A method, apparatus, device and medium for channel estimation of a MIMO system

Abstract

The application discloses a MIMO system channel estimation method, device, equipment and medium, the method comprises the following steps: obtaining the channel state information matrix of each communication terminal based on the signals including superimposed training sequences received from a plurality of communication terminals; obtaining the approximate estimation matrix corresponding to each channel state information matrix based on the calculation strategy of minimum channel estimation error variance, using an encoder to extract features and obtain the element feature map of each element; respectively fusing the element feature map of the i-th element in a plurality of approximate estimation matrices to obtain a fused feature map; calculating the weight value of each feature map layer in the fused feature map by using the L2 norm, and updating each element in a plurality of approximate estimation matrices according to the comparison result of the weight value of each feature map layer and the preset weight threshold to obtain a simplified channel estimation matrix. The application can effectively describe the change of channel fading and reduce the complexity of channel estimation.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to a channel estimation method, apparatus, terminal equipment, and computer-readable storage medium for a MIMO system. Background Technology

[0002] With the development of 5G, there may be hundreds of paths in the communication channel. Many of these paths appear in clusters. Existing tapped delay line models, clustered delay line models, and Rayleigh fading models can no longer effectively describe the changes in channel fading, resulting in low channel estimation accuracy for MIMO systems. Furthermore, with the evolution of millimeter-wave technology, there are increasingly more paths in the channel, and these paths undergo reflection, diffraction, and other effects, multiplying the number of paths received by the receiver. In this situation, existing technologies require a significant amount of computational resources to decode the signals received by the terminal at the base station. Summary of the Invention

[0003] This invention provides a channel estimation method, apparatus, device, and medium for MIMO systems. By basing the method on the fact that the channel follows a Rice distribution and using an approximate estimation matrix corresponding to the channel state information matrix of each communication terminal obtained from signals received from several communication terminals, including superimposed training sequences, the method can effectively describe the changes in channel fading, thereby improving the accuracy of channel estimation. Furthermore, by simplifying the obtained approximate estimation matrices, the complexity of channel estimation can be reduced, significantly reducing the computational resources required during the decoding process.

[0004] To address the aforementioned technical problems, a first aspect of the present invention provides a channel estimation method for a MIMO system, comprising the following steps:

[0005] Based on signals received from several communication terminals, including superimposed training sequences, a channel state information matrix for each of the communication terminals is obtained; wherein, the superimposed training sequence includes a training sequence and a communication data sequence;

[0006] Based on the calculation strategy of minimizing the variance of the channel estimation error, an approximate estimation matrix corresponding to each of the channel state information matrices is obtained;

[0007] The preset encoder is used to extract features from each of the approximate estimation matrices to obtain the element feature map of each element in each approximate estimation matrix;

[0008] The element feature maps of the i-th element in several approximate estimation matrices are fused to obtain a fused feature map with several feature layers; where i≥1;

[0009] The weight value of each feature layer in the fused feature map is calculated using the L2 paradigm, and each element in several approximate estimation matrices is updated based on the comparison result of the weight value of each feature layer with a preset weight threshold to obtain a simplified channel estimation matrix.

[0010] As a preferred embodiment, the step of updating each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with a preset weight threshold to obtain a simplified channel estimation matrix specifically includes the following steps:

[0011] If the weight value of the i-th feature layer is greater than the preset weight threshold, then the fused feature corresponding to the i-th feature layer is used to update the i-th element in the plurality of approximate estimation matrices. If the weight value of the i-th feature layer is less than or equal to the preset weight threshold, then the i-th element in the plurality of approximate estimation matrices is set to 0, until each element in the plurality of approximate estimation matrices has been updated, and the simplified channel estimation matrix is ​​obtained.

