Channel state information feedback method, apparatus, device, and storage medium

Abstract

This application discloses a channel state information feedback method, apparatus, device, and storage medium. The method includes: receiving a downlink reference signal transmitted by a base station; performing channel estimation based on the downlink reference signal to obtain a channel matrix of the downlink channel; inputting the channel matrix into a trained encoder model for feature compression to obtain a first feature vector; performing normalization and encapsulation processing on the first feature vector to generate uplink control information; and feeding back the uplink control information to the base station, which is used by the base station to reconstruct the corresponding channel matrix based on the uplink control information. This solution converts channel state information into a low-dimensional feature vector and encapsulates it into uplink control information for feedback. While ensuring that the base station can effectively reconstruct the channel state information, it significantly reduces the data overhead and feedback burden of channel state information feedback, thereby improving the efficiency of channel state information feedback.

Description

Technical Field

[0001] This application relates to the field of communication technology, and in particular to a channel state information feedback method, apparatus, device and storage medium. Background Technology

[0002] In Massive MIMO (Massive Multiple-Input Multiple-Output), accurate acquisition of downlink channel state information is crucial for base stations to perform efficient beamforming. As antenna size increases, channel dimension significantly increases, and the accuracy of channel representation directly affects overall link performance.

[0003] In existing technologies, 3GPP (3rd Generation Partnership Project) employs a codebook-based feedback mechanism, where terminal devices provide channel state information feedback by selecting a precoding matrix indication from a predefined codebook. However, this approach suffers from the following problems: the predefined codebook is difficult to accurately match complex real-world channels, resulting in quantization errors; and as antenna size increases, feedback overhead increases significantly, consuming substantial uplink resources. Summary of the Invention

[0004] This application provides a channel state information feedback method, apparatus, device, and storage medium, which can achieve efficient transmission of channel state information while ensuring channel reconstruction accuracy by inputting the downlink channel matrix into an encoder model for feature compression and generating uplink control information based on the compressed feature vector.

[0005] Firstly, this application provides a channel state information feedback method, applied to a terminal device, comprising: The downlink reference signal sent by the base station is received, and channel estimation is performed based on the downlink reference signal to obtain the channel matrix of the downlink channel. The channel matrix is ​​input into the trained encoder model for feature compression to obtain the first feature vector; The first feature vector is normalized and encapsulated to generate uplink control information, which is then fed back to the base station. The base station uses the uplink control information to reconstruct the corresponding channel matrix.

[0006] Secondly, this application provides a channel state information feedback device, applied to a terminal device, comprising: The channel estimation module is configured to receive downlink reference signals transmitted by the base station, perform channel estimation based on the downlink reference signals, and obtain the channel matrix of the downlink channel. The feature compression module is configured to input the channel matrix into the trained encoder model for feature compression to obtain a first feature vector; The information feedback module is configured to perform normalization and encapsulation processing on the first feature vector to generate uplink control information, and to feed back the uplink control information to the base station, so that the base station can restore the corresponding channel matrix based on the uplink control information.

[0007] Thirdly, this application provides a channel state information feedback device, comprising: One or more processors; A memory that stores one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the channel state information feedback method as described in the first aspect.

[0008] Fourthly, this application provides a storage medium containing computer-executable instructions, which, when executed by a computer processor, are used to perform the channel state information feedback method as described in the first aspect.

[0009] This application constructs a channel state information feedback method based on feature compression and structured encapsulation, achieving efficient expression and reconstruction of high-dimensional channel information by terminal devices under conditions of limited feedback overhead. The method receives downlink reference signals transmitted by the base station and performs channel estimation to obtain the downlink channel matrix. The channel matrix is ​​then input into a trained encoder model for feature compression to generate a first feature vector. Based on this, the first feature vector is normalized and encapsulated to generate uplink control information, which is fed back to the base station via uplink signaling, enabling the base station to reconstruct the corresponding channel matrix based on this information. This scheme significantly reduces feedback overhead by transforming the original channel matrix feedback into compressed feature feedback, while ensuring channel reconstruction accuracy. It is suitable for large-scale antenna arrays and communication scenarios with limited feedback resources. Attached Figure Description

