Method and apparatus for wireless communication

Abstract

Provided are a method and apparatus for wireless communication. The method comprises: acquiring a first measurement matrix and an observation vector of a multi-path communication system, wherein the first measurement matrix is used for representing measurement amounts corresponding to different time delays in a time delay spectrum of the multi-path communication system, and the observation vector is used for indicating an observation value of a channel frequency response (CFR) or a channel impulse response (CIR) corresponding to multiple paths in the multi-path communication system; and under the constraint of a first constraint condition, on the basis of the first measurement matrix and the observation vector, performing sparse recovery on the time delay spectrum of the multi-path communication system, wherein the first constraint condition is used for constraining the relationship between the number of non-zero elements in the time delay spectrum and the number of paths in the multi-path communication system. In the present application, a sparse recovery problem is used to solve a time delay spectrum of a multi-path communication system, thereby helping improve the resolution of a time delay estimation value of the multi-path communication system.

Description

Method and apparatus for wireless communication TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, and more particularly, to a method and apparatus for wireless communication. BACKGROUND

[0002] With the development of communication systems, a large number of high-mobility and super-density connection scenarios appear, and the influence of multipath effect on the channel is also increasingly serious. How to obtain high-resolution channel estimation results of a multipath communication system is one of the challenges currently faced.

[0003] SUMMARY

[0004] The present application provides a method and apparatus for wireless communication. The various aspects involved in the present application are introduced below.

[0005] In a first aspect, a method for wireless communication is provided, comprising: obtaining a first measurement matrix of a multipath communication system and an observation vector, wherein the first measurement matrix is used to represent measurement quantities corresponding to different time delays in a delay spectrum of the multipath communication system, and the observation vector is used to indicate observation values of channel frequency responses (CFRs) or channel impulse responses (CIRs) corresponding to multiple paths in the multipath communication system; under the constraint of a first constraint condition, performing sparse recovery on the delay spectrum of the multipath communication system according to the first measurement matrix and the observation vector, and the first constraint condition is used to constrain the relationship between the number of non-zero elements in the delay spectrum and the number of paths in the multipath communication system.

[0006] In a second aspect, an apparatus for wireless communication is provided, comprising: an obtaining unit configured to obtain a first measurement matrix of a multipath communication system and an observation vector, wherein the first measurement matrix is used to represent measurement quantities corresponding to different time delays in a delay spectrum of the multipath communication system, and the observation vector is used to indicate observation values of channel frequency responses (CFRs) or channel impulse responses (CIRs) corresponding to multiple paths in the multipath communication system; and a processing unit configured to, under the constraint of a first constraint condition, perform sparse recovery on the delay spectrum of the multipath communication system according to the first measurement matrix and the observation vector, and the first constraint condition is used to constrain the relationship between the number of non-zero elements in the delay spectrum and the number of paths in the multipath communication system.

[0007] In a third aspect, a device for wireless communication is provided, which includes a memory configured to store instructions, and a processor configured to execute the instructions stored in the memory to perform the method in any of the first aspect.

[0008] In a fourth aspect, a communication system is provided, which includes the device for wireless communication described above. In another possible design, the system can further include other devices interacting with the device for wireless communication in the solutions provided by the embodiments.

[0009] In a fifth aspect, a computer-readable storage medium is provided, which stores a computer program. The computer program causes a communication device (e.g., a terminal device or a network device) to perform some or all of the steps of the methods in the various aspects described above.

[0010] In a sixth aspect, a computer program product is provided, which includes a non-transitory computer-readable storage medium storing a computer program. The computer program is operable to cause a communication device (e.g., a terminal device or a network device) to perform some or all of the steps of the methods in the various aspects described above. In some implementations, the computer program product can be a software installation package.

[0011] In a seventh aspect, a chip is provided, which includes a memory and a processor. The processor can invoke and run a computer program from the memory to implement some or all of the steps described in the methods in the various aspects described above.

[0012] In the embodiments of the present application, the first measurement matrix and the observation vector of the multipath communication system are used to sparsely recover the delay spectrum of the multipath communication system under the constraint of the first constraint condition. Solving the delay spectrum of the multipath communication system using the sparse recovery problem helps to improve the resolution of the delay estimation value of the multipath communication system. BRIEF DESCRIPTION OF DRAWINGS

[0013] FIG. 1 is a wireless communication system 100 to which the embodiments of the present application are applied.

[0014] FIG. 2 is a schematic diagram of a signal transmission process in a wireless communication system to which the embodiments of the present application are applied.

[0015] FIG. 3 is a schematic diagram of channel estimation and signal recovery to which the embodiments of the present application are applied.

[0016] FIG. 4 is a schematic flowchart of a method for wireless communication provided by the embodiments of the present application.

[0017] FIG. 5 is a schematic diagram of a prediction method of a first neural network model in the embodiments of the present application.

[0018] FIG. 6 is a schematic flowchart of acquiring a second prediction set in an embodiment of the present application.

[0019] FIG. 7 is an example diagram of acquiring a second prediction set in an embodiment of the present application.

[0020] FIG. 8 is a schematic diagram of an apparatus for wireless communication in an embodiment of the present application.

[0021] FIG. 9 is a schematic diagram of an apparatus in an embodiment of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the present application will be described below with reference to the accompanying drawings.

[0023] Communication system architecture

[0024] FIG. 1 is a wireless communication system 100 to which embodiments of the present application are applied. The wireless communication system 100 can include a network device 110 and a terminal device 120. The network device 110 can be a device that communicates with the terminal device 120. The network device 110 can provide communication coverage for a specific geographic area and can communicate with the terminal device 120 located in the coverage area.

[0025] FIG. 1 exemplarily shows one network device and two terminals. Optionally, the wireless communication system 100 can include multiple network devices and each network device can include other numbers of terminal devices within its coverage, which is not limited in the embodiments of the present application.

[0026] Optionally, the wireless communication system 100 can further include a network controller, a mobile management entity, and other network entities, which are not limited in the embodiments of the present application.

[0027] It should be understood that the technical solutions of the embodiments of the present application can be applied to various communication systems, for example, a 5th generation (5G) system or new radio (NR), a long term evolution (LTE) system, an LTE frequency division duplex (FDD) system, an LTE time division duplex (TDD), and the like. The technical solutions provided by the present application can also be applied to future communication systems, such as a 6th generation mobile communication system, a satellite communication system, and the like.

