoencoder encoders, and transmitting the weights, parameters, or structures to the UE. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a schematic diagram illustrating an example of CSI feedback for downlink (DL) and uplink (UL) transmissions. [Figure 2] FIG. 1 is a schematic diagram illustrating an example of the basic architecture of an autoencoder. [Figure 3] Schematic diagram showing an example of the basic architecture of an autoencoder and the associated training on a dataset to determine the encoder and decoder weights / parameters. [Figure 4] FIG. 1 is a schematic diagram illustrating an example of a sandwich structure of an autoencoder in which the encoder structure and the decoder structure are symmetrical. [Figure 5] FIG. 1 is a schematic diagram illustrating an example of a general architecture using an autoencoder for CSI compression. [Figure 6] FIG. 1 is a schematic diagram illustrating an example of updating the autoencoder encoder and autoencoder decoder structures, autoencoder encoder and autoencoder decoder weights / parameters, and CSI bitstream generation and decoding schemes. [Figure 7] 1 is a flowchart illustrating an example method for managing CSI feedback compression, e.g., at a UE, in accordance with an embodiment of the present disclosure. [Figure 8] 10 is a flowchart illustrating another example method for managing CSI feedback compression, eg, at a base station, in accordance with an embodiment of the present disclosure. [Figure 9] FIG. 9 is a schematic diagram illustrating an example of updating the structure of the autoencoder and the autoencoder decoder, the weights / parameters of the autoencoder and the autoencoder decoder, and the CSI bitstream generation and decoding scheme. [Figure 10] 10 is a flowchart illustrating a further example method for managing CSI feedback compression, e.g., at a UE, in accordance with an embodiment of the present disclosure. [Figure 11]10 is a flowchart illustrating an additional example method for managing CSI feedback compression at a gNB, in accordance with an embodiment of the present disclosure. [Figure 12] FIG. 1 is a block diagram of a device according to some embodiments of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0017] Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. The following description refers to the accompanying drawings, in which like numbers in different drawings represent the same or similar elements, unless otherwise noted. The implementations described in the following description of exemplary embodiments do not represent all implementations according to the present disclosure. Rather, they are merely examples of systems, apparatus, and methods according to aspects related to the present disclosure as set forth in the appended claims.

[0018] In 3GPP NR, the allocation, configuration, and transmission scheme of radio resources for the Uu interface (the interface between the UE and the gNB) are determined by the base station. To optimize performance in various aspects on the Uu interface, the base station may dynamically adjust the allocation, configuration, and transmission scheme of radio resources based on the current channel conditions on the Uu interface. For this purpose, the base station may be aware of the current channel conditions on the Uu interface for both DL and UL transmissions. The dimensionality of the channel conditions reported in CSI may be high, making it difficult to accurately describe the channel conditions. For example, communicating the channel conditions when the dimensionality of the channel conditions is high may require a large number of bits. To address the problem of accurately describing the channel conditions using CSI feedback, one approach may include using fewer bits to report the channel conditions through CSI, for example, using CSI compression.

[0019] A second problem in reporting channel conditions via CSI feedback may involve providing a more accurate (i.e., higher precision) representation of the channel conditions, in other words, improving the accuracy of the CSI reports. The embodiments provide approaches to address the issue of CSI compression. Additionally, at least some of the disclosed embodiments provide approaches to more accurately represent channel conditions.

[0020] FIG. 1 is a schematic diagram illustrating CSI feedback for DL ​​and UL transmissions. In 3GPP NR, CSI feedback is only used for DL ​​transmissions. In DL transmissions, the base station 120 is the transmitter (TX) of downlink traffic, and the UE 110 is the receiver (RX). To estimate the channel condition in downlink transmissions, in 3GPP NR, the base station 120 uses a physical downlink shared channel (PDSCH) and a physical downlink control channel (PDCCH), as illustrated in FIG. 1. The UE 110 transmits a reference signal RS 130 on a control channel (CQI), where the sequence and format of the reference signal RS 130 are provided in the 3GPP NR standard. In response to receiving the reference signal RS 130, the UE 110 may estimate a downlink channel condition based on the distortion of the received reference signal RS 130. The UE 110 may then provide information associated with the downlink channel condition to the base station 120 (i.e., provide CSI feedback) by sending a CSI 140 to the base station 120. The content of the channel condition information 140 in 3GPP NR includes a Channel Quality Indicator (CQI), a Precoding Matrix Indicator (PMI), a CSI-RS Resource Indicator (CRI), and a Synchronization Signal and Physical Broadcast Channel Resource Block Indicator (SSBRI). The CSI may include a Layer Indicator (LI), a Layer Indicator (LI), and a Rank Indicator (RI). For uplink transmission, the UE 110 is the transmitter (TX) of uplink traffic, and the base station 120 is the receiver (RX). To estimate the channel condition in the uplink transmission, the UE 110 may also transmit a reference signal RS 150 (RS on a physical uplink shared channel (PUSCH)). Upon receiving the reference signal RS 150, the base station 120 may estimate the uplink channel condition based on the distortion of the reference signal RS 150. For the uplink, the base station 120 may not transmit CSI to the UE 110.

[0021] FIG. 2 is a schematic diagram illustrating an example of the basic architecture of an autoencoder. An autoencoder may be a type of artificial neural network that may be used for dimensionality reduction. An autoencoder may include an encoder and a decoder. The encoder may take input data and compress it into a low-dimensional representation. The decoder may then take the compressed representation and reconstruct the original input data. According to some disclosed embodiments, an autoencoder may be used to reduce the dimensionality of CSI. An autoencoder composed of an encoder and a decoder may be associated with an unsupervised learning method based on a deep neural network. As illustrated in FIG. 2, both the encoder 210 and the decoder 220 may be composed of multiple neural nodes 230 belonging to different operational stages, and neural nodes 230 belonging to the same operational stage form a layer. There may be multiple links connecting neural nodes in different layers, and each link may be multiplied by a weight / parameter 240. In each neural node 230, a nonlinear operation (e.g., a sigmoid function, a step function, etc.) may be applied to all inputs (multiplied by corresponding weights / parameters 240) to generate an output. Before the output of a neural node 230 becomes the input of a neural node 230 in the next layer, the output may be multiplied by a corresponding weight / parameter 240 .

[0022] The input of encoder 210 may be high-dimensional, and the goal of encoder 210 may be to reduce the dimensionality of the input. Thus, the output of encoder 210 may be a low-dimensional representation of the input of encoder 210, and the output of encoder 210 then becomes the input of decoder 220. The goal of decoder 220 may be to reconstruct the low-dimensional input as a high-dimensional output. It is contemplated that the output of decoder 220 and the input of encoder 210 may be identical.

[0023] FIG. 3 is a schematic diagram illustrating an example of a basic autoencoder architecture and associated training on a dataset 330 to determine the weights / parameters of the encoder 310 and decoder 320. An autoencoder's dataset 330 may refer to a collection of input data used to train and evaluate an autoencoder model. In the context of an autoencoder, the dataset 330 may include a set of unlabeled examples or samples that the autoencoder may learn to reconstruct. To make the encoder 310's input and the decoder 320's output identical, the weights / parameters in the encoder 310 and decoder 320 may be updated appropriately for any given input of the autoencoder. If the weights / parameters in the encoder 310 and decoder 320 are inappropriate based on the autoencoder's current input, a specific difference or error 340 between the encoder 310's input and the decoder 320's output may be measured with a specific form of loss function (e.g., mean squared error, root mean squared error, normalized mean squared error, etc.), as shown in FIG. 3. The weights / parameters in the encoder 310 and decoder 320 may then be updated based on the current difference or error 340. It should be understood that the input of the encoder 310 and the output of the decoder 320 may be identical, nearly identical, similar, or any other correlation between the two resulting from training or operational use. While the goal may be for them to be identical, it should be understood that in some cases they may not match. For example, the difference or error 340 may represent the deviation between the input of the encoder 310 and the output of the decoder 320.

[0024] To make the encoder's input and decoder's output identical, the weights and parameters of both the encoder and decoder may be updated appropriately for any input to the autoencoder. Many autoencoder structures have been proposed for applications in which autoencoders may be used. The autoencoder's structure may be specified based on several characteristics, including, but not limited to, the number of layers and the number of neural nodes in each layer. Returning to the example shown in FIG. 2, the illustrated encoder 210 may have two layers, with the second layer having two neural nodes. The autoencoder's structure may have different types of layers. For example, the autoencoder may include pooling layers, convolutional layers, etc. The connection architecture between layers may include partial connections, full connections, or any other connections between nodes and / or layers, as determined based on the application. Furthermore, the autoencoder's structure may include different types of connections, including forward connections, backward connections, etc.

[0025] In an autoencoder structure, neural nodes may implement different types of operations. For example, a neural node may implement a sigmoid function, a step function, or other similar data manipulation operations. The structure of the autoencoder may determine the number of weights / parameters in the encoder and decoder. For example, as shown in Figure 2, the encoder may implement a W based on the structure determined for the autoencoder. n The weights / parameters 240 may have weights 240. The autoencoder structure may determine the type of weights / parameters 240 in the implementation. For example, the weights / parameters 240 may include real numbers, complex numbers, integers, floating point numbers, or any other data type that may be determined by the structure of the autoencoder. The autoencoder structure selected for an application may determine which loss function is used, such as difference or error. It may be determined whether the signal can be used to measure

[0026] FIG. 4 is a schematic diagram illustrating an example of a sandwich structure of an autoencoder in which the encoder and decoder structures are symmetric. In the sandwich structure, the input layer 410 and the output layer 420 are positioned as the outermost layers, and the hidden layer 430 is sandwiched between them. In the outermost structure of the autoencoder shown in FIG. 4, the encoder and decoder structures are symmetric in several aspects. First, the number of layers in the encoder and the decoder may be the same, with the exception that one additional layer known as the code or bottleneck layer 440 may be present in the encoder. Second, the number of neural nodes in the encoder's input layer 410 may be the same as the number of neural nodes in the decoder's output layer 420. Third, the number of encoder hidden layers 430 and decoder hidden layers 435 (i.e., a layer may be a hidden layer if it is not the input layer 410, output layer 420, or code / bottleneck layer 440) may be the same in the encoder and the decoder. Fourth, the vector of weights / parameters between the input layer 410 and the first hidden layer 450 in the encoder may be the transpose of the vector of weights / parameters between the output layer 420 and the first hidden layer 480 in the decoder. The vector of weights / parameters between the first hidden layer 450 and the second hidden layer 460 in the encoder may be the transpose of the vector of weights / parameters between the first hidden layer 480 and the second hidden layer 470 in the decoder, etc. As shown in Figure 4, W1 T is the transpose of W1, and W2 T is the transpose of W2, etc.