[0010] Figure 1 This is a flowchart of a channel state information feedback method provided in an embodiment of this application; Figure 2 This is a flowchart of an uplink control information generation method provided in an embodiment of this application; Figure 3 This is a flowchart of a scaling factor determination method provided in an embodiment of this application; Figure 4 This is a flowchart of a first field generation method provided in an embodiment of this application; Figure 5 This is a flowchart of a second field generation method provided in an embodiment of this application; Figure 6 This is a flowchart of an uplink control information rate control method provided in an embodiment of this application; Figure 7 This is a schematic diagram of the structure of uplink control information provided in an embodiment of this application; Figure 8 This is a structural block diagram of a channel state information feedback device provided in an embodiment of this application; Figure 9 This is a schematic diagram of the structure of a channel state information feedback device provided in an embodiment of this application. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as being processed sequentially, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. A process can be terminated when its operation is completed, but it may also have additional steps not included in the drawings. A process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0012] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0013] Currently, acquiring high-precision downlink channel state information is crucial for efficient beamforming in large-scale antenna arrays. However, as antenna size continues to increase, traditional codebook-based channel state information feedback mechanisms are gradually revealing significant limitations. In existing technologies, terminals typically employ the Type I / Type II codebook method defined by the 3GPP standard, selecting the best-matching precoding matrix index from a predefined codebook set for feedback to assist the base station in beam selection. However, this mechanism has significant shortcomings in complex channel environments and large-scale antenna scenarios. Regarding representation accuracy, predefined codebooks struggle to accurately characterize the high-dimensional features of actual channels, leading to persistent quantization errors and constraining performance limits. In terms of feedback overhead, as antenna array size increases and channel dimensions grow rapidly, the number of Type II codebook feedback bits increases significantly, consuming substantial uplink resources and limiting the scalability of the antenna system. Regarding data representation, with the development of AI-based channel compression methods, compression results are typically represented as continuous feature vectors, while existing uplink control information formats are primarily designed for discrete indices, lacking an effective mechanism for carrying unstructured continuous data. Therefore, existing codebook-based channel state information feedback methods are difficult to balance feedback accuracy, overhead efficiency, and expressive power in high-dimensional and complex channel scenarios.

[0014] Therefore, this invention aims to propose a channel state information feedback method based on feature compression and structured representation. This method enables efficient compression and accurate reconstruction of downlink channel information in scenarios where high-dimensional channels and limited feedback resources coexist, thereby supporting high-quality beamforming and scheduling decisions at the base station. The method receives downlink reference signals transmitted by the base station and performs channel estimation to obtain the downlink channel matrix. The channel matrix is ​​then input into a trained encoder model for feature compression to obtain a first feature vector. Based on this, the first feature vector is normalized and encapsulated to generate uplink control information, which is then reported to the base station. The base station reconstructs the channel matrix based on this uplink control information to recover key channel features for subsequent beam selection and resource allocation. By introducing a channel feature compression mechanism based on the encoder model, this method effectively reduces feedback overhead and improves channel reconstruction accuracy, making it suitable for large-scale antenna arrays and wireless communication scenarios with limited feedback.

[0015] Figure 1 This is a flowchart of a channel state information feedback method provided in an embodiment of this application. (Reference) Figure 1 The channel state information feedback method specifically includes: S110. Receive the downlink reference signal sent by the base station, perform channel estimation based on the downlink reference signal, and obtain the channel matrix of the downlink channel.

[0016] The downlink reference signal can be a sequence of signals transmitted by the base station for channel measurement, reflecting the transmission characteristics of the wireless channel at the current moment. The channel matrix can be a parameter matrix used to describe the multiple-input multiple-output transmission relationship between the transmitter and receiver, characterizing the channel gain, phase change, and spatial correlation between each transmit and receive antenna. The downlink reference signal and the channel matrix together constitute the core basic information for characterizing the downlink transmission characteristics.

[0017] In one embodiment, the method for receiving the downlink reference signal sent by the base station may be: the terminal device receives the reference signal sent by the base station at a set time-frequency resource location to obtain the downlink reference signal.

[0018] In one embodiment, the channel estimation based on the downlink reference signal can be performed as follows: the terminal device obtains the local reference signal corresponding to the downlink reference signal, performs a complex division operation between the downlink reference signal and the local reference signal to obtain the channel coefficients corresponding to each time domain resource unit, and combines the channel coefficients to generate a channel matrix.

[0019] Through the above steps, the downlink reference signal sent by the base station is received and channel estimation is performed to obtain the channel matrix of the downlink channel. This enables the terminal equipment to accurately characterize the current downlink transmission characteristics, thereby providing a reliable basis for subsequent channel quality assessment and improving the performance and stability of the communication system.

[0020] S120. Input the channel matrix into the trained encoder model for feature compression to obtain the first feature vector.