[0012] As a preferred embodiment, the step of calculating the weight value of each feature layer in the fused feature map using the L2 paradigm specifically includes the following steps:

[0013] The following expression is calculated to make the feature similarity between the original feature map corresponding to the approximate estimation matrix and the fused feature map converge to zero, thereby obtaining the pixel weight set:

[0014]

[0015] Wherein, the original feature map includes the element feature map of each element in the approximate estimation matrix; the pixel weight set includes the weight value of each feature layer in the fused feature map;

[0016] L sim K represents the feature similarity; K represents the number of approximate estimation matrices; This represents the fused feature map; The original feature map represents the approximate estimation matrix corresponding to the k-th communication terminal; The L2 norm of the original feature map is represented by the approximate estimation matrix corresponding to the k-th communication terminal and the fused feature map.

[0017] As a preferred embodiment, obtaining the channel state information matrix of each communication terminal based on signals received from several communication terminals, including superimposed training sequences, specifically includes the following steps:

[0018] Based on signals received from several communication terminals, including superimposed training sequences, the channel state information matrix of each communication terminal is obtained through the following expression:

[0019]

[0020] in, This represents the channel state information matrix; Y This indicates a signal that includes superimposed training sequences; Represents the conjugate transpose of the training sequence; This represents the second preset constant; This indicates the preset average received signal-to-noise ratio; T Indicates the coherence time.

[0021] As a preferred embodiment, the calculation strategy based on minimizing the variance of the channel estimation error to obtain the approximate estimation matrix corresponding to each channel state information matrix specifically includes the following steps:

[0022] An optimization problem based on minimizing the variance of channel estimation error: The matrix is ​​obtained by calculating the following expression. A :

[0023]

[0024] According to the matrix A For each of the channel state information matrices, an approximate estimation matrix corresponding to each channel state information matrix is ​​calculated using the following expression:

[0025]

[0026] in, This represents the approximate estimation matrix; B Represents a large-scale fading sparse diagonal matrix of size LK×LK; tr ( B 2 ) represents the sum of the diagonal elements of the matrix obtained by multiplying the large-scale fading sparse diagonal matrix by itself; I LK L represents an L×K dimensional diagonal identity matrix; LK represents the number of signals transmitted simultaneously by K users to L cells in a MIMO system.

[0027] As a preferred embodiment, the signal comprising the superimposed training sequence is specifically represented as follows:

[0028]

[0029] in, This represents the first preset constant; GThis represents the product of an M×M dimensional small-scale fading sparse matrix and an LK×LK dimensional large-scale fading sparse diagonal matrix. S This represents the communication data sequence; P This refers to the training sequence; W The preset Gaussian white noise is represented; the sum of the first preset constant and the second preset constant is 1.

[0030] A second aspect of the present invention provides a channel estimation apparatus for a MIMO system, comprising:

[0031] A channel state information matrix acquisition module is used to obtain a channel state information matrix for each of the communication terminals based on signals received from a plurality of communication terminals, including superimposed training sequences; wherein the superimposed training sequences include training sequences and communication data sequences.

[0032] The approximate estimation matrix acquisition module is used to obtain the approximate estimation matrix corresponding to each of the channel state information matrices based on the calculation strategy of minimizing the variance of the channel estimation error.

[0033] The feature extraction module is used to extract features from each of the approximate estimation matrices using a preset encoder, and to obtain the element feature map of each element in each of the approximate estimation matrices.

[0034] The feature fusion module is used to fuse the element feature maps of the i-th element in several approximate estimation matrices to obtain a fused feature map with several feature layers; where i≥1;

[0035] The simplified channel estimation matrix acquisition module is used to calculate the weight value of each feature layer in the fused feature map using the L2 paradigm, and update each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with a preset weight threshold to obtain the simplified channel estimation matrix.

[0036] As a preferred embodiment, the simplified channel estimation matrix acquisition module is used to update each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with a preset weight threshold, to obtain a simplified channel estimation matrix, specifically including:

[0037] If the weight value of the i-th feature layer is greater than the preset weight threshold, then the fused feature corresponding to the i-th feature layer is used to update the i-th element in the plurality of approximate estimation matrices. If the weight value of the i-th feature layer is less than or equal to the preset weight threshold, then the i-th element in the plurality of approximate estimation matrices is set to 0, until each element in the plurality of approximate estimation matrices has been updated, and the simplified channel estimation matrix is ​​obtained.

[0038] A third aspect of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the MIMO system channel estimation method as described in any of the first aspects.