[0027] In the sandwich structure shown in Figure 4, the encoder and decoder of an autoencoder may synchronize their structures and weights / parameters. The design of an autoencoder may involve a trade-off between structural complexity and reconstruction performance. A sandwich structure may be designed to significantly reduce the dimensionality of the encoder input and perfectly reconstruct the compressed representation at the decoder output. However, overfitting issues may arise (i.e., a change in the autoencoder input may increase the difference / error between the autoencoder input and output). Alternatively, a structure may be designed with a simpler structure, resulting in a larger difference / error between the autoencoder input and output.

[0028] The following describes the use of an autoencoder to perform CSI compression. A general architecture for applying an autoencoder to CSI compression in 3GPP NR is shown in FIG. 5. First, a base station 520 (denoted as gNB) sends a reference signal RS 530 (on a PDSCH / PDCCH) to a UE 510. Upon receiving the reference signal RS 530, the UE 510 estimates a channel state (denoted as H′). An encoder of the autoencoder (denoted as AE encoder) may be located in the UE 510, and a decoder of the autoencoder (denoted as AE decoder) may be located in the base station 520. After estimating H′ in the UE 510, the AE encoder may compress H′ into low-dimensional information. Then, a quantization and encoding process may be performed to transfer the low-dimensional information into a bitstream of a certain number of bits, which may represent a new low-dimensional version of the CSI 540. A transceiver (TX / RX) module in the UE 510 may send the low-dimensional version of the CSI to the base station 520. After receiving the low-dimensional version of the CS 540 via the TX / RX module in the base station 520, the decoding and dequantization process in the base station 520 may proceed to transfer the bitstream to the low-dimensional information. Finally, the AE decoder in the base station 520 may reconstruct the low-dimensional information into a channel state (denoted as H'). In FIG. 5, the AE encoder and quantization & encoding in the UE 510 may perform CSI bitstream generation, and the AE decoder and decoding & dequantization in the base station 520 may perform CSI decoding to reconstruct a high-dimensional version of the CSI to determine the channel state.

[0029] There are two potentially significant challenges with this general architecture: First, the traditional applications where autoencoders are used (e.g., video / audio processing, principle specific) In some cases (e.g., performance analysis), the autoencoder including both the AE encoder and the AE decoder may be located on the same machine. In the case of CSI compression, the architecture may have information about both the input and output of the autoencoder because both reside within the architecture's devices (e.g., on the UE 510 and the base station 520). The architecture may therefore determine a measurement of the difference / error between the autoencoder's input and output. As a result, the architecture may need to update the weights / parameters of both the AE encoder and the AE decoder based on the measured difference / error. However, when an autoencoder is applied to CSI feedback compression, the UE 510 (AE encoder) may have information related to the channel state directly, and the base station 520 (AE decoder) may have information related to the channel state based only on the results of channel reconstruction. As a result, the UE 510 and the base station 520 may not have complete information about the difference / error between the autoencoder's input and output. Therefore, due to incomplete information about the channel state, the UE 510 and the base station 520 may not be able to update the weights / parameters of the AE encoder and the AE decoder.

[0030] Second, the weights / parameters of the AE encoder and AE decoder may be updated periodically, but the difference / error may not be significantly reduced. In such a situation, the current structure of the AE encoder and AE decoder may not be appropriate. In this case, the structure of the AE encoder and AE decoder may be modified. The derivation of an appropriate structure of the AE encoder and AE decoder may be based on the difference / error between the input and output of the autoencoder. If the UE 510 and the base station 520 do not have information about the difference / error between the input and output of the autoencoder, both the UE 510 and the base station 520 may not be able to determine improved weights / parameters and structures of the AE encoder and AE decoder.

[0031] To at least partially address these two challenges in applying autoencoders to CSI compression in 3GPP NR, disclosed embodiments provide methods and apparatuses that may include two stages: an initial stage and a second stage. The purpose of the initial stage may be two-fold. First, both the UE and the base station may commonly understand the supported structure of the AE encoder and AE decoder. Second, both the UE and the base station may commonly understand the adopted CSI bitstream generation and decoding scheme. In the second stage, once the structure of the UE's AE encoder and the base station's AE decoder have been adopted by the UE and the base station, at least some disclosed embodiments perform CSI compression and update the weights / parameters and structure accordingly.

[0032] In the initial stage, one or more aspects of the present application may be applicable. A specific number of autoencoder encoder / decoder structures for CSI feedback may be provided. The base station may communicate the supported autoencoder encoder / decoder structures (all or part of the structures) to the UE. This information may be conveyed, for example, by a Master Information Block (MIB) or a System Information Block (SIB) in NR. The UE may communicate the supported autoencoder encoder / decoder structures (all or part of the standard structures) to the base station. This information may be conveyed, for example, by UE capability information in radio resource control (RRC) signaling in NR.

[0033] In some embodiments, the structure of the adopted autoencoder / decoder is determined by the network (base station or core network (CN)), which may signal this configuration to the UE. In other embodiments, the structure of the adopted autoencoder / decoder is determined by the UE. The UE may signal this configuration to the network, which may then allocate radio resources for the UE to send this configuration. In other embodiments, the structure of the employed autoencoder / decoder may be specified in the standard.

[0034] In some embodiments, the employed weights / parameters of both the AE encoder and the AE decoder may be determined by the network, and the network may communicate this configuration to the UE. In other embodiments, the employed weights / parameters of both the AE encoder and the AE decoder may be determined by the UE, and the UE may communicate this configuration to the network. In this case, the network may allocate radio resources for the UE to send this configuration. In other embodiments, the employed weights / parameters of both the AE encoder and the AE decoder may be specified in a standard.

[0035] A specific number of CSI bitstream generation / decoding schemes may be provided. The base station may communicate the supported CSI bitstream generation / decoding schemes (all or part of the schemes) to the UE. This information may be conveyed, for example, by a master information block (MIB) or a system information block (SIB) in NR. The UE may communicate the supported CSI bitstream generation / decoding schemes (all or part of the schemes) to the base station. This information may be conveyed, for example, by UE capability information in RRC signaling in NR.

[0036] In some embodiments, the adopted CSI bitstream generation / decoding scheme may be determined by the network, and the network may communicate this configuration to the UE. In other embodiments, the adopted CSI bitstream generation / decoding scheme may be determined by the UE, and the UE may communicate this configuration to the network. In this case, the network may allocate radio resources for the UE to send the configuration. In other embodiments, the adopted CSI bitstream generation / decoding scheme may be specified.

[0037] According to some disclosed embodiments, in an alternative for second-stage design, after the initial stage, both the UE and the base station may have information associated with the adopted structure of the AE encoder and AE decoder, the weights / parameters of the AE encoder and AE decoder, and the CSI bitstream generation and decoding scheme. For second-stage design, the methods and apparatuses described and illustrated herein may provide two alternatives. FIG. 6 shows a block diagram of an example implementation of the first alternative for the second stage, in which the UE may provide updated weights / parameters to the base station in implementing CSI compression. The first alternative for the second stage may require several steps. First, the base station 620 (i.e., the gNB) may send a reference signal RS 630 in one or more downlink transmissions. After receiving the reference signal RS 630, the UE 610 may estimate the channel state H′. The AE encoder 650 in the UE 610 may then reduce the dimension of the estimated channel state to generate reduced-dimensional information.

[0038] The CSI bitstream generator 655 in the UE 610 may forward the low-dimensional information to a CSI bitstream generator having a specific number of bits. The TX / RX module 680 in the UE 610 may then transmit the bitstream to the base station 620 as a low-dimensional version of the CSI 660. Upon receiving the bitstream containing the low-dimensional version of the CSI 660, the TX / RX module 685 in the base station 620 may provide the bitstream as an input to a CSI bitstream decoder 675. The CSI bitstream decoder 675 in the base station 620 may convert the bitstream into low-dimensional information. The AE decoder 670 at the base station 620 may then reconstruct the channel state. The autoencoder 640 in the UE 610 may measure the difference / error between its own input and output based on a specific loss function. Based on the difference / error, the autoencoder 640 in the UE 610 The encoder 640 may update weights / parameters for both the AE encoder 650 in the UE 610 and the AE decoder 670 in the base station 620. Additionally, the autoencoder 640 in the UE 610 may select other structures for both the AE encoder 650 in the UE 610 and the AE decoder 670 in the base station 620. Accordingly, the TX / RX module in the UE 610 may send one or more of the updated weights / parameters of the AE encoder 650 in the UE 610 and / or the AE decoder 670 in the base station 620 and / or the structures of the AE encoder 650 in the UE 610 and the AE decoder 670 in the base station 620 to the base station 620.

[0039] After receiving one or more of the updated weights / parameters of the AE encoder 650 in the UE 610 and / or the AE decoder 670 in the base station 620, and / or the structure of the AE encoder 650 in the UE 610 and the AE decoder 670 in the base station 620, the base station 620 may further confirm or determine one or more of the adopted weights / parameters for the AE encoder 650 in the UE 610 and / or the AE decoder 670 in the base station 620, and / or the adopted structure of the AE encoder 650 in the UE 610 and the AE decoder 670 in the base station 620. The base station 620 may then communicate the adopted weights / parameters of the AE encoder 650 to the UE 610 to update the AE decoder 670 in the base station 620, and / or communicate the adopted structure of the AE encoder 650 to the UE 610 to update the AE decoder 670 in the base station 620. The AE decoder 670 in the base station 620 may then use the adopted weights / parameters or the structure determined by the base station 620. The AE encoder 650 in the UE 610 and the autoencoder 640 in the UE 610 may also use the updated weights / parameters or structure determined by the base station 620.

[0040] According to aspects of the disclosed embodiments of the present application, the UE may include an autoencoder (including an encoder and a decoder), an AE encoder, and a CSI bitstream generator. The base station may include an AE decoder and a CSI bitstream decoder. After the initial stage, the UE may update weights / parameters of the AE encoder in the UE and the AE decoder in the base station (using the autoencoder in the UE), and the UE may communicate the updated weights / parameters to the base station. The base station may allocate radio resources for the UE to upload the updated weights / parameters. After receiving the updated weights / parameters from the UE, the base station may further communicate or confirm information including the adopted weights / parameters to the UE.