[0021] The channel matrix can be a complex matrix obtained by processing the downlink reference signal and the local reference signal, used to represent the transmission characteristics of the communication channel. The encoder model can be a pre-trained neural network model, whose input is the channel matrix and output is a low-dimensional feature vector, used to compress the channel information. The first feature vector can be the compressed result of the encoder model output. The encoder model can be composed of a convolutional neural network, which extracts features layer by layer and reduces dimensionality so that the output first feature vector contains the main channel information. The base station is equipped with a decoder model corresponding to the encoder model, used to perform channel reconstruction processing on the uplink control information.

[0022] In one embodiment, the channel matrix can be compressed by inputting the channel matrix into the input layer of the encoder model and performing nonlinear mapping through a multi-layer neural network to obtain the compressed feature representation.

[0023] Through the above steps, the channel matrix can be input into the trained encoder model for feature compression to obtain the first feature vector, which enables the terminal device to reduce the data dimension while retaining the main channel features, thereby improving the efficiency of subsequent processing or transmission.

[0024] S130. The first feature vector is normalized and encapsulated to generate uplink control information, which is then fed back to the base station for the base station to reconstruct the corresponding channel matrix based on the uplink control information.

[0025] The first feature vector can be a low-dimensional vector data output by the encoder model after feature compression of the channel matrix, used to carry compressed feature information in the channel matrix. Normalization processing can be a scaling operation performed on the values ​​of each dimension of the first feature vector to ensure that each dimension meets a preset numerical range. Encapsulation processing can be the process of organizing and encapsulating the normalized first feature vector according to the field format of uplink control information. Uplink control information can be control information used to feed back compressed feature data to the base station.

[0026] In one embodiment, the normalization and encapsulation of the first feature vector can be performed by: scaling each element of the first feature vector, writing the scaled first feature vector into the uplink control information, and adding at least one of vector length identifier, quantization identifier, version identifier and verification information to the uplink control information to generate uplink control information that can be transmitted.

[0027] In one embodiment, the uplink control information can be fed back to the base station by sending uplink control information through the uplink control channel.

[0028] Through the above steps, the first feature vector can be normalized and encapsulated to generate uplink control information, which is then fed back to the base station. This allows the base station to reconstruct the corresponding channel matrix based on the received uplink control information, thereby achieving efficient feedback and reconstruction of compressed channel information.

[0029] Optionally, Figure 2 This is a flowchart illustrating an uplink control information generation method provided in an embodiment of this application. (Reference) Figure 2 The method for generating uplink control information specifically includes: S1301. Determine a scaling factor based on the first feature vector, and normalize the first feature vector based on the scaling factor to obtain a second feature vector.

[0030] The scaling factor can be a parameter used to adjust the magnitude of the first feature vector, and its value can be determined by the statistical properties of the first feature vector. The second feature vector can be a normalized feature vector used for subsequent encapsulation and transmission.

[0031] In one embodiment, the scaling factor can be determined based on the first feature vector by obtaining the maximum value of the elements in the first feature vector and using the obtained maximum value as the scaling factor.

[0032] In one embodiment, the normalization process for the first feature vector can be achieved by dividing each element of the first feature vector by a scaling factor, thereby compressing each element to obtain the second feature vector.

[0033] Through the above steps, a scaling factor can be determined based on the first feature vector, and the first feature vector can be normalized using the scaling factor to obtain the second feature vector, thereby providing a data foundation with a unified scale for subsequent data encapsulation and feedback.

[0034] Optionally, Figure 3 This is a flowchart of a scaling factor determination method provided in an embodiment of this application. (Reference) Figure 3 The scaling factor determination method specifically includes: S13011. Perform amplitude calculation on each element in the first feature vector to obtain the amplitude corresponding to each element.

[0035] The magnitude can be a numerical value used to represent the size of a vector element. For complex elements, it can be the modulus of their real and imaginary parts, and for real elements, it can be their absolute value.

[0036] In one embodiment, the amplitude calculation for each element in the first feature vector can be performed as follows: read each element in the first feature vector one by one, and when the element is a real number, perform an absolute value operation on it; when the element is a complex number, perform a square summation and square root operation on its real part and imaginary part to obtain the corresponding amplitude.

[0037] Through the above steps, the amplitude can be calculated for each element in the first feature vector, and the amplitude corresponding to each element can be obtained, providing basic data for subsequent scaling factor determination and normalization processing.

[0038] S13012. The maximum value among the plurality of amplitudes is determined as the scaling factor.

[0039] In one embodiment, the maximum value among multiple amplitudes can be determined as the scaling factor by performing a comparison operation on all amplitudes, filtering out the amplitude with the largest value, and using that maximum value as the scaling factor.

[0040] By following the steps above, the maximum value among multiple amplitudes can be determined as the scaling factor, providing a basis for subsequent normalization processing of the first feature vector, thereby improving the stability and consistency of feature data during transmission.

