RIS cooperative communication system, method, device, medium and product

Abstract

Provided are an RIS cooperative communication system, method, device, medium and product, belonging to the technical field of wireless communications, acquiring a first real-time pilot signal sent by a base station and a second real-time pilot signal sent by a user terminal, and determining a real-time channel vector according to the first real-time pilot signal and the second real-time pilot signal; inputting the real-time channel vector into a rate prediction model, and predicting an optimal de-noising information rate and a target phase shift matrix corresponding to the optimal de-noising information rate by the rate prediction model according to the real-time channel vector; adjusting the phase shifts of all RIS reflection units according to the target phase shift matrix, so that the RIS reflection units are used as relays to realize signal transmission between the base station and the user terminal; wherein the rate prediction model is obtained by training a preset model by de-noising information rates corresponding to all phase shift vectors in the RIS reflection codebook. The optimal denoising information rate and the target phase shift matrix are predicted through the rate prediction model, and the data rate of the user is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication, in particular to a RIS cooperative communication system, method, device, medium and product. BACKGROUND

[0002] Wireless communication systems can increase transmission capacity and reduce latency by introducing new radio spectrum and a series of emerging technologies, such as massive multiple-input multiple-output (MIMO), small cells, edge computing, and new channel coding techniques. The improvement of the physical layer of the wireless communication system depends largely on the control and change of the propagation environment, such as the use of antenna diversity in the transmitter and receiver, cooperative relay schemes, etc. Each surface reflection unit of the reconfigurable intelligent surface (RIS) can effectively change the signal propagation environment, and can also achieve efficient transmission using scattered waves.

[0003] The existing scheme uses an unsupervised learning method to design the hybrid precoding matrix of the base station and the RIS phase shift matrix, and does not consider the finiteness of the RIS implementable phase shift, so that the data rate of the user needs to be improved. SUMMARY

[0004] The present application provides a RIS cooperative communication system, method, device, medium and product, which solves the problem that the existing scheme uses an unsupervised learning method to design the hybrid precoding matrix of the base station and the RIS phase shift matrix, and does not consider the finiteness of the RIS implementable phase shift, so that the data rate of the user needs to be improved.

[0005] The present application provides a RIS cooperative communication system, which includes an RIS device and a user terminal, and the RIS device includes an RIS controller and an RIS reflection unit. The RIS controller is configured to obtain a first real-time pilot signal transmitted by a base station and a second real-time pilot signal transmitted by a user terminal, determine a real-time channel vector according to the first real-time pilot signal and the second real-time pilot information, input the real-time channel vector into a rate prediction model, the rate prediction model is configured to predict an optimal denoising information rate and a target phase shift matrix corresponding to the optimal denoising information rate according to the real-time channel vector, and adjust the phase shift of all RIS reflection units according to the target phase shift matrix, so that the RIS reflection units act as relays to realize signal transmission between the base station and the user terminal; wherein the rate prediction model is obtained by training a preset model with denoising information rates corresponding to all phase shift vectors in an RIS reflection codebook.

[0006] As an embodiment, the RIS controller is further configured to: obtaining a first pilot signal transmitted by the base station and a second pilot signal transmitted by the user terminal, and determining a channel vector according to the first pilot signal and the second pilot signal; obtaining, from the user terminal, a denoised information rate corresponding to any phase shift vector in the RIS reflection codebook; constructing a data set according to the channel vector and the denoised information rate; training a neural network model based on a regression loss function and the data set to obtain the rate prediction model.

[0007] As an embodiment, the regression loss function is used to represent the mean square error between the normalized expected denoised information rate and the predicted denoised information rate of the neural network model.

[0008] As an embodiment, the user terminal is configured to: receive a signal transmitted by the RIS device based on any phase shift vector in the RIS reflection codebook, obtain a received signal corresponding to the phase shift vector, and construct a super matrix according to the received signal; input the super matrix into a double-channel residual neural network to obtain a denoised received signal output by the double-channel residual neural network; determine the denoised information rate corresponding to the phase shift vector according to the denoised received signal.