[0041] In the UE, when the adopted weights / parameters provided by the base station are received by the UE, the AE encoder may perform CSI compression using the adopted weights / parameters provided by the base station. Then, the CSI bitstream generator may convert the compressed CSI into a bitstream having a specific number of bits. The autoencoder may further update the weights / parameters using the adopted weights / parameters provided by the base station. In the base station, the CSI bitstream decoder may convert the bitstream received from the UE to obtain compressed CSI. The AE decoder may reconstruct the CSI using the adopted weights / parameters to determine the channel state.

[0042] After the initial stage, the UE may select a different AE encoder / decoder structure from a set of AE encoder / decoder structures supported by the UE and the base station. The UE may communicate the selected AE encoder / decoder structure to the base station. The base station may allocate radio resources for the UE to communicate the selected AE encoder / decoder structure to the base station. The base station may further communicate the adopted AE encoder / decoder structure to the UE. The AE decoder in the base station may then use the adopted structure. .

[0043] 7 is a flowchart illustrating a method 700 for managing channel state information feedback compression in a UE, according to an embodiment of the disclosure described and illustrated herein. Method 700 may include an initial stage (not shown in FIG. 7) in which both the UE and the base station may be configured with information associated with the adopted structure of the AE encoder and AE decoder, weights / parameters for the AE encoder and AE decoder, and CSI bitstream generation and decoding schemes.

[0044] Referring to FIG. 7, method 700 includes step 710, in which the UE receives at least one reference signal from a base station. In embodiments, the UE may receive at least one reference signal used for various purposes, including, but not limited to, synchronization, channel estimation, and signal quality measurement. The reference signal may be transmitted by the base station and received by the UE to facilitate reliable communication therebetween (i.e., at least one reference signal may be received by the UE from the base station). According to disclosed embodiments, the at least one reference signal may be used by the UE to determine channel conditions of a Uu link between the UE and the base station.

[0045] The method 700 includes a step 720 in which the UE estimates channel conditions based on at least one reference signal. Estimating channel conditions may refer to the UE analyzing the quality and characteristics of the wireless channel between itself and the base station, including determining signal strength, interference levels, and other factors affecting communication performance. The UE may estimate the channel conditions by analyzing the at least one reference signal received from the base station.

[0046] The method 700 includes a step 730 in which the UE reduces the dimension of the estimated channel state using a first autoencoder within the UE, such that a low-dimensional channel state is generated. Reducing the dimension of the estimated channel state may refer to compressing high-dimensional data into a low-dimensional representation using an autoencoder. In some disclosed embodiments, the autoencoder within the UE may be based on a deep neural network. The deep neural network may include multiple neural nodes, each of which may include weights and parameters used to generate an output based on the neural node input. This reduction may reduce the number of bits for transmitting the estimated channel state from the UE to the base station. It should be understood that the low-dimensional representation of the estimated channel state received by the base station via CSI may be reconstructed into a high-dimensional representation using a corresponding autoencoder decoder within the base station.

[0047] The method 700 includes step 740, in which the UE calculates an error measure based on a difference between the UE channel condition input and the UE channel condition output using a second autoencoder and an autoencoder decoder within the UE. For example, the autoencoder within the UE may measure the difference or error by comparing its input to its output using a loss function. Thus, calculating the error measure may be based on a difference between the UE channel condition input and the UE channel condition output and may include performing the calculation using the loss function.

[0048] The method 700 includes a step 750 in which the UE calculates at least one of updated weights or updated parameters for an autoencoder encoder in the UE based on a value indicative of an error measure, the value indicative of a channel condition and an expected error between the autoencoder encoder of the UE and the autoencoder decoder of the base station corresponding to the value indicative of the error measure. Taking the difference into account, the UE may generate updated weights or updated parameters for its autoencoder to account for the value indicative of the error measure. The autoencoder decoder in the base station may be equipped with corresponding updated weights / parameters and structure that match the updated weights / parameters and structure for the autoencoder encoder in the UE.

[0049] The method 700 includes a step 760 in which the UE updates at least one of weights or parameters of a first autoencoder in the UE based on the calculated at least one of the updated weights or updated parameters. Based on the calculated at least one of the updated weights or updated parameters, the UE may update its first autoencoder to match the autoencoder in the base station.

[0050] The method 700 includes step 770, in which the UE generates at least one of a CSI bitstream or at least one message based on at least one of low-dimensional channel conditions or at least one of updated weights or updated parameters for a first autoencoder in the UE. The CSI bitstream may refer to a stream of bits including information about the current state of a wireless channel between the base station and the UE. The at least one message may refer to at least one packet, PDU, or other unit of data sent and / or received in a data communication. In the disclosed embodiment, the at least one message may be present in a data communication from the UE to the base station (and / or from the base station to the UE). Based on the updated weights and / or updated parameters calculated and updated based on the value indicative of the error measurement, the UE may generate either a channel condition bitstream or at least one message, which may include at least one of low-dimensional channel conditions, at least one of updated weights or updated parameters for a first autoencoder in the UE, or at least one of updated weights or updated parameters for an autoencoder decoder in the base station.

[0051] The method 700 includes a step 780 in which the UE transmits at least one of a CSI bitstream or at least one message. Once generated, the CSI may be transmitted to the base station via the CSI bitstream. In the disclosed embodiment, the stream of bits including information about the current state of the wireless channel between the base station and the UE may include information about the autoencoder components in the base station and the UE (e.g., weights, parameters, and structure information of the autoencoder components for performing CSI compression). Alternatively (or additionally), information about the autoencoder components in the base station and the UE may be conveyed in a message (e.g., at least one message) other than the CSI bitstream. For example, one or more messages may be sent in a separate data communication from the UE communicating information about the base station and the autoencoder components in the UE.

[0052] In some embodiments, the method 700 may further include the UE calculating, based on the value indicative of the error measure, at least one of updated weights or updated parameters for an autoencoder decoder in the base station, and transmitting the at least one of updated weights or updated parameters for the autoencoder decoder in the base station to the base station. Thus, the UE may communicate the updated weights and / or parameters corresponding to the autoencoder decoder in the base station to the base station.

[0053] In some embodiments, the method 700 may include a step in which the UE receives the autoencoder weights, the autoencoder parameters, the autoencoder structure, the autoencoder parameters, ... The method may further include receiving one or more of autoencoder / decoder weights, autoencoder / decoder parameters, or autoencoder / decoder structures. For example, the UE may receive the autoencoder / decoder structure and / or the autoencoder / decoder structure, e.g., included in a master information block message and / or a system information block message.

[0054] In some embodiments, method 700 may require the UE to be configured with information associated with the employed structure of the autoencoder encoder, the employed structure of the autoencoder decoder, the autoencoder encoder weights, the autoencoder decoder weights, the structure of the autoencoder encoder, the structure of the autoencoder decoder, the channel state information bitstream generation, and the decoding scheme.

[0055] In some embodiments, method 700 may further include the UE transmitting information regarding calculated weights or parameters of at least one of the first autoencoder in the UE, the second autoencoder in the UE, the autoencoder decoder in the UE, and the autoencoder decoder in the base station, or at least one of the first autoencoder in the UE, the second autoencoder in the UE, the autoencoder decoder in the UE, or the autoencoder decoder in the base station. For example, the UE may include an algorithm that enables calculation of and determination of an appropriate structure of weights / parameters in the autoencoder component to perform CSI compression, and the UE may transmit the information to the base station.

[0056] 8 is a flowchart illustrating a method 800 for managing channel state information feedback compression in a gNB, according to embodiments of the disclosure described and illustrated herein. Method 800 may involve an initial stage (not shown in FIG. 8) in which both the base station and the UE may be configured with information associated with the employed structures of the AE encoder and AE decoder, weights / parameters for the AE encoder and AE decoder, and CSI bitstream generation and decoding schemes.

[0057] 8, the method 800 includes a step 810 in which a base station transmits at least one reference signal. In an embodiment, the base station may transmit the reference signal to the UE to enable the UE to estimate channel conditions. The reference signal may provide information about channel quality, such as signal strength and interference level. The UE may use the received reference signal to optimize transmission and reception on the wireless channel.

[0058] The method 800 includes a step 820 in which the base station receives a CSI bit stream or at least one message in response to the transmitted at least one reference signal. In a disclosed embodiment, the base station may receive a stream of bits providing information regarding the current state of the channel between the base station and the UE based at least in part on estimated channel conditions, as determined by the UE. Thus, at least one reference signal may be transmitted to the UE, and a channel state information bit stream may be received from the UE.

[0059] The method 800 includes step 830, in which the base station determines at least one of updated weights, updated parameters, or a structure for at least one of a first autoencoder in the UE, a second autoencoder in the UE, or an autoencoder decoder in the UE based on at least one of the CSI bitstreams or one or more messages conveying this information. In some disclosed embodiments, the base station determines at least one of updated weights, updated parameters, or a structure for at least one of the first autoencoder in the UE, the second autoencoder in the UE, or the autoencoder decoder in the UE based on at least one of the CSI bitstreams or one or more messages separate from the CSI bitstreams conveying this information. The processor may extract information related to updated weights, updated parameters, or updated structure for the base encoder component.

[0060] Method 800 includes step 840, in which the base station calculates at least one of updated weights or updated parameters for the autoencoder decoder of the base station based on at least one of updated weights or updated parameters for the autoencoder encoder in the UE. In some disclosed embodiments, the base station may receive updated weights and / or updated parameters for the autoencoder encoder in the UE and calculate updated weights and / or updated parameters for the autoencoder decoder in the base station to correspond to the autoencoder encoder in the UE. The base station may determine associated updated weights, parameters, and / or structure of an autoencoder component used to perform CSI compression and calculate corresponding updated weights, parameters, and / or structure for the autoencoder decoder in the base station.

[0061] The method 800 includes a step 850 in which the base station updates at least one of the weights, parameters, or structure of an autoencoder decoder within the base station based on the calculated at least one of the updated weights or the updated parameters. Once the weights, parameters, or structure of the autoencoder decoder are calculated, the base station may update the weights, parameters, and / or structure of the autoencoder decoder within the base station accordingly.

[0062] The method 800 may further include receiving at least one of calculated weights, parameters, or updated structures from the UE for at least one of the autoencoder decoder in the base station, the first autoencoder in the UE, the second autoencoder in the UE, or the autoencoder decoder in the UE. In some embodiments, the updated structure may be a sandwich structure. In some embodiments, the base station may receive updated weights, parameters, and / or structures for one or more autoencoder components in the UE and the base station from the UE.

[0063] The method 800 may further include transmitting at least one of the calculated weights, parameters, or structures to the UE for at least one of the autoencoder decoder in the base station, the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, or the autoencoder decoder in the UE. The base station may transmit the weights, parameters, and / or structures back to the UE to enable updates or confirm updates of autoencoder components system-wide to ensure that the weights, parameters, and / or structures in the UE and the base station correspond.