[0041] S1302. Generate a first field based on the scaling factor and the channel matrix, and generate a second field based on the second feature vector.

[0042] The first field can be field information generated based on the scaling factor and the channel matrix, and the second field can be field information generated based on the second feature vector, which are used to construct the data structure of the uplink control information.

[0043] In one embodiment, the first field can be generated by: formatting the scaling factor, obtaining the configuration parameters of the channel matrix, and combining the formatted scaling factor and the configuration parameters to generate the first field.

[0044] In one embodiment, the second field can be generated by quantizing and encoding each element in the second feature vector, and using the quantized and encoded data as the second field.

[0045] Through the above steps, the first field can be generated based on the scaling factor and the channel matrix, and the second field can be generated based on the second feature vector, thereby constructing the uplink control information structure for feedback and providing the necessary data foundation for the base station to reconstruct the channel matrix.

[0046] Optionally, Figure 4 This is a flowchart of a first field generation method provided in an embodiment of this application. (Reference) Figure 4 The method for generating the first field specifically includes: S13021. Obtain the quantization accuracy of the base station configuration.

[0047] Quantization precision can be a parameter used by the base station to indicate the quantization method of feature data, which limits the number of bits or quantization level of each element in the feature vector during quantization encoding. The base station can be a communication node that sends configuration information to terminal devices to unify the data encoding method between the terminal devices and the base station.

[0048] In one embodiment, the quantization accuracy of the base station configuration can be obtained by parsing the quantization accuracy from the system configuration information sent by the base station.

[0049] By following the steps above, the quantization accuracy of the base station configuration can be obtained, enabling the terminal device to encode the feature data according to a unified quantization rule when generating uplink control information, thereby ensuring the accuracy of the data parsing and channel matrix reconstruction process on the base station side.

[0050] S13022. Calculate the effective rank of the channel matrix to obtain the channel rank indication information of the downlink channel.

[0051] The effective rank can be the number of independent channel dimensions with significant energy contributions in the channel matrix, reflecting the spatial multiplexing capability of the channel. The channel rank indication information can be a numerical value used to represent the effective rank.

[0052] In one embodiment, the effective rank of the channel matrix can be calculated by performing singular value decomposition on the channel matrix, obtaining multiple singular values, filtering out singular values ​​that are greater than a set energy threshold, and using the number of filtered singular values ​​as the effective rank.

[0053] Through the above steps, the effective rank of the channel matrix can be calculated, and the channel rank indication information of the downlink channel can be obtained. This enables the system to carry channel dimension information during feature feedback, thereby supporting efficient channel reconstruction and resource scheduling at the base station.

[0054] S13023. Determine the channel quality indication information of the downlink channel based on the channel matrix.

[0055] Among them, channel quality indication information can be an indicator that reflects the downlink channel quality level, which can be used by the base station for resource scheduling.

[0056] In one embodiment, the channel quality indication information of the downlink channel can be determined by: performing modulus calculation on each element in the channel matrix to obtain the channel amplitude corresponding to each element, performing a square operation on the channel amplitude to obtain the channel power value; obtaining the noise power estimate, performing a ratio operation on the channel power value and the noise power estimate to obtain the signal-to-noise ratio, and mapping the signal-to-noise ratio to the channel quality indication information.

[0057] Through the above steps, the channel quality indication information of the downlink channel can be determined according to the channel matrix, enabling the terminal device to provide channel quality-related parameters during the feedback process, thereby supporting the scheduling and optimization decisions on the base station side.

[0058] S13024. Combine the channel rank indication information, the channel quality indication information, the scaling factor, and the quantization precision to generate a first field.

[0059] In one embodiment, the method for combining channel rank indication information, channel quality indication information, scaling factor, and quantization precision to generate the first field can be: writing the channel rank indication information, channel quality indication information, scaling factor, and quantization precision into a data structure in a preset field order to form the first field.

[0060] Through the above steps, the channel rank indication information, channel quality indication information, scaling factor, and quantization precision can be combined to generate the first field, thereby constructing a data structure containing the necessary configuration information to support subsequent uplink feedback and base station parsing.

[0061] Optionally, Figure 5 This is a flowchart of a second field generation method provided in an embodiment of this application. (Reference) Figure 5 The method for generating the second field specifically includes: S13025. Obtain the quantization accuracy of the base station configuration.

[0062] Among them, quantization precision can be a parameter used by the base station to indicate the quantization method of feature data, which is used to limit the number of bits of each element in the feature vector during quantization encoding.

[0063] In one embodiment, the quantization accuracy of the base station configuration can be obtained by parsing the quantization accuracy parameter from the system broadcast information sent by the base station and using the parameter as input for subsequent data processing.