[0009] As an embodiment, the user terminal is further configured to construct the super matrix according to the real part and the imaginary part of the received signal; and correspondingly, the two neural network channels of the double-channel residual neural network are respectively configured to perform denoising processing on the real part and the imaginary part of the received signal.

[0010] As an embodiment, the two neural network channels of the double-channel residual neural network each include a plurality of denoising blocks that are continuously cascaded and have the same structure, and each of the denoising blocks includes a plurality of residual subnetworks, wherein, except for the last residual subnetwork, the other residual subnetworks are configured to extract spatial features of the real part or the imaginary part of the received signal based on convolution, normalization and correction linear operations, and the last residual subnetwork is configured to perform denoising on the spatial features of the real part or the imaginary part of the received signal based on element-wise subtraction.

[0011] The application also provides an RIS cooperative communication method, which is applicable to an RIS controller and includes: obtaining a first real-time pilot signal transmitted by the base station and a second real-time pilot signal transmitted by the user terminal, and determining a real-time channel vector according to the first real-time pilot signal and the second real-time pilot signal; inputting the real-time channel vector into a rate prediction model, the rate prediction model being configured to predict an optimal denoising information rate and a target phase shift matrix corresponding to the optimal denoising information rate according to the real-time channel vector; adjusting the phase shift of all RIS reflection units according to the target phase shift matrix, so that the RIS reflection units implement signal transmission between the base station and the user terminal as a relay; wherein the rate prediction model is trained by denoising information rates corresponding to all phase shift vectors in the RIS reflection codebook.

[0012] The application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the RIS cooperative communication method as described above when executing the computer program.

[0013] The application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and the computer program is executable by a processor to implement the RIS cooperative communication method as described above.

[0014] The application also provides a computer program product including a computer program, and the computer program is executable by a processor to implement the RIS cooperative communication method as described above.

[0015] The RIS cooperative communication system, method, device, medium and product provided by the application train the rate prediction model by the denoising information rates corresponding to all phase shift vectors in the RIS reflection codebook, so as to predict the optimal denoising information rate and the target phase shift matrix corresponding to the optimal denoising information rate, and improve the data rate of the user. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0017] Figure 1 is a structural schematic diagram of the RIS cooperative communication system provided by the application.

[0018] Figure 2 is a functional module block diagram of the RIS cooperative communication system provided by the application.

[0019] Figure 3 is a structural schematic diagram of the neural network channel provided by the application.

[0020] Figure 4 is a flowchart of the RIS cooperative communication method provided by the present application Figure 5 is a structural schematic diagram of an electronic device provided by the present application.

[0021] In the figure: 100-RIS device, 110-RIS controller, 120-RIS reflecting unit, 200-user terminal, 300-base station. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0023] It should be noted that all actions of obtaining signals, information or data in the present application are carried out in compliance with the corresponding data protection regulations and policies of the place and with the authorization given by the corresponding system owner.

[0024] With the increasing requirement of communication speed, high-density cellular structure becomes a necessary condition. However, with the construction of more and more base stations, higher hardware cost consumption and more energy consumption will be brought, which is a big pain point for large-scale use of new technology. How to ensure the quality of communication while relatively less hardware cost consumption and energy consumption is a problem to be considered. Therefore, the present application proposes to introduce a passive reconfigurable intelligent surface into the communication framework to assist communication. By using this reconfigurable intelligent surface to replace a part of high-energy-consumption, high-hardware-cost base stations, while ensuring the quality of communication signals.