[0064] According to some disclosed embodiments, both the UE and the base station may have information associated with the adopted structure of the AE encoder and AE decoder, the weights / parameters of the AE encoder and AE decoder, and the CSI bitstream generation and decoding scheme. Figure 9 is a schematic diagram illustrating an example implementation of such an arrangement. The schematic diagram of Figure 9 has some similarities with the schematic diagram of Figure 6, with some notable differences becoming apparent in the following description of Figure 9.

[0065] First, the base station 920 (i.e., the gNB) may send a reference signal RS 930 in one or more downlink transmissions. After receiving the reference signal RS 930, the UE 910 may estimate a channel state. Then, the AE encoder 950 in the UE 910 may reduce the dimension of the estimated channel state to generate low-dimensional information H′.

[0066] The AE encoder 950 in the UE 910 may forward the low-dimensional information H′ to a CSI bitstream generator 955 having a specific number of bits. The TX / RX module 990 in the UE 910 may then transmit the bitstream to the base station 920 as a low-dimensional version of the CSI 960. Upon receiving the bitstream including the low-dimensional version of the CSI 960, the TX / RX module 995 in the base station 920 may input the bitstream to a CSI bitstream decoder 975. The CSI bitstream decoder 975 in the base station 920 may convert the bitstream into low-dimensional information. The AE decoder 970 at the base station 920 may then reconstruct the channel state. The autoencoder 940 in the UE 910 may measure the difference and / or error between its own input and output based on a specific loss function. Based on the difference / error, the UE 910 may send a message regarding such difference / error to the base station 920. The UE 910 may send such a message periodically, or the message may be sent when the difference and / or error is greater than a certain threshold. The event for sending the message may be determined by the base station 920, which may communicate the configuration for the event to the UE 910.

[0067] When the base station 920 receives a message from the UE 910 communicating the difference / error, the autoencoder 980 in the base station 920 may update the weights / parameters of the AE encoder 950 in the UE 910 and the AE decoder 970 in the base station 920, or may select a different structure for the AE encoder 950 in the UE 910 and the AE decoder 970 in the base station 920. The AE decoder 970 in the base station 920 may use the updated weights / parameters or the updated structure. The base station 920 may communicate to the UE 910 the updated weights / parameters for the AE encoder 950 in the UE 910 and / or the AE decoder 970 in the base station, or the adopted structure of the AE encoder 950 in the UE 910 and / or the AE decoder 970 in the base station 920. The autoencoder 940 and the AE encoder 950 in the UE 910 may use the updated weights / parameters or the updated structure.

[0068] According to the disclosed embodiments, the base station may include an autoencoder (including an encoder and a decoder), an AE decoder, and a CSI bitstream decoder. The UE may include an autoencoder (including an encoder and a decoder), an AE encoder, and a CSI bitstream generator. After the initial stage, the base station may update the weights / parameters of the AE encoder in the UE and the AE decoder in the base station based on a message sent by the UE about the feasibility of the updated weights / parameters, and the base station may communicate the updated weights / parameters to the UE. The base station may allocate radio resources for transmitting the updated weights / parameters (e.g., conveyed via a control channel or a shared channel). The UE may send a message to the base station about the feasibility of the currently adopted weights / parameters (e.g., a difference / error between the estimated channel conditions and the output of the autoencoder in the UE). Such a difference / error may be derived based on a specific loss function. The message about the feasibility of the currently adopted weights / parameters may be sent periodically or when the difference / error is greater than a specific threshold. The events for sending the message may be determined by the base station, and the base station may communicate such configuration (for the events) to the UE. The base station may allocate radio resources for the UE to send the message.

[0069] In the UE, the autoencoder may compress the channel and reconstruct the channel using the adopted weights / parameters provided by the base station. The autoencoder may derive an error / difference between the channel state and the reconstructed channel state to measure the feasibility of the adopted weights / parameters communicated by the base station. The autoencoder may compress the CSI using adopted weights / parameters provided by the base station. The CSI bitstream generator may then convert the compressed CSI into a bitstream having a specific number of bits. In the base station, the autoencoder may update the weights / parameters of the autoencoder and AE encoder in the UE and the AE decoder in the base station. The CSI bitstream decoder may convert the bitstream received from the UE to obtain compressed CSI. The AE decoder may reconstruct the CSI using the updated weights / parameters.

[0070] After the initial stage, the base station may select other structures for the autoencoder and AE encoder / decoder (based on a message about the feasibility of the currently adopted weights / parameters sent by the UE) from a set of autoencoder and AE encoder / decoder structures supported by the UE and commonly by the base station, and the base station may communicate the selected structure to the UE. The base station may allocate radio resources for communicating the selected structure to the UE (e.g., via a control channel or a shared channel). The UE may send a message to the base station to communicate the feasibility of the currently adopted structure (e.g., a difference / error between the estimated channel condition and the output of the autoencoder in the UE). Such a difference / error may be derived based on a specific loss function. Such a message (about the feasibility of the currently adopted structure) may be sent periodically or when the difference / error is greater than a specific threshold. An event for sending a message is determined by the base station, and the base station may communicate the configuration (for the event) to the UE. The base station may allocate radio resources for the UE to send the message. The AE decoder in the base station may then use the adopted structure. The autoencoder and AE encoder in the UE may then use the adopted structure.

[0071] 10 is a flowchart illustrating a method 1000 for managing channel state information feedback compression at a UE, according to an embodiment of the disclosure described and illustrated herein. Method 1000 may include an initial stage (not shown in FIG. 10) in which both the UE and the base station may be configured with information associated with the employed structures of the AE encoder and AE decoder, the weights / parameters of the AE encoder and AE decoder, and the CSI bitstream generation and decoding scheme.

[0072] Referring to Figure 10, method 1000 includes step 1010, in which the UE receives at least one reference signal from a base station. In embodiments, the UE may receive at least one reference signal used for various purposes, including, but not limited to, synchronization, channel estimation, and signal quality measurement. These reference signals may be transmitted by the base station and received by the UE to facilitate reliable communication therebetween (i.e., at least one reference signal may be received from the base station). In accordance with the disclosed embodiments, the reference signal may be used by the UE to determine channel conditions of a Uu link between the UE and the base station.

[0073] The method 1000 includes step 1020, in which the UE estimates channel conditions based on at least one reference signal. Estimating channel conditions may refer to the UE analyzing the quality and characteristics of the wireless channel between itself and the base station, including determining signal strength, interference levels, and other factors affecting communication performance. The UE may estimate the channel conditions by analyzing the at least one reference signal received from the base station.

[0074] The method 1000 includes a step 1030 in which the UE reduces the dimension of the estimated channel state using a first autoencoder (i.e., an AE encoder) within the UE, such that a low-dimensional channel state is generated. The coder may be used for dimensionality reduction. In some disclosed embodiments, the input of the estimated channel state to the first autoencoder may be a high-dimensional channel state. In disclosed embodiments, the first autoencoder in the UE may be used to convert the high-dimensional estimated channel state to a low-dimensional estimated channel state. In some disclosed embodiments, the first autoencoder may be based on a deep neural network. The deep neural network may include multiple neural nodes, each of which may include weights and parameters used to generate an output based on the neural node input. This reduction may include a reduction in the number of bits for transmitting the estimated channel state from the UE to the base station.

[0075] Method 1000 includes step 1040, or the UE, calculating an error measure based on a difference between the UE channel condition input and the UE channel condition output using a second autoencoder encoder and an autoencoder decoder forming an autoencoder within the UE. For example, the autoencoder within the UE may measure a difference or error in a comparison between its input and its output based on a loss function. Thus, calculating the error measure may be based on a difference between the UE channel condition input and the UE channel condition output and may include performing the calculation using the loss function. Based on the difference or error determined from the error measure calculated by the autoencoder, the UE may update weights / parameters and / or structures of both the AE encoder at the UE and the AE decoder in the base station. Training the autoencoder within the UE may configure the autoencoder to determine a difference or error between its input and its output.

[0076] The method 1000 includes step 1050, in which the UE transmits at least one of a CSI bitstream or at least one message, where the at least one of the CSI bitstream or the at least one message is based on at least one of a value indicative of a low-dimensional channel state or an error measurement. The CSI bitstream may be transmitted to a base station. The transmitted CSI bitstream may be received by the base station and decoded by a CSI bitstream decoder in the base station to recover the original high-dimensional estimated channel state. Alternatively (or additionally), one or more messages may be transmitted to the base station separately from the CSI bitstream. The CSI bitstream or the one or more messages transmitted to the base station may include at least one of a value indicative of a low-dimensional channel state or an error measurement.

[0077] The method 1000 may further include the UE generating at least one of the CSI bitstream or the at least one message based on at least one of a value indicative of a low-dimensional channel condition or an error measurement. In the disclosed embodiment, the base station may use the low-dimensional channel condition and either or both of the difference or error measured by the autoencoder in the UE to determine changes to weights / parameters used by the base station autoencoder decoder (i.e., AE decoder) to match the autoencoder encoder in the UE (i.e., AE encoder).

[0078] In an initial stage, the method 1000 may further include the UE receiving an autoencoder encoder structure and the base station receiving an autoencoder decoder structure. In some disclosed embodiments, the autoencoder encoder and autoencoder decoder structures may be included in at least one of an MIB message or an SIB message. Thus, the autoencoder encoder and autoencoder structures in the UE and the autoencoder decoder structures in the base station may be distributed in an initial stage of configuring the architecture to operate according to the disclosed embodiments.

[0079] In some embodiments, the method 1000 may further include the UE receiving one or more of the autoencoder encoder weights, the autoencoder encoder parameters, the autoencoder encoder structure, the autoencoder decoder weights, the autoencoder decoder parameters, or the autoencoder decoder structure. For example, the UE may receive the autoencoder encoder structure and / or the autoencoder decoder structure included in, e.g., a master information block message and / or a system information block message.

[0080] In some embodiments, the method 1000 may include the UE being configured with information associated with the employed structure of the autoencoder encoder, the employed structure of the autoencoder decoder, the autoencoder encoder weights, the autoencoder decoder weights, the structure of the autoencoder encoder, the structure of the autoencoder decoder, the channel state information bitstream generation, and the decoding scheme.

[0081] In some embodiments, the method 1000 may further include the UE transmitting information regarding at least one of the first autoencoder in the UE, the second autoencoder in the UE, the autoencoder decoder in the UE, or the autoencoder decoder in the base station, or calculated weights or parameters of at least one of the first autoencoder in the UE, the second autoencoder in the UE, the autoencoder decoder in the UE, or the autoencoder decoder of the base station. For example, the UE may include an algorithm that enables calculation of weights / parameters in the autoencoder component and determination of an appropriate structure to perform CSI compression, and the UE may transmit the information to the base station.