[0064] By following the steps above, the quantization accuracy of the base station configuration can be obtained, enabling the terminal device to perform quantization processing according to unified rules when generating uplink control information, thereby ensuring that the base station can correctly parse the data and complete the channel matrix reconstruction.

[0065] S13026. Map the second feature vector to the bit sequence corresponding to the quantization precision.

[0066] The bit sequence can be a binary data sequence used for transmission.

[0067] In one embodiment, the second feature vector is discretized element by element according to the quantization interval corresponding to the quantization precision, each element is mapped to the corresponding quantization index value, and the quantization index value is encoded into multiple original sequences according to the preset bit width, and then concatenated in order to form a bit sequence.

[0068] Through the above steps, the second feature vector can be mapped to a bit sequence corresponding to the quantization precision, thereby realizing the discretization encoding and efficient transmission of feature data, providing a foundation for decoding and channel reconstruction on the base station side.

[0069] S13027. Encapsulate the bit sequence to generate a second field.

[0070] In one embodiment, the method for encapsulating a bit sequence to generate a second field can be: dividing the bit sequence into segments of fixed length and splicing them together to form continuous data blocks to obtain the second field.

[0071] Through the above steps, the bit sequence can be encapsulated to generate a second field, thereby constructing a data structure that meets the transmission requirements and providing support for subsequent uplink feedback and base station parsing.

[0072] S1303. Combine the first field and the second field to generate uplink control information.

[0073] In one embodiment, the method of generating uplink control information by assembling the first field and the second field can be: concatenating the first field and the second field sequentially, and adding at least one of field length information, separator identifier and verification information during the assembly process to form complete uplink control information.

[0074] Through the above steps, the first and second fields can be packaged together to generate uplink control information, thereby constructing a complete data unit for feedback compressed channel information, providing support for data parsing and channel recovery on the base station side.

[0075] Optionally, Figure 6 This is a flowchart of an uplink control information rate control method provided in an embodiment of this application. (Reference) Figure 6 The specific methods for controlling the bit rate of this line control information include: S13031. Perform channel coding processing on the first field to obtain a first coded bit sequence with a first code rate.

[0076] Channel coding can be a process of performing forward error correction coding on data to improve its error resistance during transmission. The first code rate can be a ratio parameter between the number of coded bits after channel coding and the original number of bits. The first coded bit sequence can be the bit sequence after channel coding.

[0077] In one embodiment, the channel coding process for the first field can be performed by inputting the original bit sequence corresponding to the first field into the channel coding module and generating a coded bit sequence through convolutional coding.

[0078] Through the above steps, channel coding processing can be performed on the first field to obtain the first coded bit sequence with the first code rate, thereby improving the reliability and anti-interference capability of key control information during transmission.

[0079] S13032. Perform channel coding processing on the second field to obtain a second coded bit sequence with a second code rate, wherein the first code rate is lower than the second code rate.

[0080] The second code rate can be a higher coding parameter than the first code rate, which introduces fewer redundant bits into the second field during the encoding process, thereby improving transmission efficiency. The second coded bit sequence can be the bit sequence obtained after performing channel coding on the second field.

[0081] In one embodiment, the channel coding process for the second field can be performed by inputting the bit sequence corresponding to the second field into the channel coding module and generating the second coded bit sequence through convolutional coding.

[0082] Through the above steps, channel coding processing can be performed on the second field to obtain a second coded bit sequence with a second code rate, and the first code rate can be lower than the second code rate, thereby realizing a differentiated coding strategy and improving the reliability and efficiency of the overall communication system.

[0083] Optionally, performing channel coding processing on the second field to obtain a second coded bit sequence with a second code rate includes: The bit field is obtained by concatenating the individual bits in the second field.

[0084] The second field can be a data structure encapsulated from the quantized bit sequence, which may contain multiple sub-bit sequences. The bit field can be a continuous binary data sequence formed by concatenating multiple bits in a preset order, used for subsequent encoding or transmission.

[0085] In one embodiment, the method of concatenating the bits in the second field to obtain the bit field can be: reading the bits in the second field, concatenating the bits in sequence to form a continuous bit stream, and using the bit stream as the bit field.

[0086] Through the above steps, the bits in the second field can be concatenated to obtain the bit field, thereby forming a continuous data sequence, which provides a basis for subsequent channel coding and transmission processing.

[0087] The bit field is channel-coded using polar coding to obtain a second coded bit sequence with a second code rate.

[0088] Among them, polar codes can be a forward error correction coding method based on the channel polarization principle, which achieves encoding by constructing reliable and unreliable bit positions.