[0025] Reconfigurable intelligent surface (RIS) technology has been rapidly developed in recent years. RIS is an artificial electromagnetic surface structure with programmable electromagnetic properties, which is developed from metamaterial technology. Traditional metamaterials can achieve unique physical phenomena such as electromagnetic black holes and electromagnetic invisibility cloaks, but they are described by equivalent medium parameters, which are single-function and fixed analog metamaterials. In recent years, the rapidly developing RIS technology has the characteristics of real-time programmable electromagnetic properties. Real-time programmable is a revolutionary technological leap, which allows the metasurface to change its electromagnetic properties, thereby realizing various functions that traditional metamaterials cannot achieve. RIS is usually composed of a large number of electromagnetic units arranged in an array, and by applying control signals to the adjustable elements on the electromagnetic units, the electromagnetic properties of these electromagnetic units can be dynamically controlled, thereby actively and intelligently controlling the spatial electromagnetic wave in a programmable manner to form an electromagnetic field with controllable amplitude, phase, polarization and frequency. RIS is a cheap passive artificial structure that can digitally manipulate electromagnetic waves and achieve a good electromagnetic propagation environment with limited power consumption. RIS is a planar array composed of a large number of reconfigurable passive elements, each of which can independently induce a certain phase shift signal (through a connected controller) to collectively change the propagation of the reflected signal. Compared with remote radio frequency relays, RIS does not use a transmitter module and only reflects the received signal through a passive array, so it does not produce transmission power consumption. Although the incident signal on the reflector array is not scaled and reflected, it can be used to improve the propagation environment with very low power consumption. Therefore, compared with remote relays, RIS reflection arrays are a more economical choice.

[0026] The existing scheme adopts an unsupervised learning method to design the base station hybrid precoding matrix and the RIS phase shift matrix, adopts a two-stage convolutional neural network model to segment the two-stage network for segmented training, there are many constraint conditions in the segmented training of the two-stage neural network, and the finiteness of the RIS achievable phase shift is not considered.

[0027] The present application provides a RIS cooperative communication scheme based on multiple neural networks to solve the problem of uncontrollable wireless environment in traditional communication, and faces a new type of communication scenario. A single or multiple RIS relays are deployed between the source node (base station) and the destination node (user terminal) of the communication network to create a signal transmission route in cooperation with the network. RIS realizes the reflection of the signal in the propagation process of the signal, and the reflection phase shift of RIS is controlled by the controller.

[0028] Figure 1 is a structural schematic diagram of the RIS cooperative communication system provided by the present application, as Figure 1As shown, due to the susceptibility of high-frequency beams to obstruction, in outdoor scenarios, it is assumed that there is no direct line-of-sight (LOS) link between the base station (BS) signal source and the user terminal (UE). The signal is transmitted via cooperative reflection using RIS devices deployed in the environment. The RIS devices are deployed with... A reconfigurable RIS reflection unit is used in the reflection panel. The RIS controller configures the phase shift matrix of the RIS reflection unit to generate different reflection modes, acting as a relay element in the network to transmit signals from the BS to the UE. It is assumed that all RIS reflection units are controlled and configured by the controller. In operation, the controller operates based on the RIS reflection codebook. The phase shift of each RIS reflection unit is continuously calculated and configured accordingly.

[0029] definition Let be the channels from BS to RIS and from RIS to the k-th UE, respectively. Based on this, the signal received by the user terminal can be represented as: (1) in, Let represent the phase shift matrix selected from the RIS reflection codebook. The phase shift vector of the nth RIS reflection unit in the phase shift matrix can be expressed as . , and They are the first The amplitude and phase shift of each RIS reflector unit. It is the signal sent by the base station to the k-th UE. It is received noise.

[0030] From equation (1), the achievable information rate for the signal to be reflected and propagated to the UE via the RIS device can be obtained as follows: (2) in, Indicates information rate, Signal power With noise power The ratio.

[0031] The expressions for the millimeter-wave channels from BS to RIS and from RIS to the k-th UE are as follows: (3) (4) in, It is the number of paths the signal travels through in a multi-path transmission. Indicates the first The path loss corresponding to each path. They represent the first The pitch angle and the azimuth angle of the path angle of arrival are respectively the pitch angle and the azimuth angle of the path angle of departure. are respectively the pitch angle and the azimuth angle of the path angle of departure.