[0028] The terminal device in the embodiments of the present application can also be referred to as a user equipment (UE), an access terminal, a user unit, a user station, a mobile station, a mobile station (MS), a mobile terminal (MT), a remote station, a remote terminal, a mobile device, a user terminal, a terminal, a wireless communication device, a user agent or a user apparatus. The terminal device in the embodiments of the present application can refer to a device providing voice and / or data connectivity for a user, and can be used to connect people, things and machines, for example, handheld devices with wireless connection function, vehicle-mounted devices, etc. The terminal device in the embodiments of the present application can be a mobile phone, a tablet computer (Pad), a notebook computer, a palm computer, a mobile internet device (MID), a wearable device, a virtual reality (VR) device, an augmented reality (AR) device, a wireless terminal in industrial control, a wireless terminal in self driving, a wireless terminal in remote medical surgery, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, etc. Optionally, the UE can be used to act as a base station. For example, the UE can act as a scheduling entity, which provides sidelink signals between UEs in V2X or D2D, etc. For example, a cellular phone and a car communicate with each other using sidelink signals. The cellular phone and the smart home device communicate with each other without relaying the communication signals through the base station.

[0029] The network device in the embodiments of the present application can be a device for communicating with a terminal device, which can also be referred to as an access network device or a radio access network device, such as a network device, which can be a base station. The network device in the embodiments of the present application can refer to a radio access network (RAN) node (or device) that accesses a terminal device to a wireless network. The base station can broadly cover various names in the following or be replaced by the following names, such as: Node B (NodeB), evolved Node B (eNB), next generation Node B (gNB), relay station, access point, transmitting and receiving point (TRP), transmitting point (TP), master station MeNB, auxiliary station SeNB, multi-standard radio (MSR) node, home base station, network controller, access node, wireless node, access point (AP), transmission node, transceiver node, baseband unit (BBU), remote radio unit (RRU), active antenna unit (AAU), remote radio head (RRH), central unit (CU), distributed unit (DU), positioning node, etc. The base station can be a macro base station, a micro base station, a relay node, a donor node or the like, or a combination thereof. The base station can also refer to a communication module, modem or chip for being arranged in the foregoing device or apparatus. The base station can also be a mobile switching center and a device that undertakes the function of a base station in device-to-device (D2D), vehicle-to-everything (V2X), machine-to-machine (M2M) communication, network side device in 6G network, device that undertakes the function of a base station in future communication system, etc. The base station can support networks of the same or different access technologies. The embodiments of the present application do not limit the specific technology and specific device form adopted by the network device.

[0030] The base station can be fixed or mobile. For example, a helicopter or a drone can be configured to act as a mobile base station, and one or more cells can move according to the location of the mobile base station. In other examples, a helicopter or a drone can be configured to act as a device that communicates with another base station.

[0031] In some deployments, the network device in the embodiments of the present application can refer to a CU or a DU, or the network device includes a CU and a DU. The gNB can also include an AAU.

[0032] The network device and the terminal device can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; can also be deployed on the water surface; and can also be deployed on aircraft, balloons and satellites in the air. The scenarios in which the network device and the terminal device are located are not limited in the embodiments of the present application.

[0033] It should be understood that all or part of the functions of the communication device in the present application can also be implemented by software functions running on hardware, or by virtualized functions instantiated on a platform (such as a cloud platform).

[0034] Signal transmission process in a wireless communication system

[0035] FIG. 2 is a flowchart of signal transmission in a wireless communication system to which the embodiments of the present application are applicable. As shown in FIG. 2, the signal transmission process in the wireless communication system can be roughly divided into various signal processing processes S211 to S218 shown in FIG. 2. Part or all of the signal processing processes shown in FIG. 2 can be implemented by a single artificial intelligence (AI) model.

[0036] The transmitter performs channel coding on the information to be transmitted in the channel coding process S211, to obtain a coded bit stream. The information to be transmitted can be in the form of a bit stream.

[0037] In the modulation process S212, the bit stream is modulated into modulation symbols.

[0038] In the pilot insertion process S213, pilot symbols are inserted into the modulation symbols to form a signal to be transmitted, wherein the pilot symbols can be used for channel estimation and symbol detection by the receiver.

[0039] In the transmission signal S214, the signal is transmitted to the receiver on a channel. During the transmission of the signal on the channel, noise is usually superimposed on the signal.

[0040] In the channel estimation process S215, the receiver can perform channel estimation based on the pilot signal to obtain channel state information (channel state information-reference signal, CSI), and feed back the CSI to the transmitter through a feedback link for the transmitter to adjust the channel coding, modulation, precoding and the like.

[0041] In the symbol detection process S216, the received modulation symbols are subjected to symbol detection to obtain a detection result.

[0042] In the demodulation process S217, the received modulated symbols are demodulated based on the detection result to obtain a code stream.

[0043] In the channel decoding process S218, the code stream is decoded to obtain recovered information, which can be in the form of a bit stream.

[0044] It should be understood that the signal processing processes S211 to S218 shown in FIG. 2 are only exemplary and common signal processing processes in a wireless communication system, and the wireless communication system can also include resource mapping, precoding, interference cancellation, CSI measurement, etc. Signal processing processes can also be implemented by separate AI models. For the sake of brevity, the present application will not be described again.

[0045] Channel estimation

[0046] Due to the complexity and time-varying nature of the wireless channel environment, in a wireless communication system (for example, the wireless communication system introduced above), the receiver needs to recover the received signal based on the estimation result of the channel. FIG. 3 is a schematic diagram of channel estimation and signal recovery applicable to embodiments of the present application.

[0047] As shown in FIG. 3, in step S310, the transmitter transmits a series of pilot signals known to the receiver, such as channel state information-reference signals (CSI-RS), demodulation reference signals (DMRS), etc. in addition to the data signal on the time-frequency resource.

[0048] In step S311, the transmitter transmits the above-mentioned data signal and pilot signal to the receiver through the channel.

[0049] Wherein, the time-frequency resource occupied by the pilot signal is different from the time-frequency resource occupied by the data signal.