[0082] 11 is a flowchart illustrating a method 1100 for managing CSI feedback compression in a gNB according to embodiments of the disclosure described and illustrated herein. Method 1100 may include an initial stage (not shown in FIG. 11) in which both the base station and the UE may be configured with information associated with the employed structures of the AE encoder and AE decoder, the weights / parameters of the AE encoder and AE decoder, and the CSI bitstream generation and decoding scheme.

[0083] 11, method 1100 includes step 1110 in which a base station transmits at least one reference signal. In an embodiment, the base station may transmit the reference signal to the UE to enable the UE to estimate channel conditions. The reference signal may provide information about channel quality, such as signal strength and interference level. The UE can use the received reference signal to optimize its transmission and reception.

[0084] The method 1100 includes step 1120, in which the base station receives at least one of the CSI bit streams in response to the transmitted at least one reference signal or at least one message. In the disclosed embodiments, the base station may receive a stream of bits providing information about the current state of the channel between the base station and the UE based at least in part on estimated channel conditions determined by the UE. Thus, the at least one reference signal may be transmitted to the UE, and the CSI bit stream may be received from the UE. Alternatively, the base station may receive at least one message from the UE including information described in the disclosed embodiments.

[0085] The method 1100 includes step 1130, in which the base station determines, based on the CSI bitstream or the at least one message, at least one of a value indicative of a low-dimensional channel condition or an error measure. As described and illustrated herein, the base station may use a CSI bitstream decoder to determine the low-dimensional channel condition and / or the error measure received from the UE. The base station may extract a value indicative of the received error measurement from the CSI bitstream. Alternatively, the base station may also receive a value indicative of the error measurement from one or more messages separate from the CSI bitstream. The base station may input the recovered low-dimensional channel state information to an autoencoder decoder (i.e., an AE decoder) to recover the original high-dimensional channel state determined by the UE. Thus, an estimated channel state may be determined using an autoencoder decoder in the base station to decode the low-dimensional channel state and recover the original estimated channel state. In some disclosed embodiments, the autoencoder decoder in the base station may be based on a deep neural network. A deep neural network implementing the autoencoder decoder may include multiple neural nodes, each of which may include weights and parameters used to generate an output based on the neural node input. In some embodiments, the low-dimensional channel state may be converted to a high-dimensional channel state using an autoencoder decoder in the base station.

[0086] The method 1100 includes step 1140, in which the base station calculates at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on at least one of estimated channel conditions or a value indicating an error measurement, where the estimated channel conditions may be based on a low-dimensional channel condition. In the disclosed embodiment, the UE may provide information about the channel conditions via the CSI bitstream so that differences caused by changes in the channel conditions are compensated for to correct inaccuracies in CSI compression by updating the weights / parameters and / or structure of the autoencoder components in the UE and the base station. Thus, calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station may include calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station based on a value indicating the error measurement. For example, the error measurement may be used to determine changes needed to the weights or parameters of the autoencoder decoder to match the autoencoder encoder in the UE. Thus, the error measurement may be used in an algorithm implemented in the base station to generate updated weights and / or updated parameters for the autoencoder decoder in the base station. At least one of updated weights or updated parameters for the autoencoder decoder in the base station may correspond to at least one of updated weights or updated parameters for the autoencoder encoder in the UE, and at least one of a structure of the autoencoder encoder or a structure of the autoencoder decoder may be transmitted to the UE.

[0087] Method 1100 includes step 1150, in which the base station updates at least one of weights or parameters of an autoencoder decoder in the base station based on the calculated at least one of the updated weights or updated parameters. Step 1150 performs an update of the autoencoder decoder in the base station to correspond to an update of the autoencoder encoder in the UE and to compensate for changes in channel conditions that may have caused differences or errors in CSI compression between the UE and the base station.

[0088] In some disclosed embodiments, calculating at least one of the updated weights or updated parameters may further include calculating at least one of the updated weights or updated parameters for each of a first autoencoder encoder in the UE (i.e., an AE encoder), a second autoencoder encoder in the UE, and an autoencoder decoder in the UE, where the second autoencoder encoder and the autoencoder decoder in the UE belong to the same autoencoder pair. For example, the autoencoder components in the UE and the base station require weights and parameters to be calculated to provide identical (or nearly identical) outputs for a particular input. Thus, each autoencoder component in the UE and the base station may require updated weights, updated parameters, and / or an updated structure to maintain consistent outputs. Further, calculating at least one of the updated weights or updated parameters may include updating at least one of the weights or parameters of the first autoencoder in the UE, the second autoencoder in the UE, and the autoencoder decoder in the UE. Once the weights and / or parameters are calculated, the weights and parameters can be used to update the weights and parameters of the associated autoencoder component.

[0089] In some embodiments, the method 1100 may include the base station being configured with information associated with the employed structure of the autoencoder encoder, the employed structure of the autoencoder decoder, the autoencoder encoder weights, the autoencoder decoder weights, the structure of the autoencoder encoder, the structure of the autoencoder decoder, the channel state information bitstream generation, and the decoding scheme.

[0090] FIG. 12 is a block diagram of a device 1200 according to some embodiments of the present disclosure. The device 1200 may be a network node, a roadside unit, a relay node, a base station (e.g., the base station 620 of FIG. 6 or the base station 920 of FIG. 9), or a UE (e.g., the UE 610 of FIG. 6 or the UE 910 of FIG. 9). The device 1200 may take any form, including, but not limited to, a computer system, a vehicle, a component mounted on a vehicle, a roadside unit, a laptop computer, a wireless terminal including a mobile phone, a wireless handheld device, or a wireless personal device, or any other form. The device 1200 may include an antenna 1202 used to transmit or receive electromagnetic signals to / from a base station, a UE, or another device. The antenna 1202 may include one or more antenna elements and may enable different input / output antenna configurations, such as a multiple input multiple output (MIMO) configuration, a multiple input single output (MISO) configuration, and a single input multiple output (SIMO) configuration. In some embodiments, antenna 1202 may include multiple (e.g., tens or hundreds) antenna elements and may enable multiple antenna functions such as beamforming. In some embodiments, antenna 1202 is a single antenna.

[0091] Device 1200 may include a transceiver 1204 coupled to antenna 1202. Transceiver 1204 may be a wireless transceiver in device 1200 and may communicate bidirectionally with a base station, a UE, or another device. For example, transceiver 1204 (e.g., TX / RX module 680 or TX / RX module 685 in FIG. 6 or TX / RX module 990 or TX / RX module 995 in FIG. 9) may receive / transmit wireless signals from / to a UE or gNB in ​​Uu communications. Transceiver 1204 may include a modem that modulates packets, provides the modulated packets to antenna 1202 for transmission, and demodulates packets received from antenna 1202.

[0092] The device 1200 may include memory 1206. The memory 1206 may be any type of computer-readable storage medium, including volatile or non-volatile memory devices, or a combination thereof. Computer-readable storage media include, but are not limited to, non-transitory computer storage media. Non-transitory storage media may be accessed by a general-purpose or special-purpose computer. Examples of non-transitory storage media are portable computer diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), and the like. Non-transitory media include, but are not limited to, computer programmable read-only memory (EEPROM), electrically erasable programmable ROM (EEPROM), digital versatile disk (DVD), flash memory, compact disk (CD) ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, etc. Non-transitory media are used to carry or store desired program code means (e.g., instructions and / or data structures) and may be accessed by a general-purpose or special-purpose computer or a general-purpose or special-purpose processor. In some examples, software / program code may be transmitted from a remote source (e.g., a website, server, etc.) using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, microwave, etc. In such examples, coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, microwave, etc. are within the definition of medium. Combinations of the above examples are also within the scope of computer-readable media.

[0093] The memory 1206 may store information related to the identity of the device 1200 and signals and / or data received by the antenna 1202. The memory 1206 may also store post-processed signals and / or data. The memory 1206 may also store computer-readable program instructions, mathematical models, and algorithms used for signal processing in the transceiver 1204 and calculations in the processor 1208. The memory 1206 may further store computer-readable program instructions for execution by the processor 1208 that operate the device 1200 to perform various functions described in this disclosure. In some examples, the memory 1206 may include a basic input / output system (BIOS) that controls basic hardware or software operations, such as interaction with peripheral components or devices.

[0094] The computer-readable program instructions of the present disclosure may be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-set data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​and conventional procedural programming languages. The computer-readable program instructions may be executed entirely on a computing device, as a standalone software package, or may execute partially on a first computing device and partially on a second computing device that is remote from the first computing device. In the latter scenario, the second, remote computing device may be connected to the first computing device through any type of network, including a local area network (LAN) or a wide area network (WAN).

[0095] The device 1200 may include a processor 1208, which may include a hardware device having processing capabilities. The processor 1208 may be a general-purpose processor, a digital signal processor (DSP), a central processing unit (CPU), a microcontroller, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a programmable logic device, discrete gates, transistor logic components, discrete hardware components, or other processors. The processor 1208 may include at least one of a programmable logic device (PGD), a processor architecture, ...

[0096] The device 1200 may include a global positioning system (GPS) 1210. The GPS 1210 may be used to enable location-based or other services based on the geographic location of the device 1200. The GPS 1210 receives a global navigation satellite system (GNSS) signal from a single satellite via an antenna 1202. GPS 1210 may receive a GPS (Global Positioning System) signal or multiple satellite signals to provide the geographic location of device 1200 (e.g., the coordinates of device 1200). In some embodiments, GPS 1210 may be omitted, such as when device 1200 is a base station in a substantially fixed location.

[0097] The device 1200 may include input / output (I / O) devices 1212 that may be used to communicate the results of signal processing and calculations to a user or another device. The I / O devices 1212 may include a user interface including a display and input devices for sending user commands to the processor 1208. The display may be configured to display the status of signal reception at the device 1200, data stored in the memory 1206, the status of signal processing, and calculation results. The display may include, but is not limited to, a cathode ray tube (CRT), a liquid crystal display (LCD), a light-emitting diode (LED), a gas plasma display, a touchscreen, or other image projection devices for displaying information to a user. The input devices may be any type of computer hardware equipment used to receive data and control signals from a user. Input devices include, but are not limited to, a keyboard, a mouse, a scanner, a digital camera, a joystick, a trackball, cursor direction keys, a touchscreen monitor, or an audio / video commander.