[0039] A fourth aspect of the present invention provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the MIMO system channel estimation method as described in any of the first aspects.

[0040] Compared with the prior art, the beneficial effects of the embodiments of the present invention are that, based on the fact that the channel follows a Rice distribution, and based on the approximate estimation matrix corresponding to the channel state information matrix of each communication terminal obtained from signals received from several communication terminals including superimposed training sequences, the changes in channel fading can be effectively described, thereby improving the accuracy of channel estimation; in addition, by simplifying the obtained approximate estimation matrices, the complexity of channel estimation can be reduced, and the computational resources required in the decoding process can be significantly reduced. Attached Figure Description

[0041] Figure 1 This is a flowchart illustrating the channel estimation method for a MIMO system in an embodiment of the present invention.

[0042] Figure 2 This is a schematic diagram of the structure of the MIMO system channel estimation device in an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] See Figure 1 The first aspect of this invention provides a channel estimation method for a MIMO system, comprising the following steps S1 to S5:

[0045] Step S1: Based on signals received from several communication terminals, including superimposed training sequences, obtain a channel state information matrix for each of the communication terminals; wherein, the superimposed training sequence includes a training sequence and a communication data sequence;

[0046] Step S2: Based on the calculation strategy of minimizing the variance of the channel estimation error, obtain the approximate estimation matrix corresponding to each of the channel state information matrices;

[0047] Step S3: Use a preset encoder to extract features from each of the approximate estimation matrices to obtain the element feature map of each element in each of the approximate estimation matrices;

[0048] Step S4: The element feature maps of the i-th element in several approximate estimation matrices are fused to obtain a fused feature map with several feature layers; where i≥1;

[0049] Step S5: Calculate the weight value of each feature layer in the fused feature map using the L2 paradigm, and update each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with the preset weight threshold to obtain a simplified channel estimation matrix.

[0050] It is worth noting that, based on the channel following a Rice distribution, the channel estimation method is divided into two types: (1) time-division training sequence, where the communication terminal first sends a training sequence, and the base station estimates the channel based on the training sequence and the communication data sequence sent by the communication terminal, thereby achieving signal decoding; (2) superimposed training sequence, where the user terminal first superimposes the training sequence and the communication data sequence and then sends them to the base station, and the base station uses the received superimposed training sequence to estimate the channel, thereby achieving signal decoding. This embodiment of the invention is based on the channel following a Rice distribution, dividing the channel estimation into LOS components and random variables of the inference distribution of the dispersed multipath component signal NLOS. In the case of unknown LOS components, an approximate estimation matrix corresponding to the channel state information matrix of each communication terminal is obtained.

[0051] In steps S1 and S2, the channel state information matrix of each communication terminal is first obtained based on the signals received from several communication terminals, including superimposed training sequences. In order to further improve the accuracy of channel estimation, an approximate estimation matrix corresponding to each channel state information matrix is ​​obtained based on the calculation strategy of minimizing the variance of channel estimation error.

[0052] Assume that the approximate estimation matrix corresponding to the channel state information matrix of each communication terminal is as follows: The approximate estimation matrix for different communication terminals is represented as follows:

[0053]

[0054] In step S3, this embodiment considers that the approximate estimation matrix corresponding to each channel state information matrix may fluctuate to a certain extent due to random errors caused by differences in communication terminal chips when the communication terminal sends signals to the base station. This fluctuation may have a significant impact on the signals sent by some communication terminals. Therefore, this embodiment proposes an optimization scheme for the approximate estimation matrix. First, the obtained approximate estimation matrices are used as input to a preset encoder. The encoder is used to extract features to obtain the element feature map of each element in each approximate estimation matrix. It can be understood that different elements in the approximate estimation matrix represent different channel state information.

[0055] For example, the original feature representation corresponding to the approximate estimation matrix of the k-th communication terminal is as follows:

[0056]

[0057] In step S4, the element feature maps of the i-th element in several approximate estimation matrices are fused to obtain a fused feature map with several feature layers. The feature similarity between the fused feature map and the original feature map corresponding to the approximate estimation matrix should be consistent.