[0050] In step S312, the receiver can perform channel estimation after receiving the pilot signal. In one possible implementation, the receiver can estimate the channel information of the channel transmitting the pilot signal based on the pre-stored pilot signal and the received pilot signal through a channel estimation algorithm (for example, least squares method (LS) channel estimation).

[0051] In step S313, the receiver can recover the channel information on the full time-frequency resource according to the channel information of the channel transmitting the pilot sequence using an interpolation algorithm, which is used for subsequent CSI feedback or data recovery, etc.

[0052] The result of channel estimation can not only be applied to signal recovery, but also be applied to indoor positioning, environment perception and other fields, and serve new task targets. For example, by estimating the multipath delay in a multipath communication system, channel equalization can be achieved, the multipath effect of the channel can be compensated to enable energy to be concentrated in the main path, and the robustness of communication can be improved. The relative delay in the multipath delay estimation in the multipath communication system can be regarded as the time difference of arrival (TDOA) between the original node and the virtual node symmetrical along the reflecting surface, the normal position of the reflecting surface can be obtained by geometric calculation for environment perception, and single base station indoor positioning function can also be realized based on multipath assistance.

[0053] Sparse recovery algorithm

[0054] The sparse recovery problem model is By the known measurement matrix A and the observation vector y, the sparse delay vector x is reconstructed. Common sparse recovery algorithms include greedy algorithms, convex optimization algorithms, and algorithms based on the Bayesian framework.

[0055] The greedy algorithm performs inner product of the residual and each column of the measurement matrix each time, selects the basis vector with the highest matching degree with the residual, iteratively finds the support set of the sparse vector, and updates the residual. Finally, the sparse signal is reconstructed by using the least square estimation of the support set. Based on the greedy idea, there are matching pursuit (MP), orthogonal matching pursuit (OMP), sparsity adaptive matching pursuit (SAMP), etc. For example, the MP algorithm performs inner product of the aliasing signal and each column of the measurement matrix each time, selects the position with the maximum inner product in the pseudo spectrum, that is, the basis with the largest correlation degree, and sets the delay at the maximum inner product position as the prediction value of this round. The limitation of this idea is that the algorithm is too greedy, the error accumulation of the sparse recovery problem is particularly serious, and the iterative delay estimation strategy has poor robustness.

[0056] The sparse recovery problem with l0-norm constraint is a non-deterministic polynomial (NP)-Hard problem. A convex optimization algorithm considers relaxing the l0-norm constraint, so that an approximate solution of the non-convex problem can be obtained by an optimization method, and the sparse delay vector can be predicted. The l0-norm is relaxed to the l1-norm, and the sparse recovery problem is converted into a convex problem. The least absolute shrinkage and selection operator (Lasso) regression model is used for solving, and a common algorithm is a basis pursuit (BP) algorithm. Alternatively, the l0-norm can be relaxed to the l p The Focuss algorithm iteratively solves the l p The l1-norm optimization approximation solves the original l0-norm sparse recovery problem.

[0057] An algorithm based on a Bayesian framework transforms the original problem into a parameterized maximum a posteriori (MAP) estimation problem by giving a Bayesian prior assumption to the sparse recovery problem. The sparse Bayesian learning (SBL) assumes that any element of the sparse vector to be recovered satisfies a Gaussian prior distribution, and the parameters of the Gaussian prior distribution are subject to a conjugate prior Gamma distribution. By calculating the marginal distribution of the hyperparameters in the Gamma distribution, the mean and variance of the unknown parameters are obtained by MAP iteration, so that an accurate estimation of the sparse delay vector is obtained.

[0058] As described above, the channel estimation result can be applied to signal recovery, indoor positioning, environment perception, and other scenarios in a wireless communication system. With the development of communication systems, a large number of high-mobility and super-density connection scenarios appear, and the influence of multipath effect on the channel is also increasingly serious. How to obtain a high-resolution channel estimation result of a multipath communication system is one of the challenges.

[0059] To solve the above problems, the present application provides a method for wireless communication. The method uses a first measurement matrix and an observation vector of a multipath communication system to perform sparse recovery on a delay spectrum of the multipath communication system under a first constraint condition. Solving the delay spectrum of the multipath communication system by using the sparse recovery problem helps to improve the resolution of the delay estimation value of the multipath communication system.

[0060] The method for wireless communication of the embodiments of the present application is described below in conjunction with FIG. 4. The method shown in FIG. 4 includes steps S410 to S420.

[0061] In step S410, a first measurement matrix and an observation vector of a multipath communication system are obtained.

[0062] In some implementations, the observation vector is used to indicate the observed value of the CFR or CIR corresponding to multiple paths in the multipath communication system, or in other words, the observation vector is used to indicate the observed value of the CFR or CIR corresponding to the composite channel in the multipath communication system.

[0063] In some implementations, the observation vector is determined based on a pilot signal transmitted by a transmitter in a multipath communication system. For example, the pilot signal and the received signal are known, and the observation signal is obtained using LS.

[0064] The following describes how to obtain the observation vector, taking the observation vector used to indicate the CFR observation value as an example.

[0065] Assume that the multipath communication system is an indoor wireless communication system based on orthogonal frequency division multiplexing (OFDM) architecture. The wireless signal sent by the source node through a single antenna can be received by the user node's single antenna. The frequency domain pilot signal transmitted by the transmitter is x = [x1, x2, ..., x N ] T After removing the cyclic prefix and performing discrete Fourier transform (DFT) at the receiving end, the received signal corresponding to the pilot segment is intercepted as y = [y1, y2, ..., y N ] T , where N is the number of subcarriers. The CFR of the composite channel in a multipath communication system is represented by h, which can be expressed as a matrix y = Xh + z, where X = diag{x} and z is white Gaussian noise (AWGN). The LS algorithm is used to calculate the CFR observation vector

[0066] In some implementations, the first measurement matrix is ​​used to represent measurement quantities corresponding to different delays in a delay profile of a multipath communication system.

[0067] In some implementations, the delay domain is discretized into a delay grid, and the delay corresponding to the multipath is a portion of the delay in the delay grid. The delay spectrum has non-zero elements only at the delay grid positions where complex amplitudes exist. For example, the delay grid is T, T = [T1, T2, ..., T P ], P is the delay grid length, satisfying P is greater than the number of subcarriers N. The delay spectrum is u, u is a vector of length P, satisfying wherein τ l is the delay of the lth path in the multipath.