[0089] In one embodiment, the method of channel coding the bit field using polar codes can be as follows: mapping the bit field to the input bit position of the polar code, filling the information bits into the information bits, filling the frozen bits with preset values ​​to obtain the input vector, and multiplying the input vector by the polar matrix to obtain the second coded bit sequence.

[0090] Through the above steps, the bit field can be channel-coded using polar codes to obtain a second coded bit sequence with a second code rate, thereby improving the anti-interference capability of the data while meeting the transmission efficiency requirements.

[0091] S13033. Packet the first encoded bit sequence and the second encoded bit sequence together to obtain uplink control information.

[0092] In one embodiment, the method of combining the first coded bit sequence and the second coded bit sequence to obtain uplink control information can be: concatenating the first coded bit sequence and the second coded bit sequence, and adding at least one of length identifier, separator identifier and check information during the concatenation process to generate complete uplink control information.

[0093] Through the above steps, the first coded bit sequence and the second coded bit sequence can be packaged to obtain uplink control information, thereby constructing a complete feedback data structure and improving the reliability and efficiency of data transmission.

[0094] Optionally, Figure 7 This is a schematic diagram of the structure of uplink control information provided in an embodiment of this application. (Reference) Figure 7 The uplink control information 100 includes a first field 110 and a second field 120.

[0095] The first field 110 contains channel rank indication information 111, channel quality indication information 112, scaling factor 113, and quantization precision 114. The second field 120 contains a bit sequence 121 and a polar code parity bit 122. The channel rank indication information 111 indicates the number of spatial multiplexing layers supported by the current channel; the channel quality indication information 112 characterizes the transmission quality level of the current channel; the scaling factor 113 is used to perform amplitude normalization processing on the channel feature vector; the quantization precision 114 indicates the number of quantization levels used when quantizing the channel feature vector; the bit sequence 121 represents the discrete bit data obtained after quantization mapping of the channel feature vector; and the polar code parity bit 122 is used for redundancy check information generated after channel coding of the bit sequence.

[0096] Based on the above embodiments, Figure 8 This is a structural block diagram of a channel state information feedback device provided in an embodiment of this application. (Reference) Figure 8 The channel state information feedback device provided in this embodiment specifically includes: a channel estimation module 21, a feature compression module 22, and an information feedback module 23.

[0097] The channel estimation module 21 is configured to receive downlink reference signals sent by the base station, perform channel estimation based on the downlink reference signals, and obtain the channel matrix of the downlink channel. The feature compression module 22 is configured to input the channel matrix into the trained encoder model for feature compression to obtain a first feature vector. The information feedback module 23 is configured to perform normalization and encapsulation processing on the first feature vector to generate uplink control information, and feed back the uplink control information to the base station for the base station to reconstruct the corresponding channel matrix based on the uplink control information.

[0098] Based on the above embodiments, the information feedback module 23 includes: a normalization unit configured to determine a scaling factor based on the first feature vector, and normalize the first feature vector based on the scaling factor to obtain a second feature vector; a field generation unit configured to generate a first field based on the scaling factor and the channel matrix, and generate a second field based on the second feature vector; and a packet assembly unit configured to assemble the first field and the second field into a packet to generate uplink control information.

[0099] Based on the above embodiments, the normalization unit includes: an amplitude calculation subunit, configured to perform amplitude calculation on each element in the first feature vector to obtain the amplitude corresponding to each element; and a scaling factor subunit, configured to determine the maximum value among the plurality of amplitudes as a scaling factor.

[0100] Based on the above embodiments, the field generation unit includes: a quantization precision subunit configured to obtain the quantization precision configured by the base station; an effective rank subunit configured to calculate the effective rank of the channel matrix to obtain the channel rank indication information of the downlink channel; a channel quality subunit configured to determine the channel quality indication information of the downlink channel based on the channel matrix; and an information combination subunit configured to combine the channel rank indication information, the channel quality indication information, the scaling factor, and the quantization precision to generate a first field.

[0101] Based on the above embodiments, the field generation unit includes: a quantization precision subunit, configured to obtain the quantization precision configured by the base station; a vector mapping subunit, configured to map the second feature vector to a bit sequence corresponding to the quantization precision; and a sequence encapsulation subunit, configured to encapsulate the bit sequence to generate a second field.

[0102] Based on the above embodiments, the packet assembly unit includes: a first coding subunit configured to perform channel coding processing on the first field to obtain a first coded bit sequence with a first code rate; a second coding subunit configured to perform channel coding processing on the second field to obtain a second coded bit sequence with a second code rate, wherein the first code rate is lower than the second code rate; and a sequence packet assembly subunit configured to assemble the first coded bit sequence and the second coded bit sequence into a packet to obtain uplink control information.