[0058] For example, the fusion feature representation of several approximate estimation matrices is as follows:

[0059]

[0060] In step S5, the similarity between multiple element feature maps and the fused feature map is solved using the L2 paradigm. Finally, the weight value of each feature layer in the fused feature map is obtained, which is used to characterize the weight of each element in the approximate estimation matrix. Then, based on the comparison result of the weight value of each feature layer with the preset weight threshold, each element in several approximate estimation matrices is updated, thereby obtaining a simplified channel estimation matrix and reducing the complexity of channel estimation.

[0061] As a preferred embodiment, the step of updating each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with a preset weight threshold to obtain a simplified channel estimation matrix specifically includes the following steps:

[0062] If the weight value of the i-th feature layer is greater than the preset weight threshold, then the fused feature corresponding to the i-th feature layer is used to update the i-th element in the plurality of approximate estimation matrices. If the weight value of the i-th feature layer is less than or equal to the preset weight threshold, then the i-th element in the plurality of approximate estimation matrices is set to 0, until each element in the plurality of approximate estimation matrices has been updated, and the simplified channel estimation matrix is ​​obtained.

[0063] Specifically, if the weight value of the i-th feature layer is greater than the preset weight threshold, it indicates that the i-th element in the approximate estimation matrix is ​​an important element. Therefore, the i-th element in several approximate estimation matrices is updated to the fused feature corresponding to the i-th feature layer. If i=1, referring to the expression for the fused feature above, the first element in several approximate estimation matrices is updated to the fused feature corresponding to the first feature layer. If the weight value of the i-th feature layer is less than or equal to the preset weight threshold, it indicates that the i-th element in the approximate estimation matrix is ​​an unimportant element. Therefore, the i-th element in several approximate estimation matrices is set to 0 until each element in several approximate estimation matrices is updated, and a simplified channel estimation matrix is ​​obtained.

[0064] As a preferred embodiment, the step of calculating the weight value of each feature layer in the fused feature map using the L2 paradigm specifically includes the following steps:

[0065] The following expression is calculated to make the feature similarity between the original feature map corresponding to the approximate estimation matrix and the fused feature map converge to zero, thereby obtaining the pixel weight set:

[0066]

[0067] Wherein, the original feature map includes the element feature map of each element in the approximate estimation matrix; the pixel weight set includes the weight value of each feature layer in the fused feature map;

[0068] L sim K represents the feature similarity; K represents the number of approximate estimation matrices; This represents the fused feature map; The original feature map represents the approximate estimation matrix corresponding to the k-th communication terminal; The L2 norm of the original feature map representing the approximate estimation matrix corresponding to the fused feature map and the k-th communication terminal; This represents the preset hyperparameters.

[0069] It is worth noting that the expression is divided into three parts. The first part represents the approximate estimation matrix of K communication terminals. If the original feature map of the approximate estimation matrix of each communication terminal has the same similarity to the fused feature map (the second part), then the cumulative value of the K similarities in the second part is 1. The third part is the L2 norm. In order to avoid overfitting, the L2 norm is needed to limit the error between the feature maps before and after fusion. Therefore, the goal of the whole expression is to make the cumulative value of the K similarities in the second part as close to 1 as possible, and to make the error between the feature maps before and after fusion as small as possible.

[0070] As a preferred embodiment, obtaining the channel state information matrix of each communication terminal based on signals received from several communication terminals, including superimposed training sequences, specifically includes the following steps:

[0071] Based on signals received from several communication terminals, including superimposed training sequences, the channel state information matrix of each communication terminal is obtained through the following expression:

[0072]

[0073] in, This represents the channel state information matrix; Y This indicates a signal that includes superimposed training sequences; Represents the conjugate transpose of the training sequence; This represents the second preset constant; This indicates the preset average received signal-to-noise ratio; T Indicates the coherence time.