[0068] In the embodiments of the present application, the measurement quantity corresponding to different time delays in the delay spectrum can be a measurement quantity in the time domain or a measurement quantity in the frequency domain.

[0069] In some implementations, the measurement quantity corresponding to different time delays in the delay spectrum is associated with a filter at the transmitter and / or a filter at the receiver in the multipath communication system.

[0070] In some implementations, the measurement quantity corresponding to different time delays in the delay spectrum is associated with a time-domain convolution of a filter at the transmitter and a filter at the receiver in the multipath communication system. For example, the filter at the transmitter is a pulse shaping filter f t (τ), the filter at the receiver is a matched filter f r (τ), and the time-domain convolution of the transmit-receive filter is The filter waveform can be measured in an open outdoor environment, and the time-domain convolution of the transmit-receive filter (for example, in the form of a raised cosine filter) can be calculated.

[0071] In some implementations, the time-domain convolution of the filter at the transmitter and the filter at the receiver can also be referred to as the time-domain convolution of the transmit-receive filter, or the time-domain waveform of the transmit-receive filter.

[0072] In some implementations, the measurement quantity corresponding to different time delays in the delay spectrum indicates the time-domain convolution of the filter at the transmitter and the filter at the receiver in the multipath communication system, which can be understood as that the first measurement matrix is used to represent the time-domain waveforms of the transmit-receive filter corresponding to different time delays in the delay spectrum of the multipath communication system.

[0073] In some other implementations, the measurement quantity corresponding to different time delays in the delay spectrum indicates the discrete frequency-domain waveform of the transmit-receive filter, which can be understood as that the first measurement matrix is used to represent the discrete frequency-domain waveforms of the transmit-receive filter corresponding to different time delays in the delay spectrum of the multipath communication system.

[0074] Hereinafter, the first measurement matrix is introduced by taking an example of the first measurement matrix being used to represent the discrete frequency-domain waveforms of the transmit-receive filter corresponding to different time delays in the delay spectrum of the multipath communication system.

[0075] Suppose the filter at the transmitter is a pulse shaping filter f t (τ), the filter at the receiver is a matched filter f r (τ), and the time-domain convolution of the transmit-receive filter is The time-domain representation of the composite channel without considering the filter effect is where L is the number of paths in the composite channel, a l , and τ l are the complex amplitude and absolute delay of the lth path, respectively. Thus, the composite channel can be expressed in time domain as By sampling the time τ with a period T s and by DFT processing, the composite channel can be expressed in frequency domain as where n denotes the n-th discrete point in time domain. Assume that the delay spectrum u is expressed as The observation vector h is the observation of the CFR, which can be written as a linear relationship between the delay spectrum u and the frequency domain waveform set A corresponding to the delay of the transmit-receive filter through the delay grid T, i.e., h = Au + n, where n is the noise introduced in the observation process, and a first measurement matrix A = [DFT{f(nT s -Τ1), DFT{f(nT s -Τ2), …, DFT{f(nT s -Τ P )] e C N×P .

[0076] In some implementations, the first measurement matrix is used to represent the measurement quantity corresponding to different power in the energy spectrum of the multi-path communication system. For example, the energy spectrum can be a power spectrum.

[0077] In some implementations, the energy domain is discretized into an energy grid, and the spectrum intensity corresponding to the multi-path is part of the energy in the energy grid, and the energy spectrum is only a non-zero element at the energy grid position where the spectrum intensity exists.

[0078] In some implementations, the energy spectrum is associated with the delay corresponding to the multi-path in the multi-path communication system.

[0079] In some implementations, based on the first parameter and / or the second parameter, the first measurement matrix can be adjusted to obtain an adjusted second measurement matrix.

[0080] In some implementations, the first parameter and / or the second parameter represent the error introduced by the hardware imperfection in the multi-path communication system, or in other words, the first parameter and / or the second parameter represent the error caused by the hardware defects in the multi-path communication system. Adjusting the first measurement matrix using the first parameter and / or the second parameter helps to improve the accuracy of the first measurement matrix.

[0081] In some implementations, the first parameter represents the antenna gain corresponding to the subcarrier in the multi-path communication system.

[0082] Since the curve of the antenna gain variation is not absolutely flat in the frequency range, there can be differences in the antenna gain at the subcarrier frequencies of the OFDM signal. In some implementations, the first parameter is represented using a vector. For example, the first parameter can be represented using a vector g, where g i is the actual gain for the i-th subcarrier.

[0083] In some implementations, the second parameter represents a sampling time offset (STO) between the receiving end and the transmitting end in the multipath communication system.

[0084] In some implementations, the STO is caused by the sampling clock offset between the digital-to-analog converter in the source node and the analog-to-digital converter in the user node due to the clock desynchronization between the transmitting end and the receiving end.

[0085] In some implementations, the second parameter is used to indicate that a first time offset is added to the CIR of the channel in the multipath communication system in the time domain. For example, the second parameter represents the STO, and the time domain response of the original composite channel in the second row and the second column in Table 1, i.e., the CIR h[n], is added with a first time offset δ to obtain the response h[n+δ] in the second row and the third column considering the STO = δ.

[0086] In some other implementations, the second parameter is used to indicate that a first rotation phase is added to the phase of the CFR of the channel in the multipath communication system in the frequency domain. For example, the second parameter represents the STO, and the frequency domain response of the original composite channel in the third row and the second column in Table 1, i.e., the CFR H[k], is added with a first rotation phase e j2πkδ / N to obtain the response H[k]e j2πkδ / N in the third row and the third column considering the STO = δ.

[0087] Table 1

[0088] The following describes a method for adjusting the first measurement matrix to obtain the second measurement matrix based on the first parameter and the second parameter.