[0098] The device 1200 may further include a mechanical interface 1214 , such as an electrical bus, that connects the transceiver 1204 , the memory 1206 , the processor 1208 , the GPS 1210 , and the I / O device(s) 1212 .

[0099] In some embodiments, the device 1200 may be configured or programmed to manage channel state information feedback compression. The processor 1208 may be configured to execute instructions stored in the memory 1206 to perform at least one of the methods 700, 800, 1000, or 1110 described in connection with Figures 7, 8, 10, and 11, respectively.

[0100] In some embodiments, device 1200 may be configured or programmed to transmit CSI over a Uu interface. For example, device 1200 may be a UE in Uu interface communications, and processor 1208 may be configured to execute instructions stored in memory 1206 to determine channel state information, store the channel state information, and transmit the channel state information to a base station. Device 1200 may include any well-known elements of a UE.

[0101] In some embodiments, the device 1200 may be a UE, and the processor 1208 may execute instructions stored in the memory 1206 to receive at least one reference signal, estimate a channel state based on the at least one reference signal, reduce, using a first autoencoder in the UE, a dimension of the estimated channel state to generate a low-dimensional channel state, use a second autoencoder and an autoencoder decoder in the UE to calculate an error measure based on a difference between a UE channel state input and a UE channel state output, and transmit a CSI bitstream, the CSI bitstream being based on at least one of the values ​​indicative of the low-dimensional channel state or the error measure.

[0102] In some embodiments, the device 1200 may be a base station (e.g., a gNB), and the processor 1208 may execute instructions stored in the memory 1206 to transmit at least one reference signal, receive a channel state information bitstream in response to the transmitted at least one reference signal, determine at least one of values ​​indicative of a low-dimensional channel condition or an error measure based on the CSI bitstream, calculate at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on the at least one of the values ​​indicative of the estimated channel condition or the error measure, the estimated channel condition being based on the low-dimensional channel condition, and update at least one of the weights or parameters of the autoencoder decoder in the base station based on the calculated at least one of the updated weights or updated parameters.

[0103] One or more aspects of the present disclosure may relate to or otherwise incorporate features of 3GPP Release 18 "Research on AI / ML for the NR Air Interface."

[0104] All of the processes described herein may be fully automated via software code modules, including one or more specific computer-executable instructions executed by a computing system. The computing system may include one or more computers or processors. The code modules may be stored on any type of non-transitory computer-readable medium or other computer storage device. Some or all of the methods may be embodied in specialized computer hardware. In at least some embodiments according to the present disclosure, the weights / parameters of the AE encoder and AE decoder are updated on both the UE side and the gNB side.

[0105] Many variations beyond those described herein will be apparent from this disclosure. For example, depending on the embodiment, certain acts, events, or functions of any of the algorithms described herein may be performed in a different sequence, added, combined, or entirely omitted (e.g., not all acts or events described may be necessary to practice the algorithm). Also, in some embodiments, acts or events may be performed simultaneously rather than sequentially, for example, through multithreading, interrupt processing, or multiple processors or processor cores, or on other parallel architectures. Furthermore, different tasks or processes may be performed by different machines and / or computing systems that can function together.

[0106] The various illustrative logic blocks and modules described in connection with the embodiments disclosed herein may be implemented or performed by a machine such as a processing unit or processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The processor may be a microprocessor, but alternatively, the processor may be a controller, microcontroller, state machine, or combination thereof, etc. The processor may include electrical circuitry configured to process computer-executable instructions. In another embodiment, the processor includes an FPGA or other programmable device that performs logical operations without processing computer-executable instructions. A processor may also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. While described herein primarily with reference to digital technology, a processor may also include primarily analog components. The computing environment may include any type of computer system, including, but not limited to, a microprocessor-based computer system, a mainframe computer, a digital signal processor, a portable computing device, a device controller, or a computational engine within an appliance.

[0107] In particular, conditional language such as "can," "could," "could," or "might" be understood within the context of their common usage to convey that certain embodiments include certain features, elements, and / or steps and other embodiments do not, unless specifically stated otherwise. Thus, such conditional language is not intended to generally imply that features, elements, and / or steps are in any way required by one or more embodiments, or that one or more embodiments necessarily include logic for determining whether those features, elements, and / or steps are included in or performed in any particular embodiment, with or without user input or prompting.

[0108] It should be understood that any process descriptions, elements, or blocks of the flow diagrams described herein and / or illustrated in the accompanying figures may represent modules, segments, or portions of code that comprise one or more executable instructions for implementing specific logical functions or elements within the process. Alternative implementations are within the scope of the embodiments described herein, including those in which elements or functions are omitted, performed out of order from that shown, or described, substantially simultaneously or in reverse order, depending on the functionality involved as would be understood by one of ordinary skill in the art.

[0109] Although the exemplary embodiments provided in this disclosure refer to examples of 3GPP NR radio access technologies, the scope of the solutions in this disclosure is by no means limited to these examples. They may be applicable to, for example, 3GPP LTE or 3GPP 6G. They may also be applicable to non-3GPP radio access technologies, for example, IEEE 802.11 technologies (e.g., but not limited to, 802.11n, 802.11u, or 802.11p).

[0110] As used in this disclosure, the use of the term "or" in a list of items indicates an inclusive list. A list of items may be preceded by a phrase such as "at least one of" or "one or more of." For example, a list of at least one of A, B, or C includes A or B or C or AB (i.e., A and B), or AC or BC or ABC (i.e., A and B and C). Also, as used in this disclosure, beginning a list of conditions with the phrase "based on" should not be construed as "based only on" the set of conditions, but rather as "based at least in part on" the set of conditions. For example, a conclusion stated as "based on condition A" could be based on both condition A and condition B without departing from the scope of this disclosure.

[0111] As used herein, the terms "comprise," "include," or "contain" are used interchangeably, can have the same meaning, and are to be interpreted in an inclusive and broad manner. The terms "comprise," "include," or "contain" can be used before a list of elements to indicate that at least all of the listed elements in the list are present, but that other elements not in the list may also be present. For example, if A comprises B and C, then both {B,C} and {B,C,D} are within the scope of A.

[0112] The present disclosure, along with the accompanying drawings, describes illustrative configurations that are not representative of all examples that may be implemented or of all configurations within the scope of the present disclosure. The term "exemplary" should not be interpreted as "preferred" or "advantageous over other examples," but rather as "illustration, instance, or example." By reading this disclosure, including the description of the embodiments and figures, those skilled in the art will recognize that the technology disclosed herein may be implemented using alternative embodiments. Those skilled in the art will recognize that the embodiments, or specific features of the embodiments, described herein may be combined to arrive at still other embodiments for practicing the technology described in the present disclosure. Thus, the present disclosure is not limited to the examples and designs described herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0113] The flowcharts and block diagrams in the figures illustrate example architecture, functionality, and operation of possible implementations of systems, methods, and devices according to various embodiments. It should be noted that in some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending on the functionality involved. Likewise, additional steps may be included in such methods, and certain steps may be omitted or combined, in methods consistent with various embodiments.

[0114] It is understood that the described embodiments are not mutually exclusive, and that elements, components, materials, or steps described in connection with one illustrative embodiment may be combined with or removed from other embodiments in any suitable manner to achieve desired design objectives.

[0115] References herein to "some embodiments" or "some exemplary embodiments" mean that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment. The appearances of the phrases "one embodiment," "some embodiments," or "another embodiment" in various places in this disclosure do not necessarily all refer to the same embodiments, nor do separate or alternative embodiments necessarily exclude other embodiments from one another.

[0116] Additionally, the articles "a" and "an," as used in this disclosure and the appended claims, should generally be construed to mean "one or more," unless otherwise specified or unless it is clear from the context that the singular form is intended.

[0117] Unless expressly stated otherwise, each numerical value and range should be interpreted as an approximation, such as by the word "about" or "approximately" preceding the value or range value.

[0118] Although elements in the following method claims, if any, are recited in a particular order, these elements are not necessarily intended to be limited to being performed in this particular order, unless the recitation of a claim otherwise suggests a particular order for performing some or all of these elements.

[0119] It is understood that certain features of the present disclosure, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features herein that are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable subcombination, or as appropriate in any other described embodiment herein. Certain features described in the context of various embodiments are not essential features of those embodiments, unless noted as such.

[0120] It will be further understood that various changes, substitutions, and variations in the details, materials, and arrangements of parts described and illustrated to explain the nature of the described embodiments may be made by those skilled in the art without departing from the scope of the present disclosure, and it is therefore intended that the following claims encompass all such alternatives, modifications, and variations as fall within the scope of the claims.

[0121] Supplementary Note 1: A method for managing channel state information feedback compression, the method comprising: receiving at least one reference signal; estimating a channel condition based on at least one reference signal; reducing the dimension of the estimated channel state using a first autoencoder in a user equipment (UE) to generate a low-dimensional channel state; using a second autoencoder and an autoencoder decoder in the UE to calculate an error measure based on a difference between the UE channel condition input and the UE channel condition output; and transmitting at least one of a channel state information bit stream or the at least one message, wherein the at least one of the channel state information bit stream or the at least one message is based on at least one of a value indicative of a low-dimensional channel state or an error measurement.

[0122] Supplementary Note 2: The method of Supplementary Note 1, further comprising generating at least one of a channel state information bit stream or at least one message based on at least one of a value indicating a low-dimensional channel state or an error measurement.

[0123] Supplementary Note 3: The method of Supplementary Note 1, wherein the at least one reference signal is received from a base station.

[0124] Supplementary Note 4: The method of Supplementary Note 1, wherein at least one of the channel state information bit stream or the at least one message is transmitted to a base station.

[0125] Supplementary Note 5: The method of Supplementary Note 4, wherein at least one of the channel state information bit stream or the at least one message transmitted to the base station includes at least one of a value indicative of a low-dimensional channel state or an error measure.

[0126] Appendix 6: The first autoencoder is based on a deep neural network. The method according to claim 1.

[0127] Appendix 7: The method of Appendix 6, wherein the deep neural network includes a plurality of neural nodes, each of which includes weights and parameters used to generate an output based on the neural node input.

[0128] Appendix 8: The method of Appendix 1, further comprising inputting the estimated channel state to a first autoencoder, wherein the estimated channel state input to the first autoencoder is a high-dimensional channel state.

[0129] Appendix 9: The method of Appendix 1, wherein calculating an error measure based on a difference between a UE channel condition input and a UE channel condition output includes using a loss function.

[0130] Appendix 10: The method of Appendix 1, further comprising calculating updated weights and parameters for the first autoencoder encoder and the autoencoder decoder based on the error measurement.