[0074] As a preferred embodiment, the calculation strategy based on minimizing the variance of the channel estimation error to obtain the approximate estimation matrix corresponding to each channel state information matrix specifically includes the following steps:

[0075] An optimization problem based on minimizing the variance of channel estimation error: The matrix is ​​obtained by calculating the following expression. A :

[0076]

[0077] According to the matrix A For each of the channel state information matrices, an approximate estimation matrix corresponding to each channel state information matrix is ​​calculated using the following expression:

[0078]

[0079] in, This represents the approximate estimation matrix; B Represents a large-scale fading sparse diagonal matrix of size LK×LK; tr ( B 2 ) represents the sum of the diagonal elements of the matrix obtained by multiplying the large-scale fading sparse diagonal matrix by itself; I LK L represents an L×K dimensional diagonal identity matrix; LK represents the number of signals transmitted simultaneously by K users to L cells in a MIMO system.

[0080] As a preferred embodiment, the signal comprising the superimposed training sequence is specifically represented as follows:

[0081]

[0082] in, This represents the first preset constant; G This represents the product of an M×M dimensional small-scale fading sparse matrix and an LK×LK dimensional large-scale fading sparse diagonal matrix. S This represents the communication data sequence; P This refers to the training sequence; W The preset Gaussian white noise is represented; the sum of the first preset constant and the second preset constant is 1.

[0083] It's worth noting that in a MIMO scenario, assuming an area is divided into L cells, each cell randomly distributes K communication terminals, and each base station is equipped with M antennas, then after the superimposed training sequence X is transmitted from the communication terminal side, it undergoes large-scale fading (such as shadowing and path loss) and small-scale fading (such as multipath fading) before reaching the base station, forming the signal received by the base station. The fading signal is typically described as the channel gain between the nth user antenna and the mth base station antenna, denoted as... .in, The small-scale fading factor is set according to different scenarios. During the setting process, the small-scale fading factor changes significantly when the scenario changes. It is generally quite complex and dynamic. This represents the large-scale fading factor of the channel; this part of the data does not change over a long period of time. Generally speaking, This information is usually known to the base station as a constant. Combined with the average received signal-to-noise ratio... Given Gaussian white noise W, the signal received by the base station is represented as follows:

[0084]

[0085] in, H The M×M dimensional small-scale fading sparse matrix can be further decomposed into: H LOS and H NLOS ,in, H LOS Represented as the direct wave channel response, H NLOS The multipath component channel response is expressed as follows:

[0086]

[0087] It is the Rice fading factor, defined as the ratio of the power of the direct component to the power of the multipath component, and is usually easy to obtain.

[0088] X Let LK represent the transmitted signal in LK×T dimensions, where the user's signal is a vector of length T×1, and LK represents the number of signals transmitted simultaneously by K communication terminals to L cells in a MIMO system. Y This represents the M×T dimensional signal received by a base station equipped with M antennas. B This represents the LK×LK dimensional large-scale fading sparse diagonal matrix. Typically, the large-scale fading sparse diagonal matrix can be pre-measured and remains essentially unchanged. B They are considered as known variables. To simplify formula (1), this embodiment assumes... G = HB Then the signal received by the base station is represented as:

[0089]

[0090] Due to signal X It consists of a training sequence and a communication data sequence, which transmits the signal. X After being split into training sequences and communication data sequences, the signal received by the base station can be represented as:

[0091]

[0092] It is worth noting that in formula (3), the first term on the right side of the equation represents the communication signal, the second term represents the training signal, and the third term represents the error caused by noise.

[0093] training sequence P satisfy:

[0094]

[0095] Among them, training sequences P Let be a T-dimensional Hadamard matrix, and be known variables.

[0096] Combining formula (4), in large-scale MIMO systems, since the channel vectors are asymptotically orthogonal, therefore, , After simplifying formula (3), the channel state information matrix is ​​approximately expressed as:

[0097]

[0098] Furthermore, to improve the accuracy of channel estimation, multiple samplings can yield an approximate estimation matrix corresponding to the channel state information matrix:

[0099]

[0100] An optimization problem based on minimizing the variance of channel estimation error: The solution to this optimization problem is obtained using the following expression:

[0101]

[0102] When the number of antennas M approaches infinity, the direct wave channel response H LOS Having deterministic characteristics, in this case, .