[0032] and are respectively the steering vectors of the signal sender base station and the signal receiver user terminal, and the expressions are as follows: (5) (6) wherein, represents the wavelength, in the formula 5, the value is the signal wavelength of the base station, in the formula 6, the value is the signal wavelength of the user terminal, represents the antenna spacing, in the formula 5, the value is the antenna spacing of the base station, in the formula 6, the value is the antenna spacing of the user terminal, represents the direct product of the matrix.

[0033] Figure 2 is the functional module block diagram of the RIS cooperative communication system provided by the present application, as shown in Figure 2 The present application provides a RIS cooperative communication system, which comprises a RIS device 100 and a user terminal 200, and the RIS device 100 comprises a RIS controller 110 and a RIS reflecting unit 120.

[0034] The RIS controller 110 is configured to acquire a first real-time pilot signal sent by a base station 300 and a second real-time pilot signal sent by a user terminal 200, determine a real-time channel vector according to the first real-time pilot signal and the second real-time pilot information, input the real-time channel vector into a rate prediction model, the rate prediction model is configured to predict an optimal denoising information rate and a target phase shift matrix corresponding to the optimal denoising information rate according to the real-time channel vector, and adjust the phase shift of all RIS reflecting units 120 according to the target phase shift matrix, so that the RIS reflecting units 120 realize signal transmission between the base station 300 and the user terminal 200 as a relay; wherein the rate prediction model is obtained by training a preset model with denoising information rates corresponding to all phase shift vectors in a RIS reflecting codebook.

[0035] The pilot signal is mainly used for signal synchronization, frequency estimation and channel state estimation, which is a known signal sent at a specific time or frequency, and the receiving end can obtain the characteristics of the channel by analyzing these known signals. After acquiring the first real-time pilot signal and the second real-time pilot signal, the real-time channel vector for real-time signal transmission is estimated by applying the channel reciprocity principle.

[0036] Before practical application, the rate prediction model is pre-trained in the present application, and the rate prediction model is obtained based on the preset model training of the channel vector, all phase shift vectors in the RIS reflection codebook, and the denoising information rate that can be achieved by the user terminal 200 corresponding to all phase shift vectors in the RIS reflection codebook. The rate prediction model contains the correlation between the above data.

[0037] It can be understood that the rate prediction model is trained by the denoising information rate corresponding to all phase shift vectors in the RIS reflection codebook in the present application, so as to predict the optimal denoising information rate and the target phase shift matrix corresponding to the optimal denoising information rate, thereby improving the data rate of the user.

[0038] On the basis of the above-mentioned embodiments, as an optional embodiment, the RIS controller 110 is further configured to train the rate prediction model, and the training learning process of the rate prediction model includes the following steps.

[0039] The first pilot signal transmitted by the base station 300 and the second pilot signal transmitted by the user terminal 200 are acquired, and the channel vector is determined according to the first pilot signal and the second pilot information.

[0040] The denoising information rate corresponding to any phase shift vector in the RIS reflection codebook is acquired from the user terminal 200.

[0041] According to the channel vector and the denoising information rate, a data set is constructed.

[0042] Based on the regression loss function and the data set, the neural network model is trained to obtain the rate prediction model.

[0043] The first pilot signal transmitted by the base station 300 and the second pilot signal transmitted by the user terminal 200 are acquired, and the channel vector is determined according to the first pilot signal and the second pilot information. Specifically, the first pilot signal transmitted by the base station 300 and the second pilot signal transmitted by the user terminal 200 are acquired based on the RIS reflection unit 120, and the channel vector is estimated based on the channel reciprocity principle. The channel vector can be represented as .

[0044] In the step of acquiring the denoising information rate corresponding to any phase shift vector in the RIS reflection codebook from the user terminal 200, the RIS controller 110 performs exhaustive beam training using the RIS reflection codebook. Any phase shift vector corresponds to a reflection beamforming vector. The signal is transmitted to the user terminal 200. The user terminal 200 receives the signal corresponding to the phase shift vector, calculates the achievable denoising information rate based on formulas 1 and 2 after denoising the received signal, and feeds back the denoising information rate to the RIS controller 110.m The denoising information rate corresponding to the phase shift vector can be expressed as .