[0131] Appendix 11: The method of Appendix 1, further comprising receiving at least one of autoencoder encoder weights, autoencoder encoder parameters, autoencoder encoder structure, autoencoder decoder weights, autoencoder decoder parameters, or autoencoder decoder structure.

[0132] Appendix 12: The method described in Appendix 1, wherein the UE is comprised of information associated with the adopted structure of the autoencoder encoder, the adopted structure of the autoencoder decoder, the autoencoder encoder weights, the autoencoder decoder weights, the structure of the autoencoder encoder, the structure of the autoencoder decoder, the channel state information bitstream generation, and the decoding scheme.

[0133] Supplementary Note 13: The method of Supplementary Note 1, further comprising receiving at least one of an autoencoder encoder structure or an autoencoder decoder structure.

[0134] Supplementary Note 14: The method of Supplementary Note 13, wherein at least one of the autoencoder encoder structure or the autoencoder decoder structure is included in at least one of the master information block message or the system information block message.

[0135] Supplementary Note 15: An apparatus for channel state information feedback compression, comprising: a memory for storing instructions; and a processor configured to execute instructions stored in a memory to receive at least one reference signal, estimate a channel state based on the at least one reference signal, reduce a dimension of the estimated channel state using a first autoencoder in a user equipment (UE) so that a low-dimensional channel state is generated, calculate an error measure based on a difference between a UE channel state input and a UE channel state output using a second autoencoder and an autoencoder decoder in the UE, and transmit at least one of a channel state information bit stream or the at least one message, wherein at least one of the channel state information bit stream or the at least one message is based on at least one of a value indicative of the low-dimensional channel state or the error measure.

[0136] Supplementary Note 16: The processor executes instructions stored in the memory to generate a channel state information bitstream based on at least one of a value indicative of a low-dimensional channel state or an error measurement. 16. The apparatus of claim 15, further configured to generate at least one of a stream or at least one message.

[0137] Supplementary Note 17: The apparatus of Supplementary Note 15, wherein the at least one reference signal is received from a base station.

[0138] Supplementary Note 18: The apparatus of Supplementary Note 15, wherein at least one of the channel state information bit stream or the at least one message is transmitted to a base station.

[0139] Supplementary Note 19: The apparatus of Supplementary Note 18, wherein at least one of the channel state information bit stream or at least one message transmitted to the base station includes at least one of a value indicative of a low-dimensional channel state or an error measure.

[0140] Addendum 20: The apparatus of Addendum 15, wherein the first autoencoder is based on a deep neural network.

[0141] Appendix 21: The apparatus of Appendix 20, wherein the deep neural network includes a plurality of neural nodes, each of the neural nodes including weights and parameters used to generate an output based on the neural node input.

[0142] Supplementary Note 22: The apparatus of Supplementary Note 15, wherein the processor is further configured to execute instructions stored in the memory to input the estimated channel state to a first autoencoder, and the estimated channel state input to the first autoencoder is a high-dimensional channel state.

[0143] 23. The apparatus of claim 15, wherein calculating the error measure based on a difference between the UE channel condition input and the UE channel condition output includes using a loss function.

[0144] Supplementary Note 24: The apparatus of Supplementary Note 15, wherein the processor is further configured to execute instructions stored in the memory to calculate updated weights and parameters for the first autoencoder encoder and the autoencoder decoder based on the error measurement.

[0145] 25. The apparatus of claim 15, wherein the processor is further configured to execute instructions stored in the memory to receive at least one of autoencoder encoder weights, autoencoder encoder parameters, autoencoder encoder structure, autoencoder decoder weights, autoencoder decoder parameters, or autoencoder decoder structure.

[0146] Appendix 26: The apparatus of Appendix 15, wherein the UE is configured with information associated with an adopted structure of the autoencoder encoder, an adopted structure of the autoencoder decoder, autoencoder encoder weights, autoencoder decoder weights, an autoencoder encoder structure, an autoencoder decoder structure, channel state information bitstream generation, and a decoding scheme.

[0147] Supplementary Note 27: The apparatus of Supplementary Note 15, wherein the processor is further configured to execute instructions stored in the memory to receive at least one of an autoencoder encoder structure or an autoencoder decoder structure.

[0148] Supplementary Note 28: At least one of the autoencoder encoder structure or the autoencoder decoder structure may include a master information block message or a system information block message. 28. The apparatus of claim 27, wherein the apparatus is included in at least one of the messages.

[0149] Supplementary Note 29: A non-transitory computer-readable medium storing instructions executable by one or more processors of an apparatus for managing channel state information feedback compression to implement a method, the method comprising: receiving at least one reference signal; estimating a channel condition based on at least one reference signal; reducing the dimension of the estimated channel state using a first autoencoder in the user equipment UE so that a low-dimensional channel state is generated; using a second autoencoder and an autoencoder decoder in the UE to calculate an error measure based on a difference between the UE channel condition input and the UE channel condition output; transmitting a channel state information bit stream or at least one message, wherein at least one of the channel state information bit stream or the at least one message is based on at least one of a value indicative of a low-dimensional channel state or an error measurement.

[0150] Supplementary Note 30: A method for managing channel state information feedback compression, the method comprising: transmitting at least one reference signal; receiving at least one of a channel state information bit stream or at least one message in response to the transmitted at least one reference signal; determining at least one of a value indicative of a low-dimensional channel state or an error measure based on at least one of the channel state information bit stream or the at least one message; calculating at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on at least one of an estimated channel state or a value indicative of an error measure, wherein the estimated channel state is based on a low-dimensional channel state; and updating at least one of the weights or parameters of an autoencoder decoder in the base station based on the calculated at least one of the updated weights or the updated parameters.

[0151] 31. The method of claim 30, wherein the estimated channel state is determined using an autoencoder decoder in the base station to decode the low-dimensional channel state.

[0152] Supplementary Note 32: The method of Supplementary Note 30, wherein at least one reference signal is transmitted to a user equipment (UE), and at least one of the channel state information bit stream or the at least one message is received from the UE.

[0153] Supplementary Note 33: The method of Supplementary Note 30, wherein determining at least one of the values ​​indicative of the low-dimensional channel state or error measure includes extracting at least one of the values ​​indicative of the low-dimensional channel state or error measure from at least one of the channel state information bit stream or the at least one message.

[0154] Addendum 34: The method of Addendum 30, wherein the autoencoder decoder in the base station is based on a deep neural network.

[0155] 35. The method of claim 34, wherein the deep neural network includes a plurality of neural nodes, each of the neural nodes including at least one of weights or parameters used to generate an output based on the neural node input.

[0156] 36. The method of claim 30, further comprising converting the low-dimensional channel state to a high-dimensional channel state using an autoencoder decoder.

[0157] Supplementary Note 37: The method of Supplementary Note 30, wherein calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station includes calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station based on a value indicative of the error measure.

[0158] Addendum 38: The method of Addendum 37, wherein at least one of the updated weights or updated parameters for the autoencoder decoder corresponds to at least one of the updated weights or updated parameters for the autoencoder encoder in the user equipment.

[0159] Supplementary Note 39: The method of Supplementary Note 30, wherein calculating at least one of updated weights or updated parameters further includes calculating at least one of updated weights or updated parameters for each of a first autoencoder encoder in a user equipment (UE), a second autoencoder encoder in the UE, and an autoencoder decoder in the UE, wherein the second autoencoder encoder in the UE and the autoencoder decoder in the UE belong to the same autoencoder pair.

[0160] Supplementary Note 40: The method of Supplementary Note 39, wherein calculating at least one of updated weights or updated parameters further includes updating at least one of weights or parameters of a first autoencoder encoder in the UE, a second autoencoder encoder in the UE, and an autoencoder decoder in the UE.

[0161] Supplementary Note 41: The method of Supplementary Note 30, wherein the base station is configured with information associated with the adopted structure of the autoencoder encoder, the adopted structure of the autoencoder decoder, the autoencoder encoder weights, the autoencoder decoder weights, the structure of the autoencoder encoder, the structure of the autoencoder decoder, the channel state information bitstream generation, and the decoding scheme.

[0162] Supplementary Note 42: The method of Supplementary Note 30, further comprising transmitting at least one of an autoencoder encoder structure or an autoencoder decoder structure to a user equipment.

[0163] Supplementary Note 43: An apparatus for managing channel state information feedback compression, the apparatus comprising: a memory for storing instructions; and a processor configured to execute instructions stored in a memory to transmit at least one reference signal, receive at least one of a channel state information bit stream or the at least one message in response to the transmitted at least one reference signal, determine at least one of values ​​indicative of a low-dimensional channel condition or an error measure based on the at least one of the channel state information bit stream or the at least one message, calculate at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on the at least one of the values ​​indicative of the estimated channel or error measure, wherein the estimated channel condition is based on the low-dimensional channel condition, and update at least one of the weights or parameters of the autoencoder decoder in the base station based on the calculated at least one of the updated weights or updated parameters.

[0164] 44. The apparatus of claim 43, wherein the estimated channel state is determined using an autoencoder decoder in a base station to decode the low-dimensional channel state.

[0165] Supplementary Note 45: The apparatus of Supplementary Note 43, wherein at least one reference signal is transmitted to a user equipment (UE), and at least one of the channel state information bit stream or the at least one message is received from the UE.

[0166] Attachment 46: The apparatus of Attachment 43, wherein determining at least one of the values ​​indicative of the low-dimensional channel condition or the error measure includes extracting at least one of the values ​​indicative of the low-dimensional channel condition or the error measure from at least one of the channel state information bit stream or the at least one message.

[0167] Addendum 47: The apparatus of Addendum 43, wherein the autoencoder decoder in the base station is based on a deep neural network.

[0168] 48. The apparatus of claim 47, wherein the deep neural network includes a plurality of neural nodes, each of the neural nodes including at least one of weights or parameters used to generate an output based on the neural node input.

[0169] 49. The apparatus of claim 43, wherein the processor is further configured to execute instructions stored in the memory to transform the low-dimensional channel state into a high-dimensional channel state using an autoencoder decoder.

[0170] 50. The apparatus of claim 43, wherein calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station includes calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station based on a value indicative of the error measure.

[0171] Addendum 51: The apparatus of Addendum 50, wherein at least one of the updated weights or updated parameters for the autoencoder decoder corresponds to at least one of the updated weights or updated parameters for the autoencoder encoder in the user equipment.

[0172] Supplementary Note 52: The apparatus of Supplementary Note 43, wherein calculating at least one of updated weights or updated parameters further includes calculating at least one of updated weights or updated parameters for each of a first autoencoder encoder in a user equipment (UE), a second autoencoder encoder in the UE, and an autoencoder decoder in the UE, wherein the second autoencoder encoder in the UE and the autoencoder decoder in the UE belong to the same autoencoder pair.