[0103] Based on the above embodiments, the second coding subunit includes: a bit concatenation component configured to concatenate each bit in the second field to obtain a bit field; and a polar coding component configured to perform channel coding processing on the bit field using polar codes to obtain a second coded bit sequence with a second code rate.

[0104] The channel state information feedback device provided in this application embodiment, by constructing a hierarchical processing channel state information feedback system composed of a channel estimation module 21, a feature compression module 22, and an information feedback module 23, achieves full-process processing capabilities for downlink channel sensing, feature compression expression, and uplink feedback generation. This ensures that the terminal device can efficiently express and accurately reconstruct high-dimensional channel information under limited feedback overhead constraints, thereby improving the accuracy and transmission efficiency of channel state information feedback. The channel estimation module 21 undertakes the channel sensing and modeling tasks. It is configured to receive the downlink reference signal sent by the base station at a set time-frequency resource location, obtain the local reference signal corresponding to the downlink reference signal, and perform complex division operation on the received signal and the local reference signal to obtain the channel matrix characterizing the downlink channel response relationship. This channel matrix is ​​used to characterize the channel gain and phase characteristics between the transmitter and receiver in each subcarrier and each antenna dimension, providing basic input for subsequent feature extraction. Feature compression module 22, designed for compact representation of high-dimensional channel information, inputs the channel matrix into a pre-trained encoder model. Through nonlinear mapping and dimensionality compression mechanisms within the model, it extracts features and compresses information from the channel matrix to obtain a first feature vector. This first feature vector characterizes the main features of the original channel matrix in a low-dimensional space, significantly reducing the amount of feedback data while maintaining reconstruction accuracy. Information feedback module 23 performs feature normalization and feedback encapsulation. It normalizes the amplitude of the first feature vector, ensuring that the values ​​of each dimension of the feature vector meet a preset range, and performs bit-level encapsulation on the normalized feature vector to generate uplink control information. Subsequently, this uplink control information is sent to the base station via the uplink, enabling the base station to input the corresponding decoder model based on the uplink control information and reconstruct the channel reconstruction result corresponding to the original channel matrix. By leveraging the channel matrix acquisition in the channel estimation module, the low-dimensional feature representation in the feature compression module, and the normalization and encapsulation feedback coordination in the information feedback module, this embodiment can achieve efficient compression and reliable feedback of high-dimensional channel information in scenarios with rapidly changing channels and limited feedback resources. This effectively reduces feedback overhead and improves channel reconstruction accuracy, providing refined channel state support for adaptive link scheduling and beam management on the base station side, thereby achieving overall optimization of channel state information feedback performance.

[0105] The channel state information feedback device provided in this application embodiment can be used to execute the channel state information feedback method provided in the above embodiment, and has corresponding functions and beneficial effects.

[0106] Figure 9 This is a schematic diagram of the structure of a channel state information feedback device provided in an embodiment of this application, with reference to... Figure 9The channel status information feedback device includes a processor 31, a memory 32, a communication device 33, an input device 34, and an output device 35. The number of processors 31 and the number of memories 32 in the channel status information feedback device can be one or more. The processor 31, memory 32, communication device 33, input device 34, and output device 35 of the channel status information feedback device can be connected via a bus or other means.

[0107] The memory 32, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as program instructions / modules corresponding to the channel state information feedback method in any embodiment of this application (e.g., channel estimation module 21, feature compression module 22, and information feedback module 23 in the channel state information feedback device). The memory 32 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the device, etc. Furthermore, the memory 32 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0108] The communication device 33 is used for data transmission.

[0109] The processor 31 executes various functional applications and data processing of the device by running software programs, instructions and modules stored in the memory 32, thereby realizing the channel state information feedback method described above.

[0110] Input device 34 can be used to receive input digital or character information, and to generate key signal inputs related to user settings and function control of the device. Output device 35 may include display devices such as a display screen.

[0111] The channel state information feedback device provided above can be used to execute the channel state information feedback method provided in the above embodiments, and has corresponding functions and beneficial effects.

[0112] This application embodiment also provides a storage medium containing computer-executable instructions. When executed by a computer processor, the computer-executable instructions are used to execute a channel state information feedback method. The channel state information feedback method includes: receiving a downlink reference signal sent by a base station; performing channel estimation based on the downlink reference signal to obtain a channel matrix of the downlink channel; inputting the channel matrix into a trained encoder model for feature compression to obtain a first feature vector; performing normalization and encapsulation processing on the first feature vector to generate uplink control information; and feeding back the uplink control information to the base station, so that the base station can reconstruct the corresponding channel matrix based on the uplink control information.