[0045] The channel vector and the denoising information rate are added to the data set D. After obtaining all the denoising information rates, the deep learning model is trained using the entire data set D, and in the training process, the RIS phase shift vector corresponding to the best denoising information rate that can be achieved by the user terminal 200 is taken as the target, and the deep learning model is trained using a regression loss function, and in each training, the loss function is minimized.

[0046] As an optional embodiment, the regression loss function is used to represent the mean square error between the normalized expected denoising information rate and the predicted denoising information rate of the neural network model.

[0047] The expression of the regression loss function is as follows: wherein, represents the set of all neural network parameters, denotes the normalized expected denoising information rate and the predicted denoising information rate .

[0048] It can be understood that by training the rate prediction model, the best beam is predicted, the requirement for RIS reflection unit 120 array geometry knowledge and the beam training overhead are reduced, the best RIS phase shift design that maximizes the user terminal 200 receiving rate is obtained, and the best information rate of the user terminal 200 is achieved.

[0049] On the basis of the above-mentioned embodiments, as an optional embodiment, the user terminal 200 is configured to perform denoising processing on the received signal, and the denoising processing steps include the following steps.

[0050] The RIS device 100 transmits a signal based on any phase shift vector in the RIS reflection codebook, obtains a received signal corresponding to the phase shift vector, and constructs a super matrix according to the received signal.

[0051] The super matrix is input into a double-channel residual neural network to obtain a denoised received signal output by the double-channel residual neural network.

[0052] According to the denoised received signal, the denoising information rate corresponding to the phase shift vector is determined.

[0053] Optionally, the user terminal 200 is further configured to construct the supermatrix based on the real and imaginary parts of the received signal; correspondingly, the two neural network channels of the dual-channel residual neural network are respectively used to perform noise reduction processing on the real and imaginary parts of the received signal.

[0054] The signal received by user terminal 200 is a noisy matrix, which can be represented as follows: The received signal is a complex-valued matrix with independent real and imaginary parts. Combining the real and imaginary parts of the received signal yields a supermatrix, which can be represented as... .

[0055] Optionally, each of the two neural network channels of the dual-channel residual neural network includes multiple cascaded denoising blocks with the same structure. Each denoising block includes multiple layers of residual subnetworks. Except for the last layer of residual subnetworks, the other residual subnetworks are used to extract the spatial features of the real or imaginary part of the received signal based on convolution, normalization and correction linear operations. The last layer of residual subnetworks is used to denoise the spatial features of the real or imaginary part of the received signal based on element-wise subtraction.

[0056] This application employs two neural network channels to process the real and imaginary parts of the received signal respectively. The expression for the input of the dual-channel residual neural network is as follows: (7) in, This represents a mapping function.

[0057] like Figure 3 As shown, each channel of the dual-channel residual neural network includes multiple cascaded denoising blocks with identical structures, used to progressively improve denoising performance. Each denoising block consists of... It consists of layered residual subnetworks. The front of the residual subnetwork... The layer is used to perform the "Conv+BN+ReLU" operation, which refers to a continuous cascaded combination of convolution (Conv), normalization (BN), and correction linear unit (ReLU). Conv and ReLU are jointly used to explore the spatial features of the channel matrix, and a normalization function is added between Conv and ReLU to improve network stability and training speed. The last layer of the residual subnetwork uses convolution to combine spatial features and element-wise subtraction to utilize the additive nature of noise to denoise the noisy channel matrix.

[0058] set up For the first The functional expressions of each residual subnetwork, where These are the parameters of the residual subnetwork, corresponding to the first... The expression of the de-noising block is as follows: (8) wherein, , respectively represent the input and output of the Bth de-noising block.

[0059] The output result of the Bth de-noising block is the output of the entire de-noising network, as shown in the following formula: (9) The output signal of the dual-channel residual neural network can be obtained from formula (8) and formula (9) as follows: (10) wherein, represents the expression of the dual-channel residual neural network with parameters , represents the residual noise part.

[0060] According to the output signal of the dual-channel residual neural network, the user's de-noised received signal and the corresponding de-noising information rate can be obtained.