[0103] The present invention provides a channel estimation method for a MIMO system. Based on the assumption that the channel follows a Rice distribution, and using an approximate estimation matrix corresponding to the channel state information matrix of each communication terminal obtained from signals received from several communication terminals, including superimposed training sequences, the method can effectively describe the changes in channel fading, thereby improving the accuracy of channel estimation. Furthermore, by simplifying the obtained approximate estimation matrices, the complexity of channel estimation can be reduced, significantly reducing the computational resources required during the decoding process.

[0104] See Figure 2 A second aspect of the present invention provides a channel estimation apparatus for a MIMO system, comprising:

[0105] The channel state information matrix acquisition module 201 is used to obtain the channel state information matrix of each of the communication terminals based on signals received from a plurality of communication terminals, including superimposed training sequences; wherein the superimposed training sequences include training sequences and communication data sequences.

[0106] The approximate estimation matrix acquisition module 202 is used to obtain the approximate estimation matrix corresponding to each of the channel state information matrices based on the calculation strategy of minimizing the variance of the channel estimation error.

[0107] The feature extraction module 203 is used to extract features from each of the approximate estimation matrices using a preset encoder, and to obtain the element feature map of each element in each of the approximate estimation matrices.

[0108] The feature fusion module 204 is used to fuse the element feature maps of the i-th element in several approximate estimation matrices to obtain a fused feature map with several feature layers; where i≥1;

[0109] The simplified channel estimation matrix acquisition module 205 is used to calculate the weight value of each feature layer in the fused feature map using the L2 paradigm, and update each element in several approximate estimation matrices according to the comparison result of the weight value of each feature layer with a preset weight threshold to obtain the simplified channel estimation matrix.

[0110] As a preferred embodiment, the simplified channel estimation matrix acquisition module 205 is used to update each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with a preset weight threshold, to obtain a simplified channel estimation matrix, specifically including:

[0111] If the weight value of the i-th feature layer is greater than the preset weight threshold, then the fused feature corresponding to the i-th feature layer is used to update the i-th element in the plurality of approximate estimation matrices. If the weight value of the i-th feature layer is less than or equal to the preset weight threshold, then the i-th element in the plurality of approximate estimation matrices is set to 0, until each element in the plurality of approximate estimation matrices has been updated, and the simplified channel estimation matrix is ​​obtained.

[0112] As a preferred embodiment, the simplified channel estimation matrix acquisition module 205 is used to calculate the weight value of each feature layer in the fused feature map using the L2 paradigm, specifically including:

[0113] The following expression is calculated to make the feature similarity between the original feature map corresponding to the approximate estimation matrix and the fused feature map converge to zero, thereby obtaining the pixel weight set:

[0114]

[0115] Wherein, the original feature map includes the element feature map of each element in the approximate estimation matrix; the pixel weight set includes the weight value of each feature layer in the fused feature map;

[0116] L sim K represents the feature similarity; K represents the number of approximate estimation matrices; This represents the fused feature map; The original feature map represents the approximate estimation matrix corresponding to the k-th communication terminal; The L2 norm of the original feature map is represented by the approximate estimation matrix corresponding to the k-th communication terminal and the fused feature map.

[0117] As a preferred embodiment, the channel state information matrix acquisition module 201 is used to obtain the channel state information matrix of each of the communication terminals based on signals received from several communication terminals, including superimposed training sequences, specifically including:

[0118] Based on signals received from several communication terminals, including superimposed training sequences, the channel state information matrix of each communication terminal is obtained through the following expression:

[0119]

[0120] in, This represents the channel state information matrix; Y This indicates a signal that includes superimposed training sequences; Represents the conjugate transpose of the training sequence; This represents the second preset constant; This indicates the preset average received signal-to-noise ratio; T Indicates the coherence time.