[0173] Addendum 53: The apparatus of Addendum 52, wherein calculating at least one of updated weights or updated parameters further includes updating at least one of weights, parameters, or structures of a first autoencoder encoder in the UE, a second autoencoder encoder in the UE, and an autoencoder decoder in the UE.

[0174] 54. The apparatus of claim 43, wherein the base station is configured with information associated with the employed structure of the autoencoder encoder, the employed structure of the autoencoder decoder, the autoencoder encoder weights, the autoencoder decoder weights, the structure of the autoencoder encoder, the structure of the autoencoder decoder, the channel state information bitstream generation, and the decoding scheme.

[0175] Addendum 55: The apparatus of Addendum 43, wherein the processor is further configured to execute instructions stored in the memory to transmit at least one of an autoencoder encoder structure or an autoencoder decoder structure to a user equipment.

[0176] Appendix 56: A non-transitory computer-readable medium storing instructions executable by one or more processors of an apparatus for managing channel state information feedback compression to perform a method, the method comprising: transmitting at least one reference signal; receiving at least one of a channel state information bit stream or at least one message in response to the transmitted at least one reference signal; determining at least one of a value indicative of a low-dimensional channel state or an error measure based on at least one of the channel state information bit stream or the at least one message; calculating at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on at least one of an estimated channel state or a value indicative of an error measure, wherein the estimated channel state is based on a low-dimensional channel state; and updating at least one of the weights or parameters of an autoencoder decoder in the base station based on the calculated at least one of the updated weights or the updated parameters.

[0177] Supplementary Note 57: A method for managing channel state information feedback compression, the method comprising: receiving at least one reference signal; estimating a channel condition based on at least one reference signal; reducing the dimension of the estimated channel state using a first autoencoder in a user equipment (UE) to generate a low-dimensional channel state; using a second autoencoder and an autoencoder decoder in the UE to calculate an error measure based on a difference between the UE channel condition input and the UE channel condition output; calculating at least one of updated weights or updated parameters for a first autoencoder in the UE based on the value indicative of the error measure; updating at least one of weights or parameters of a first autoencoder in the UE based on the calculated at least one of the updated weights or the updated parameters; generating at least one of a channel state information bit stream or at least one message based on at least one of a low-dimensional channel state or at least one of updated weights or updated parameters for a first autoencoder in the UE; and transmitting at least one of a channel state information bit stream or at least one message.

[0178] Addendum 58: Computing at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on the value indicative of the error measure; 58. The method of claim 57, further comprising: transmitting at least one of updated weights or updated parameters for an autoencoder decoder in the base station to the base station.

[0179] Supplementary Note 59: First autoencoder in UE, second autoencoder in UE 58. The method of claim 57, further comprising transmitting information about at least one of the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, the autoencoder decoder in the base station, or calculated weights, parameters, or structures of at least one of the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, the autoencoder decoder in the UE, or the autoencoder decoder in the base station.

[0180] Supplementary Note 60: The method of Supplementary Note 59, further comprising: receiving a response including a confirmation including calculated weights, parameters, or structures of at least one of the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, the autoencoder decoder in the UE, or the autoencoder decoder in the base station; and updating at least one of the structures or calculated weights or parameters of at least one of the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, or the autoencoder decoder in the UE based on the confirmation.

[0181] Supplementary Note 61: A method for managing channel state information feedback compression, the method comprising: transmitting at least one reference signal; receiving at least one of a channel state information bit stream or at least one message in response to the transmitted at least one reference signal; determining at least one of updated weights, updated parameters, or a structure for at least one of a first autoencoder in a user equipment (UE), a second autoencoder in the UE, or an autoencoder decoder in the UE; calculating at least one of updated weights or updated parameters for an autoencoder decoder in the base station based on at least one of updated weights or updated parameters for the autoencoder encoder in the UE; and updating at least one of the weights, parameters or structure of an autoencoder decoder in the base station based on the calculated at least one of the updated weights or the updated parameters or structure.

[0182] Attachment 62: The method of attachment 61, further comprising receiving from the UE at least one of calculated weights, parameters, or updated structures of at least one of an autoencoder decoder in the base station, a first autoencoder encoder in the UE, a second autoencoder encoder in the UE, or an autoencoder decoder in the UE.

[0183] Addendum 63: The method of Addendum 62, wherein the updated structure is a sandwich structure.

[0184] Attachment 64: The method of attachment 61, further comprising transmitting to the UE at least one of the calculated weights, parameters, or structures of at least one of the autoencoder decoder in the base station, the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, or the autoencoder decoder in the UE. Explanation of Abbreviations

[0185] 3GPP 3rd Generation Partnership Project AE Autoencoder CSI Channel State Information DL Downlink gNB or next-generation base station LTE long-term evolution NR New Radio PDU Protocol Data Unit PSCCH Physical Sidelink Control Channel PSSCH Physical Sidelink Shared Channel RS reference signal TX / RX Transmit / Receive UE User Equipment UL Uplink Uu UMTS Air Interface

Claims

1. 1. A method for managing channel state information feedback compression, comprising: receiving at least one reference signal; estimating a channel condition based on the at least one reference signal; reducing the dimension of the estimated channel state using a first autoencoder in a user equipment (UE) to generate a low-dimensional channel state; calculating an error measure based on a difference between a UE channel condition input and a UE channel condition output using a second autoencoder and an autoencoder decoder within the UE; transmitting at least one of a channel state information bit stream or at least one message, wherein the at least one of the channel state information bit stream or the at least one message is based on at least one of a value indicative of the low-dimensional channel state or the error measurement. method.

2. generating the at least one of the channel state information bit stream or the at least one message based on at least one of the values ​​indicative of the low-dimensional channel state or the error measure. The method of claim 1.

3. the at least one of the channel state information bit stream or the at least one message is transmitted to a base station. The method of claim 1.

4. The first autoencoder is based on a deep neural network. The method of claim 1.

5. and inputting the estimated channel state to the first autoencoder, wherein the estimated channel state input to the first autoencoder is a high-dimensional channel state. The method of claim 1.

6. calculating the error measure based on a difference between a UE channel condition input and a UE channel condition output includes using a loss function. The method of claim 1.

7. The method of claim 1 , further comprising: calculating updated weights and parameters for the first autoencoder encoder and the autoencoder decoder based on the error measure.

8. further comprising receiving at least one of autoencoder encoder weights, autoencoder encoder parameters, autoencoder encoder structure, autoencoder decoder weights, autoencoder decoder parameters, or autoencoder decoder structure. The method of claim 1.

9. 1. A method for managing channel state information feedback compression, comprising: transmitting at least one reference signal; receiving at least one of a channel state information bit stream or at least one message in response to the at least one transmitted reference signal; determining at least one of a value indicative of a low-dimensional channel state or an error measure based on the at least one of the channel state information bit stream or the at least one message; calculating at least one of updated weights or updated parameters for an autoencoder decoder in a base station based on at least one of estimated channel conditions or the value indicative of the error measure, wherein the estimated channel conditions are based on the low-dimensional channel conditions; updating at least one of weights or parameters of the autoencoder decoder in the base station based on the calculated at least one of updated weights or updated parameters; A method comprising:

10. the estimated channel state is determined using the autoencoder decoder in the base station to decode the low-dimensional channel state.

10. The method of claim 9.

11. the at least one reference signal is transmitted to a user equipment (UE), and the at least one of the channel state information bitstream or the at least one message is received from the UE.

10. The method of claim 9.

12. determining at least one of the values ​​indicative of the low-dimensional channel state or the error measure comprises extracting at least one of the values ​​indicative of the low-dimensional channel state or the error measure from the at least one of the channel state information bit stream or the at least one message.

10. The method of claim 9.

13. the autoencoder decoder in the base station is based on a deep neural network; 10. The method of claim 9.

14. and further comprising: converting the low-dimensional channel state to a high-dimensional channel state using the autoencoder decoder.

10. The method of claim 9.

15. calculating at least one of updated weights or updated parameters for the autoencoder decoder in the base station based on the value indicative of the error measure; 10. The method of claim 9.

16. Calculating at least one of updated weights or updated parameters further includes calculating at least one of updated weights or updated parameters for each of a first autoencoder in a user equipment (UE), a second autoencoder in the UE, and an autoencoder decoder in the UE, wherein the second autoencoder in the UE and the autoencoder decoder in the UE belong to the same autoencoder pair.

10. The method of claim 9.

17. 1. A method for managing channel state information feedback compression, comprising: receiving at least one reference signal; estimating a channel condition based on the at least one reference signal; reducing the dimension of the estimated channel state using a first autoencoder in a user equipment (UE) to generate a low-dimensional channel state; calculating an error measure based on a difference between a UE channel condition input and a UE channel condition output using a second autoencoder and an autoencoder decoder within the UE; calculating at least one of updated weights or updated parameters for the first autoencoder within the UE based on a value indicative of the error measure; updating at least one of weights or parameters of the first autoencoder in the UE based on the calculated at least one of updated weights or updated parameters; generating at least one of a channel state information bitstream or at least one message based on the at least one of the low-dimensional channel conditions or updated weights or updated parameters for the first autoencoder in the UE; transmitting the at least one of the channel state information bitstream or the at least one message; A method comprising:

18. calculating at least one of updated weights or updated parameters for an autoencoder decoder in a base station based on the value indicative of the error measure; transmitting the at least one of updated weights or updated parameters for the autoencoder decoder in the base station to a base station.

18. The method of claim 17.

19. transmitting information about at least one of the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, the autoencoder decoder in the UE, or an autoencoder decoder in a base station, or calculated weights, parameters, or structures of at least one of the first autoencoder encoder in the UE, the second autoencoder encoder in the UE, the autoencoder decoder in the UE, or the autoencoder decoder in the base station.

18. The method of claim 17.

20. 1. A method for managing channel state information feedback compression, comprising: transmitting at least one reference signal; receiving at least one of a channel state information bit stream or at least one message in response to the at least one transmitted reference signal; determining at least one of updated weights, updated parameters, or a structure for at least one of a first autoencoder in a user equipment (UE), a second autoencoder in the UE, or an autoencoder decoder in the UE; calculating at least one of updated weights or updated parameters for an autoencoder decoder in a base station based on the at least one of the updated weights or the updated parameters for an autoencoder decoder in the UE; updating at least one of weights, parameters or structure of the autoencoder decoder in the base station based on the calculated at least one of updated weights or updated parameters or structure; A method comprising:

Images