[0061] It can be understood that the dual-channel residual neural network is used to de-noise the received signal of the user terminal 200, which can improve the accuracy of the RIS controller 110 estimating the channel based on the UE uplink pilot signal.

[0062] The RIS cooperative communication method provided by the present application will be described below. The RIS cooperative communication method described below can be mutually corresponding to the RIS cooperative communication system described above.

[0063] Figure 4 is a flowchart of the RIS cooperative communication method provided by the present application, as shown in Figure 4 The present application also provides an RIS cooperative communication method suitable for an RIS controller, which includes steps S100-S300.

[0064] Step S100, acquiring a first real-time pilot signal sent by a base station and a second real-time pilot signal sent by a user terminal, and determining a real-time channel vector according to the first real-time pilot signal and the second real-time pilot information; Step S200, inputting the real-time channel vector into a rate prediction model, the rate prediction model being used to predict a best de-noising information rate and a target phase shift matrix corresponding to the best de-noising information rate according to the real-time channel vector; ​Step S300, adjusting the phase shift of all RIS reflection units according to the target phase shift matrix, so that the RIS reflection units realize signal transmission between the base station and the user terminal as a relay. The rate prediction model is trained by the denoised information rate corresponding to all phase shift vectors in the RIS reflection codebook and a preset model.

[0065] Figure 5 An example of an electronic device is shown in the physical structure diagram, as Figure 5 The electronic device can include a processor 510, a communications interface 520, a memory 530, and a communications bus 540, wherein the processor 510, the communications interface 520, and the memory 530 communicate with each other through the communications bus 540. The processor 510 can invoke the logic instructions in the memory 530 to execute the RIS cooperative communication method, which includes: Obtaining a first real-time pilot signal sent by a base station and a second real-time pilot signal sent by a user terminal, and determining a real-time channel vector according to the first real-time pilot signal and the second real-time pilot information; Inputting the real-time channel vector into a rate prediction model, the rate prediction model being used to predict an optimal denoised information rate and a target phase shift matrix corresponding to the optimal denoised information rate according to the real-time channel vector; Adjusting the phase shift of all RIS reflection units according to the target phase shift matrix, so that the RIS reflection units realize signal transmission between the base station and the user terminal as a relay; The rate prediction model is trained by the denoised information rate corresponding to all phase shift vectors in the RIS reflection codebook and a preset model.

[0066] In addition, the logic instructions in the memory 530 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium, including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0067] In another aspect, the present application also provides a computer program product, which comprises a computer program, the computer program being stored in a non-transitory computer-readable storage medium, and the computer program being executable by a processor to enable the computer to perform the RIS cooperative communication method provided by the above method, the method comprising: obtaining a first real-time pilot signal transmitted by a base station and a second real-time pilot signal transmitted by a user terminal, and determining a real-time channel vector according to the first real-time pilot signal and the second real-time pilot signal; inputting the real-time channel vector into a rate prediction model, the rate prediction model being configured to predict an optimal denoising information rate and a target phase shift matrix corresponding to the optimal denoising information rate according to the real-time channel vector; adjusting the phase shift of all RIS reflection units according to the target phase shift matrix, so that the RIS reflection units realize signal transmission between the base station and the user terminal as a relay; wherein the rate prediction model is trained by denoising information rates corresponding to all phase shift vectors in an RIS reflection codebook.

[0068] In yet another aspect, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, the computer program being executable by a processor to implement the RIS cooperative communication method provided by the above method, the method comprising: obtaining a first real-time pilot signal transmitted by a base station and a second real-time pilot signal transmitted by a user terminal, and determining a real-time channel vector according to the first real-time pilot signal and the second real-time pilot signal; inputting the real-time channel vector into a rate prediction model, the rate prediction model being configured to predict an optimal denoising information rate and a target phase shift matrix corresponding to the optimal denoising information rate according to the real-time channel vector; adjusting the phase shift of all RIS reflection units according to the target phase shift matrix, so that the RIS reflection units realize signal transmission between the base station and the user terminal as a relay; wherein the rate prediction model is trained by denoising information rates corresponding to all phase shift vectors in an RIS reflection codebook.