[0089] Suppose the first parameter is the antenna gain g corresponding to the subcarrier in the multipath communication system, where g i is the actual gain for the i-th subcarrier. The second parameter is a sampling time offset δ f between the receiving end and the transmitting end in the multipath communication system, where δ f indicates that a first rotation phase is added to the phase of the CFR of the channel in the multipath communication system in the frequency domain. Suppose the first measurement matrix is A = [DFT{f(nT s - T1), DFT{f(nT sDFT {f(nT s -Τ p )]∈C N×P . Adjust A based on g, which is to add unknown weight g to each row of A i Adjust A based on δ f , which is to add first rotation phase to the phase of A The adjusted second measurement matrix is

[0090] In some implementations, the second measurement matrix is represented as where g i represents the antenna gain corresponding to the i-th subcarrier, δ f represents the sampling clock deviation, f i represents the i-th subcarrier, T j represents the j-th value of the time delay spectrum, F i represents the i-th value of F, f(nT s ) represents the time-domain convolution of the transceiver filter at time nT s , T S is the sampling period, and n represents the n-th sampling point in the time domain. Since f i and f can be obtained by measurement, T j is a known value, and the calculation of only needs to solve g and δ f , a total of N+1 unknown parameters, which helps to reduce the difficulty of obtaining the measurement matrix compared to solving N+P values of the first measurement matrix.

[0091] In the embodiments of the present application, the acquisition method of the above-mentioned first measurement matrix or second measurement matrix is not limited. In some implementations, the method of obtaining the first measurement matrix or the second measurement matrix can be the gradient descent method. In some implementations, the method of obtaining the first measurement matrix or the second measurement matrix can be the Newton method. In some implementations, the method of obtaining the first measurement matrix or the second measurement matrix can be the Adam optimization algorithm. In some implementations, the method of obtaining the first measurement matrix or the second measurement matrix can be the Adagrad optimization algorithm. In some implementations, the method of obtaining the first measurement matrix or the second measurement matrix can be the genetic algorithm.

[0092] In some implementations, the gradient descent method used to obtain the first measurement matrix or the second measurement matrix may be a mini-batch gradient descent method, which helps to increase the speed of obtaining the first measurement matrix or the second measurement matrix. In some implementations, the gradient descent method may be a batch gradient descent (BGD) method. In other implementations, the gradient descent method may be a stochastic gradient descent (SGD) method.

[0093] In some implementations, when the second measurement matrix is ​​obtained using a mini-batch gradient descent method, the second measurement matrix is ​​optimized based on a loss function, where the loss function is used to indicate a reconstruction error of the second measurement matrix. For example, the loss function is represented by L, where L is defined as The reconstructed mean-square error (MSE) is Where θ is The unknown parameters in C N×P It is extremely difficult to directly perform gradient optimization on N×P unknown parameters based on the second measurement matrix. However, by utilizing the inherent constraints of the second measurement matrix and reducing the unknown parameters through the idea of ​​local parameterization, it helps to reduce the computational complexity.

[0094] The following describes a method for obtaining the second measurement matrix by taking the mini-batch gradient descent method as an example.

[0095] Assume that the second measurement matrix is ​​expressed as where g i represents the antenna gain corresponding to the i-th subcarrier, δ f represents the sampling clock deviation, f i represents the i-th subcarrier, F i represents the i-th value of F, where F is the discrete frequency domain waveform of the transceiver filter, and T j represents the jth value of the delay spectrum. g is the antenna gain corresponding to the subcarrier in the multipath communication system, where g i is the actual gain of the i-th subcarrier. f is the sampling clock deviation between the receiver and transmitter in the multipath communication system. i and F can be obtained by measuring, T j is a known value, so we can obtain The unknown parameters g and δ in f Sure The parameters θ={g,δ f} represents the parameter to be obtained. Since the hardware parameter is independent of the environment, a large number of known delay spectrum u and observation vector h can be obtained through multiple experiments in different positions and scenes, and a data set is constructed where l is the total number of data sets. When the parameter θ = {g, δ f} is obtained using the mini-batch gradient descent method, the loss function L can be defined as the MSE of the reconstruction, that is, In the process of obtaining by the mini-batch gradient descent method, the update process of the parameter θ is where m is the size of the batch, i is the corresponding index in the data set D, j is the round of parameter iteration, and a is the learning rate. By setting appropriate batch size and learning rate, the loss function gradually converges after multiple iterations, at which time the parameter θ also converges to the true value around, and the optimal estimation value of the parameters g and δ f can be obtained. Based on the estimation value of the parameters g and δ f obtained by the mini-batch gradient descent method, the second measurement matrix is calculated. The above method of obtaining the second measurement matrix based on the mini-batch gradient descent method can also be referred to as a "hardware error correction feedforward network based on gradient optimization".

[0096] Referring back to FIG. 4, in step S420, the delay spectrum of the multi-path communication system is sparsely recovered under the constraint of the first constraint condition according to the first measurement matrix and the observation vector.

[0097] In some implementations, the delay spectrum of the multi-path communication system can be sparsely recovered under the constraint of the first constraint condition according to the second measurement matrix and the observation vector.

[0098] In some implementations, the first constraint condition is used to constrain the relationship between the number of non-zero elements in the delay spectrum and the number of paths in the multi-path communication system.

[0099] In some implementations, the number of paths in the multi-path communication system can be the maximum number of paths allowed to exist. In other implementations, the number of paths in the multi-path communication system can be a maximum value of the total number of paths set.

[0100] In some implementations, the first constraint condition is used to constrain the number of non-zero elements in the delay spectrum to be less than the maximum value of the total number of paths set in the multi-path communication system. In other implementations, the first constraint condition is used to constrain the number of non-zero elements in the delay spectrum to be less than the maximum number of paths allowed to exist in the multi-path communication system. For example, the first constraint condition is represented as ‖u‖0<k, where k represents the maximum number of paths allowed to exist in the multi-path communication system.

[0101] ​In some implementations, the sparse recovery of the delay profile of the multi-path communication system according to the first measurement matrix and the observation vector can be understood as that a sparse recovery model of the delay profile is established based on the first measurement matrix and the observation vector, and a sparse vector, i.e., the delay profile, is solved based on the sparse recovery model.

[0102] In some implementations, if the first constraint condition is represented as ‖u‖0 the delay profile u of the multi-path communication system is sparsely recovered.

[0103] In some implementations, the sparse recovery of the delay profile u can be understood as that the formula the delay profile u is sparsely recovered, which can be understood as that the formula is a sparse recovery model of the delay profile u, the MSE of the delay profile u is calculated, and the delay profile u at which the MSE is the smallest is found.