[0121] As a preferred embodiment, the approximate estimation matrix acquisition module 202 is used to obtain the approximate estimation matrix corresponding to each of the channel state information matrices based on a calculation strategy that minimizes the variance of the channel estimation error, specifically including:

[0122] An optimization problem based on minimizing the variance of channel estimation error: The matrix is ​​obtained by calculating the following expression. A :

[0123]

[0124] According to the matrix A For each of the channel state information matrices, an approximate estimation matrix corresponding to each channel state information matrix is ​​calculated using the following expression:

[0125]

[0126] in, This represents the approximate estimation matrix; B Represents a large-scale fading sparse diagonal matrix of size LK×LK; tr ( B 2 ) represents the sum of the diagonal elements of the matrix obtained by multiplying the large-scale fading sparse diagonal matrix by itself; I LK L represents an L×K dimensional diagonal identity matrix; LK represents the number of signals transmitted simultaneously by K users to L cells in a MIMO system.

[0127] As a preferred embodiment, the signal comprising the superimposed training sequence is specifically represented as follows:

[0128]

[0129] in, This represents the first preset constant; G This represents the product of an M×M dimensional small-scale fading sparse matrix and an LK×LK dimensional large-scale fading sparse diagonal matrix. S This represents the communication data sequence; P This refers to the training sequence;W The preset Gaussian white noise is represented; the sum of the first preset constant and the second preset constant is 1.

[0130] It should be noted that the MIMO system channel estimation device provided in this embodiment of the invention can implement all the processes of the MIMO system channel estimation method described in any of the above embodiments. The functions and technical effects of each module in the device are the same as those of the MIMO system channel estimation method described in the above embodiments, and will not be repeated here.

[0131] A third aspect of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the MIMO system channel estimation method as described in any embodiment of the first aspect.

[0132] A third aspect of the present invention provides a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the real-time video AR interaction method as described in any embodiment of the first aspect.

[0133] The terminal device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and memory. The terminal device may also include input / output devices, network access devices, buses, etc.

[0134] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0135] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the terminal device by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0136] A fourth aspect of the present invention provides a computer-readable storage medium comprising a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform a MIMO system channel estimation method as described in any embodiment of the first aspect.

[0137] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus necessary hardware platforms, and of course, it can also be implemented entirely by hardware. Based on this understanding, all or part of the technical solution of the present invention that contributes to the background art can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0138] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A channel estimation method for a MIMO system, characterized in that, Includes the following steps: Based on signals received from several communication terminals, including superimposed training sequences, a channel state information matrix for each of the communication terminals is obtained; wherein, the superimposed training sequence includes a training sequence and a communication data sequence; Based on the calculation strategy of minimizing the variance of the channel estimation error, an approximate estimation matrix corresponding to each of the channel state information matrices is obtained; The preset encoder is used to extract features from each of the approximate estimation matrices to obtain the element feature map of each element in each approximate estimation matrix; The element feature maps of the i-th element in several approximate estimation matrices are fused to obtain a fused feature map with several feature layers; where i≥1; The weight value of each feature layer in the fused feature map is calculated using the L2 paradigm, and each element in several approximate estimation matrices is updated based on the comparison result of the weight value of each feature layer with a preset weight threshold to obtain a simplified channel estimation matrix.

2. The MIMO system channel estimation method as described in claim 1, characterized in that, The step of updating each element in several approximate estimation matrices based on the comparison result between the weight value of each feature layer and the preset weight threshold to obtain a simplified channel estimation matrix specifically includes the following steps: If the weight value of the i-th feature layer is greater than the preset weight threshold, then the fused feature corresponding to the i-th feature layer is used to update the i-th element in the plurality of approximate estimation matrices. If the weight value of the i-th feature layer is less than or equal to the preset weight threshold, then the i-th element in the plurality of approximate estimation matrices is set to 0, until each element in the plurality of approximate estimation matrices has been updated, and the simplified channel estimation matrix is ​​obtained.

3. The MIMO system channel estimation method as described in claim 1, characterized in that, The step of calculating the weight value of each feature layer in the fused feature map using the L2 paradigm specifically includes the following steps: The following expression is calculated to make the feature similarity between the original feature map corresponding to the approximate estimation matrix and the fused feature map converge to zero, thereby obtaining the pixel weight set: Wherein, the original feature map includes the element feature map of each element in the approximate estimation matrix; the pixel weight set includes the weight value of each feature layer in the fused feature map; L sim K represents the feature similarity; K represents the number of approximate estimation matrices; This represents the fused feature map; The original feature map represents the approximate estimation matrix corresponding to the k-th communication terminal; The L2 norm of the original feature map is represented by the approximate estimation matrix corresponding to the k-th communication terminal and the fused feature map.