[0069] The system embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0070] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary universal hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, and the computer software products can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and include a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0071] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A RIS cooperative communication system, characterized in that, The system includes a RIS device and a user terminal, wherein the RIS device includes a RIS controller and a RIS reflection unit; The RIS controller is configured to acquire a first real-time pilot signal sent by the base station and a second real-time pilot signal sent by the user terminal, determine a real-time channel vector based on the first real-time pilot signal and the second real-time pilot signal, input the real-time channel vector into a rate prediction model, the rate prediction model being configured to predict an optimal denoising information rate and a target phase shift matrix corresponding to the optimal denoising information rate based on the real-time channel vector, and adjust the phase shift of all RIS reflection units according to the target phase shift matrix so that the RIS reflection units act as relays to realize signal transmission between the base station and the user terminal; wherein, the rate prediction model is obtained by training a preset model with the denoising information rates corresponding to all phase shift vectors in the RIS reflection codebook.

2. The RIS cooperative communication system according to claim 1, characterized in that, The RIS controller is also used for: The first pilot signal transmitted by the base station and the second pilot signal transmitted by the user terminal are acquired, and the channel vector is determined based on the first pilot signal and the second pilot information; The rate of denoising information corresponding to any phase shift vector in the RIS reflection codebook is obtained from the user terminal; A dataset is constructed based on the channel vector and the denoising information rate; The neural network model is trained based on the regression loss function and the dataset to obtain the rate prediction model.

3. The RIS cooperative communication system according to claim 2, characterized in that, The regression loss function is used to characterize the mean square error between the normalized expected denoising rate and the predicted denoising rate of the neural network model.

4. The RIS cooperative communication system according to claim 2, characterized in that, The user terminal is used for: The system receives signals transmitted by the RIS device based on any phase shift vector in the RIS reflection codebook, obtains the received signal corresponding to the phase shift vector, and constructs a supermatrix based on the received signal. The supermatrix is ​​input into a dual-channel residual neural network to obtain the denoised received signal output by the dual-channel residual neural network. Based on the denoised received signal, the denoised information rate corresponding to the phase shift vector is determined.

5. The RIS cooperative communication system according to claim 4, characterized in that, The user terminal is further configured to construct the supermatrix based on the real and imaginary parts of the received signal; correspondingly, the two neural network channels of the dual-channel residual neural network are respectively used to perform noise reduction processing on the real and imaginary parts of the received signal.

6. The RIS cooperative communication system according to claim 4 or 5, characterized in that, The two neural network channels of the dual-channel residual neural network each include multiple cascaded denoising blocks with the same structure. Each denoising block includes multiple layers of residual subnetworks. Except for the last layer of residual subnetworks, the other residual subnetworks are used to extract the spatial features of the real or imaginary part of the received signal based on convolution, normalization and correction linear operations. The last layer of residual subnetworks is used to denoise the spatial features of the real or imaginary part of the received signal based on element-wise subtraction.

7. A RIS cooperative communication method, characterized in that, Applicable to RIS controllers, the method includes: The system acquires a first real-time pilot signal sent by the base station and a second real-time pilot signal sent by the user terminal, and determines a real-time channel vector based on the first real-time pilot signal and the second real-time pilot information. The real-time channel vector is input into the rate prediction model, which is used to predict the optimal denoising information rate and the target phase shift matrix corresponding to the optimal denoising information rate based on the real-time channel vector. The phase shift of all RIS reflection units is adjusted according to the target phase shift matrix so that the RIS reflection units can act as relays to realize signal transmission between the base station and the user terminal. The rate prediction model is obtained by training a preset model with the denoised information rates corresponding to all phase shift vectors in the RIS reflection codebook.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the RIS cooperative communication method as described in claim 7.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the RIS cooperative communication method as described in claim 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the RIS cooperative communication method as described in claim 7.

Images