[0113] Storage medium—any type of memory device or storage device. The term "storage medium" is intended to include: mounting media, such as CD-ROM, floppy disk, or magnetic tape devices; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, etc.; non-volatile memory, such as flash memory, magnetic media (e.g., hard disk or optical storage); registers or other similar types of memory elements, etc. Storage medium may also include other types of memory or combinations thereof. Furthermore, storage medium may reside in a first computer system in which a program is executed, or it may reside in a different second computer system connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media residing in different locations (e.g., in different computer systems connected via a network). Storage medium may store program instructions (e.g., specifically implemented as a computer program) executable by one or more processors.

[0114] Of course, the computer-executable instructions provided in the embodiments of this application are not limited to the channel state information feedback method described above, but can also execute related operations in the channel state information feedback method provided in any embodiment of this application.

[0115] The channel state information feedback device, storage medium, and channel state information feedback equipment provided in the above embodiments can execute the channel state information feedback method provided in any embodiment of this application. For technical details not described in detail in the above embodiments, please refer to the channel state information feedback method provided in any embodiment of this application.

[0116] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of this application. The scope of this application is determined by the scope of the claims.

Claims

1. A channel state information feedback method, applied to a terminal device, characterized in that, include: The downlink reference signal sent by the base station is received, and channel estimation is performed based on the downlink reference signal to obtain the channel matrix of the downlink channel. The channel matrix is ​​input into the trained encoder model for feature compression to obtain the first feature vector; The first feature vector is normalized and encapsulated to generate uplink control information, which is then fed back to the base station. The base station uses the uplink control information to reconstruct the corresponding channel matrix.

2. The channel state information feedback method according to claim 1, characterized in that, The normalization and encapsulation processing of the first feature vector to generate uplink control information includes: A scaling factor is determined based on the first feature vector, and the first feature vector is normalized based on the scaling factor to obtain a second feature vector. A first field is generated based on the scaling factor and the channel matrix, and a second field is generated based on the second feature vector; The first field and the second field are combined to generate uplink control information.

3. The channel state information feedback method according to claim 2, characterized in that, Determining the scaling factor based on the first feature vector includes: Perform amplitude calculation on each element in the first feature vector to obtain the amplitude corresponding to each element; The maximum value among the multiple amplitudes is determined as the scaling factor.

4. The channel state information feedback method according to claim 2, characterized in that, The step of generating the first field based on the scaling factor and the channel matrix includes: Obtain the quantization precision configured for the base station; Calculate the effective rank of the channel matrix to obtain the channel rank indication information of the downlink channel; The channel quality indication information of the downlink channel is determined based on the channel matrix; The channel rank indication information, the channel quality indication information, the scaling factor, and the quantization precision are combined to generate the first field.

5. The channel state information feedback method according to claim 2, characterized in that, The step of generating the second field based on the second feature vector includes: Obtain the quantization precision configured for the base station; Map the second feature vector to the bit sequence corresponding to the quantization precision; The bit sequence is encapsulated to generate a second field.

6. The channel state information feedback method according to claim 2, characterized in that, The step of combining the first field and the second field to generate uplink control information includes: Perform channel coding processing on the first field to obtain a first coded bit sequence with a first code rate; Channel coding processing is performed on the second field to obtain a second coded bit sequence with a second code rate, wherein the first code rate is lower than the second code rate; The first encoded bit sequence and the second encoded bit sequence are combined to obtain uplink control information.

7. The channel state information feedback method according to claim 6, characterized in that, The process of performing channel coding on the second field to obtain a second coded bit sequence with a second code rate includes: The bits in the second field are concatenated to obtain the bit field; The bit field is channel-coded using polar coding to obtain a second coded bit sequence with a second code rate.

8. A channel state information feedback device, applied to a terminal device, characterized in that, include: The channel estimation module is configured to receive downlink reference signals transmitted by the base station, perform channel estimation based on the downlink reference signals, and obtain the channel matrix of the downlink channel. The feature compression module is configured to input the channel matrix into the trained encoder model for feature compression to obtain a first feature vector; The information feedback module is configured to perform normalization and encapsulation processing on the first feature vector to generate uplink control information, and to feed back the uplink control information to the base station, so that the base station can restore the corresponding channel matrix based on the uplink control information.

9. A channel state information feedback device, characterized in that, include: One or more processors; A memory that stores one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the channel state information feedback method as described in any one of claims 1-7.

10. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the channel state information feedback method as described in any one of claims 1-7.

Images