[0104] The sparse recovery of the delay profile can be understood as solving a sparse recovery problem. However, the influence of the construction mode of the measurement matrix on the recovery accuracy is great in the traditional sparse recovery algorithm. The correlation between atoms in the first measurement matrix is high, and the restricted eigenvalue condition (REC) and the restricted isometry property (RIP) condition can not be met. It is difficult to obtain a delay profile with high accuracy by using the sparse recovery algorithm in the traditional compressed sensing reconstruction, and the delay profile estimation task of a multi-path communication system with a variable number and super-resolution cannot be implemented.

[0105] For example, the sparse recovery method based on the greedy algorithm has a serious performance decline in the case of high coupling of matching bases, tends to identify two close paths as one path, and returns a delay estimation value with large error, which has low prediction accuracy and is greatly affected by noise. The sparse recovery method based on the convex optimization algorithm only proves that the norm relaxation solution is strictly consistent with the original problem solution when the measurement matrix satisfies the RIP condition, and it is difficult to obtain an accurate sparse solution in the multi-path delay estimation scene, and the estimation accuracy and robustness are poor for the variable number of multi-paths and serious aliasing. The sparse recovery method based on the Bayesian framework uses a statistical prior method to establish a more cautious parameter estimation strategy, but also introduces a large calculation complexity. Moreover, this method is still difficult to cope with the severe scene of aliasing multi-paths that breaks through the Nyquist sampling theorem, and cannot implement the super-resolution multi-path identification task.

[0106] To solve the above problems, in some implementations, the observation vector is input into a first neural network model for model prediction to obtain a total number of paths in the multipath communication system and a first time delay estimation value corresponding to each path in the multipath communication system. Compared with a traditional sparse recovery algorithm, since the neural network has strong denoising capability and can perform sparse recovery on the time delay spectrum in the case of high correlation between atoms in the measurement matrix, it helps to improve the resolution of the first time delay estimation value corresponding to each path in the multipath communication system.

[0107] A first network model prediction method of an embodiment of the present application is described below in combination with FIG. 5. In the method of FIG. 5, the original time delay spectrum u, the observation signal h, and the measurement matrix A can be inputs of the method.

[0108] In step S510, a time delay pseudo-spectrum is calculated based on the observation vector and the measurement matrix.

[0109] In some implementations, the time delay pseudo-spectrum can be expressed as A is the first measurement matrix or the second measurement matrix, h is the observation vector, and the element in the time delay pseudo-spectrum is the inner product of the observation vector and the vector of each column of the measurement matrix. Since the columns of A are not orthogonal and the correlation degree monotonically decays as the corresponding time delay difference increases, the time delay pseudo-spectrum is not sparse, but is similar to the diffusion and leakage of the original time delay spectrum u around the grid points corresponding to the time delays of the sparse paths.

[0110] In step S520, the time delay pseudo-spectrum is rearranged.

[0111] In some implementations, the time delay pseudo-spectrum is rearranged into a two-dimensional vector close to a square matrix, so that the time delay pseudo-spectrum is close to the reconstruction problem of the original time delay spectrum u, which is similar in structure to the denoising problem of natural images.

[0112] In step S530, the observation vector is input into a first neural network model for model prediction to obtain a reconstructed time delay spectrum close to the original time delay spectrum.

[0113] In an embodiment of the present application, the training set of the first neural network model includes the original time delay spectrum, the observation signal, and the measurement matrix, which is the first measurement matrix or the second measurement matrix. For example, the training set is expressed as D train , which includes l pairs of the original time delay spectrum u and the observation signal h and the first measurement matrix A.

[0114] In an embodiment of the present application, the test set of the first neural network model includes the original time delay spectrum and the observation signal. For example, the test set which includes s pairs of the original time delay spectrum u and the observation signal h.

[0115] In some implementations, the first neural network model is trained based on the loss function to make the reconstructed delay spectrum closer to the original delay spectrum of the multipath communication system. For example, by setting appropriate learning parameters and optimizers for the first neural network model, multiple model training is performed using the first neural network model, and during the training process, part of the network parameters are automatically adjusted until the loss function converges, and the trained first neural network model can represent a nonlinear mapping relationship from the delay pseudo spectrum to the original delay spectrum, and the reconstructed delay spectrum obtained is closer to the original delay spectrum.

[0116] In some implementations, the loss function is used to calculate the difference between the reconstructed delay spectrum and the original delay spectrum of the multipath communication system. For example, the difference (e.g., MSE) between the original delay spectrum u and the reconstructed delay spectrum is defined as the reconstruction error, denoted as L rec , where P is the total number of atoms in the reconstructed delay spectrum .

[0117] In some implementations, the loss function is used to calculate the asymmetric error between the reconstructed delay spectrum and the original delay spectrum. For example, the asymmetric error between the original delay spectrum u and the reconstructed delay spectrum is defined as L asymm , where sign(·) is a sign function, and a is a set parameter satisfying 0 < a < 1, P is the total number of atoms in the reconstructed delay spectrum , which is used to punish the underestimated part in the delay spectrum recovery process. Since the original delay spectrum u is composed of a large number of zero values and sparse non-zero elements corresponding to multipath, the proportion of the two kinds of samples is very different, and the deep network tends to directly give zero prediction value. The first neural network model is trained based on the asymmetric error, which has certain performance improvement for the prediction scene with too large proportion of zero values, and helps to further improve the accuracy of the reconstructed delay spectrum.

[0118] In some implementations, the loss function is used to calculate the difference between the reconstructed delay spectrum and the original delay spectrum of the multipath communication system, and the asymmetric error between the reconstructed delay spectrum and the original delay spectrum. For example, the loss function is obtained by weighted summation of the reconstruction error and the asymmetric error, using L total to represent, L total = L rec + λ asymm L asymm , L rec identifies the reconstruction error, and L asymm represents the asymmetric error.

[0119] In the embodiments of the present application, the first neural network model is not limited. In some implementations, the first neural network model can be a convolutional neural network (CNN) model. In other implementations, the first neural network model can be a deep network model based on a residual network, UNet, or an autoencoder.

[0120] In some implementations, the first neural network model has 10 layers. For example, referring to FIG. 5, the first neural network model is stacked by 10 layers of CNN.

[0121] In some implementations, the first layer (input layer) of the first neural network model uses 64 7*7 convolution kernels to receive 1-channel input, so as to obtain a larger receptive field, which helps the first neural network model adapt to complex delay pseudo-spectrum structures.