4. The MIMO system channel estimation method as described in claim 1, characterized in that, The step of obtaining the channel state information matrix for each communication terminal based on signals received from several communication terminals, including superimposed training sequences, specifically includes the following steps: Based on signals received from several communication terminals, including superimposed training sequences, the channel state information matrix of each communication terminal is obtained through the following expression: in, This represents the channel state information matrix; Y This indicates a signal that includes superimposed training sequences; Represents the conjugate transpose of the training sequence; This represents the second preset constant; This indicates the preset average received signal-to-noise ratio; T Indicates the coherence time.

5. The MIMO system channel estimation method as described in claim 4, characterized in that, The calculation strategy based on minimizing the variance of the channel estimation error, which obtains the approximate estimation matrix corresponding to each channel state information matrix, specifically includes the following steps: An optimization problem based on minimizing the variance of channel estimation error: The matrix is ​​obtained by calculating the following expression. A : According to the matrix A For each of the channel state information matrices, an approximate estimation matrix corresponding to each channel state information matrix is ​​calculated using the following expression: in, This represents the approximate estimation matrix; B Represents a large-scale fading sparse diagonal matrix of size LK×LK; tr ( B 2 ) represents the sum of the diagonal elements of the matrix obtained by multiplying the large-scale fading sparse diagonal matrix by itself; I LK L represents an L×K dimensional diagonal identity matrix; LK represents the number of signals transmitted simultaneously by K users to L cells in a MIMO system.

6. The MIMO system channel estimation method as described in claim 4, characterized in that, The signal including the superimposed training sequence is specifically represented as follows: in, This represents the first preset constant; G This represents the product of an M×M dimensional small-scale fading sparse matrix and an LK×LK dimensional large-scale fading sparse diagonal matrix. S This represents the communication data sequence; P This refers to the training sequence; W The preset Gaussian white noise is represented; the sum of the first preset constant and the second preset constant is 1.

7. A channel estimation device for a MIMO system, characterized in that, include: A channel state information matrix acquisition module is used to obtain a channel state information matrix for each of the communication terminals based on signals received from a plurality of communication terminals, including superimposed training sequences; wherein the superimposed training sequences include training sequences and communication data sequences. The approximate estimation matrix acquisition module is used to obtain the approximate estimation matrix corresponding to each of the channel state information matrices based on the calculation strategy of minimizing the variance of the channel estimation error. The feature extraction module is used to extract features from each of the approximate estimation matrices using a preset encoder, and to obtain the element feature map of each element in each of the approximate estimation matrices. The feature fusion module is used to fuse the element feature maps of the i-th element in several approximate estimation matrices to obtain a fused feature map with several feature layers; where i≥1; The simplified channel estimation matrix acquisition module is used to calculate the weight value of each feature layer in the fused feature map using the L2 paradigm, and update each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with a preset weight threshold to obtain the simplified channel estimation matrix.

8. The MIMO system channel estimation apparatus as described in claim 7, characterized in that, The simplified channel estimation matrix acquisition module is used to update each element in several approximate estimation matrices based on the comparison result of the weight value of each feature layer with a preset weight threshold, to obtain a simplified channel estimation matrix, specifically including: If the weight value of the i-th feature layer is greater than the preset weight threshold, then the fused feature corresponding to the i-th feature layer is used to update the i-th element in the plurality of approximate estimation matrices. If the weight value of the i-th feature layer is less than or equal to the preset weight threshold, then the i-th element in the plurality of approximate estimation matrices is set to 0, until each element in the plurality of approximate estimation matrices has been updated, and the simplified channel estimation matrix is ​​obtained.

9. A terminal device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the MIMO system channel estimation method as described in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein, when the computer program is executed, it controls the device on which the computer-readable storage medium is located to perform the MIMO system channel estimation method as described in any one of claims 1 to 6.

Images