[0122] In some implementations, the middle layer of the first neural network model uses a CNN deep connection form, which helps improve the ability of the model to express complex problems.

[0123] In some implementations, the middle layer of the first neural network model uses 64 3*3 convolution kernels to receive 64-channel input, and sets the edge padding mode to equal padding, so as to maintain the two-dimensional delay pseudo-spectrum size unchanged in the deep CNN, and help reduce the degree of padding operation changing the edge structure of the two-dimensional delay pseudo-spectrum.

[0124] In some implementations, the last layer (output layer) of the first neural network model uses 1 7*7 convolution kernel to receive 64-channel input, so that the output result of the first neural network model is consistent with the shape of the two-dimensional delay pseudo-spectrum input into the first neural network model.

[0125] In some implementations, the first neural network model is a CNN using deep connection, and is therefore also referred to as a “deep CNN framework”.

[0126] In some implementations, the input layer and the middle layer of the first neural network model include an activation layer. For example, referring to FIG. 5, an activation layer is added after the convolution layer, and the activation layer uses a linear rectification function (RELU) to enhance the non-linear explanation ability of the first network model. In some implementations, the output layer of the first neural network model does not add an activation layer, so as to reduce the probability of gradient disappearance in the back propagation process caused by the activation function (for example, ReLU function, Sigmoid function, etc.).

[0127] In some implementations, the intermediate layer of the first neural network model includes a batch normalization (BN) layer for performing batch normalization processing, which helps to improve the convergence speed and learning stability of the first neural network model for model prediction.

[0128] In some implementations, the BN layer of the first neural network model introduces noise, which helps to increase the system robustness of the first network model and prevent overfitting.

[0129] Since the atomic values of the delay spectrum correspond to the complex amplitudes of different paths in the multi-path communication system, the mean and variance of the pseudo delay spectrum and the reconstructed delay spectrum have physical meanings. In some implementations, no regularization part is added to the input layer and the output layer of the first neural network model.

[0130] In step S540, the reconstructed delay spectrum is rearranged into a vector

[0131] In some implementations, based on the rearranged reconstructed delay spectrum, the total number of paths in the multi-path communication system and the first delay estimation value corresponding to each path in the multi-path communication system are determined.

[0132] In some implementations, based on the rearranged reconstructed delay spectrum, the total number of paths in the multi-path communication system and the first delay estimation value corresponding to each path in the multi-path communication system are determined by a hard threshold decision module.

[0133] In some implementations, if the first measurement matrix is used to represent the measurement quantities corresponding to different powers in the energy spectrum of the multi-path communication system, under the constraint of the first constraint condition, the energy spectrum of the multi-path communication system is sparsely recovered according to the first measurement matrix and the observation vector.

[0134] In some implementations, based on the energy spectrum of the multi-path communication system, the first delay estimation value corresponding to each path in the multi-path communication system is determined.

[0135] In the above embodiments, the delay spectrum or the energy spectrum is obtained after discretization of the delay or the energy, which inevitably introduces a grid-off effect. To solve this problem, in some implementations, a plurality of grids are determined based on the first delay estimation value, and the reconstructed delay spectrum is further optimized by traversing the grids to suppress the off-grid error, which helps to further improve the accuracy of the delay estimation value of the multi-path communication system.

[0136] The method for optimizing the reconstructed delay spectrum is described below in conjunction with FIG. 6.

[0137] In step S610, a first prediction set corresponding to the multipath communication system is obtained, and a second delay estimation value corresponding to the multipath communication system is obtained.

[0138] In some implementations, the first prediction set includes a first delay estimation value corresponding to each path in the multipath communication system. For example, the first prediction set is τ pre (1) (2) (L) , τ (i) , …, τ pre , where L is the total number of paths, and τ (1) is a first delay estimation value corresponding to the i-th path in the multipath communication system.

[0139] In some implementations, the second delay estimation value is a delay estimation value with the minimum recovery residual.

[0140] In some implementations, based on the first delay estimation value corresponding to each path in the first prediction set, a plurality of first grids are determined, and the first grids are traversed to obtain the second delay estimation value.

[0141] In some implementations, the first grid is also referred to as a “local search grid”.

[0142] In step S620, the first delay estimation value corresponding to the path in the first prediction set is replaced by the second delay estimation value to obtain a second prediction set.

[0143] In some implementations, the second prediction set can be used to optimize the reconstruction delay spectrum described above. In other implementations, the second prediction set can be used as the first prediction set in step S610 to obtain a new second delay estimation value.

[0144] Hereinafter, with reference to FIG. 7, a method for obtaining the second prediction set is described by taking the second delay estimation value as an example of a delay estimation value with the minimum recovery residual.

[0145] In step S710, based on the i-th first delay value in the first prediction set, a local search grid of the k-th iteration is determined.

[0146] Suppose the first prediction set is τ pre (1) (2) (L) , where the local search grid of the k-th iteration of the i-th first delay estimation value is 2SΔτ k , where S is the range of local search, and Δτ k is the interval delay.

[0147] In step S720, one atomic value of the local search grid is used to replace the i-th first delay estimation value in the first prediction set.​​

[0148] Using the grid of τ (i) + sΔτ k (s = 0, ±1, …, ±S) to replace the first delay estimation value τ in the first prediction set (i) to form a new prediction set

[0149] In step S730, the atomic value in the new first prediction set is used to replace the atomic value corresponding to the delay grid T in the first prediction set, to obtain a new first measurement matrix.

[0150] Using the atomic value in the delay set to replace the corresponding T in the first measurement matrix A p to obtain a new first measurement matrix

[0151] In step S740, a convex optimization algorithm is used to solve the grid delay with the smallest recovery residual in multiple rounds of iteration, that is, the i-th second delay estimation value.

[0152] Definition solved by a convex optimization algorithm wherein the initial value of is the reconstruction delay spectrum obtained by the first neural network model above. Repeat steps S710 to S740, and after K rounds of iteration, solve the minimum R to obtain the grid delay τ (i) + sΔτ k which is the i-th second delay estimation value.

[0153] In step S750, the first delay estimation value corresponding to the path in the first prediction set is replaced by the second delay estimation value to obtain a second prediction set.

[0154] Replace the i-th second delay estimation value with the i-th first delay estimation value in the first prediction set to obtain a second prediction set.

[0155] In step S760, based on the second prediction set, a least squares method is used to solve a new reconstruction delay spectrum.

[0156] Based on the second delay estimation value, the LS algorithm can be used to calculate a new Repeat steps S710-S760 for L rounds of iteration until the second delay estimation value corresponding to each path in the first prediction set and the optimized

[0157] The method embodiments of the present application are described in detail above in combination with FIG. 1 to FIG. 7, and the device embodiments of the present application are described in detail below in combination with FIG. 8 to FIG. 9. It should be understood that the description of the method embodiments corresponds to the description of the device embodiments, and therefore, the parts not described in detail can be referred to the foregoing method embodiments.

[0158] FIG. 8 is a schematic diagram of a device for side channel analysis according to an embodiment of the present application. The device 800 shown in FIG. 8 includes an obtaining unit 810 and a processing unit 820.

[0159] The obtaining unit 810 is configured to obtain a first measurement matrix and an observation vector of a multipath communication system, where the first measurement matrix is used to represent measurement quantities corresponding to different time delays in a delay spectrum of the multipath communication system, and the observation vector is used to indicate observation values of channel frequency responses (CFRs) or channel impulse responses (CIRs) corresponding to multiple paths in the multipath communication system.

[0160] The processing unit 820 is configured to perform sparse recovery on the delay spectrum of the multipath communication system under a first constraint condition, according to the first measurement matrix and the observation vector, where the first constraint condition is used to constrain a relationship between a number of elements that are not 0 in the delay spectrum and a number of paths in the multipath communication system.

[0161] In some implementations, the processing unit 820 is configured to perform sparse recovery on the delay spectrum of the multipath communication system under the first constraint condition, according to the first measurement matrix and the observation vector, where the device includes: performing sparse recovery on the delay spectrum u of the multipath communication system under the first constraint condition, according to the first measurement matrix A and the observation vector h, by using the formula where the first constraint condition is represented as ‖u‖0<k, where k represents a maximum value of a number of allowed paths in the multipath communication system.

[0162] In some implementations, the device further includes: adjusting the first measurement matrix based on a first parameter and / or a second parameter to obtain an adjusted second measurement matrix, where the first parameter represents antenna gains corresponding to subcarriers in the multipath communication system, and the second parameter represents a sampling clock deviation between a receiving end and a sending end in the multipath communication system.

[0163] In some implementations, the second parameter is used to indicate that a channel impulse response (CIR) of a channel in the multipath communication system is increased by a first time offset in a time domain; and / or the second parameter is used to indicate that a phase of a channel frequency response (CFR) of the channel in the multipath communication system is increased by a first rotation phase in a frequency domain.

[0164] In some implementations, the second measurement matrix is represented as where g i represents the antenna gain corresponding to the i th subcarrier, δ f represents the sampling clock offset, f i represents the i th subcarrier, T j represents the j th value of the delay spectrum, F i represents the i th value of F, f (nT s ) represents the time-domain convolution of the transceiver filter at time nT s , T S is the sampling period, and n represents the n th sampling point in the time domain.

[0165] In some implementations, the apparatus further includes inputting the observation vector into a first neural network model for model prediction to obtain a total number of paths in the multi-path communication system and a first delay estimation value corresponding to each path in the multi-path communication system.

[0166] In some implementations, the first neural network model is trained based on a loss function, and the loss function is used to calculate the difference between a reconstructed delay spectrum and an original delay spectrum of the multi-path communication system; and / or the loss function is used to calculate the asymmetric error between the reconstructed delay spectrum and the original delay spectrum.

[0167] In some implementations, the apparatus further includes obtaining a first prediction set corresponding to the multi-path communication system, and a second delay estimation value corresponding to the multi-path communication system, wherein the first prediction set includes a first delay estimation value corresponding to each path in the multi-path communication system, and the second delay estimation value is a delay estimation value with the smallest recovery residual; and replacing the first delay estimation value corresponding to the path in the first prediction set with the second delay estimation value to obtain a second prediction set.

[0168] In optional embodiments, the obtaining unit 810 and the processing unit 820 can be a processor 910. The apparatus 800 can further include a transceiver 930 and a memory 920, as shown in FIG. 9.

[0169] FIG. 9 is a schematic structural diagram of a communication apparatus according to an embodiment of the present application. The dashed line in FIG. 9 indicates that the unit or module is optional. The apparatus 900 can be used to implement the method described in the above method embodiments. The apparatus 900 can be a chip, a terminal device, or a network device.

[0170] The apparatus 900 can include one or more processors 910. The processor 910 can support the apparatus 900 to implement the methods described in the foregoing method embodiments. The processor 910 can be a general purpose processor or a special purpose processor. For example, the processor can be a central processing unit (CPU). Alternatively, the processor can also be other general purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic, discrete hardware components, etc. The general purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0171] The apparatus 900 can also include one or more memories 920. The memory 920 stores programs, which can be executed by the processor 910, so that the processor 910 performs the methods described in the foregoing method embodiments. The memory 920 can be independent of the processor 910 or integrated in the processor 910.

[0172] The apparatus 900 can also include a transceiver 930. The processor 910 can communicate with other devices or chips through the transceiver 930. For example, the processor 910 can perform data transceiving with other devices or chips through the transceiver 930.

[0173] The embodiments of the present application also provide a computer readable storage medium for storing programs. The computer readable storage medium can be applied to the terminal or network device provided by the embodiments of the present application, and the programs make the computer execute the methods performed by the terminal or network device in the embodiments of the present application.

[0174] The embodiments of the present application also provide a computer program product. The computer program product includes programs. The computer program product can be applied to the terminal or network device provided by the embodiments of the present application, and the programs make the computer execute the methods performed by the terminal or network device in the embodiments of the present application.

[0175] The embodiments of the present application also provide a computer program. The computer program can be applied to the terminal or network device provided by the embodiments of the present application, and the computer program makes the computer execute the methods performed by the terminal or network device in the embodiments of the present application.

[0176] It should be understood that the terms "system" and "network" can be used interchangeably in this application. In addition, the terms used in this application are only used to explain the specific embodiments of the application, and are not intended to limit the application. The terms "first", "second", "third", and "fourth" and the like in the specification and claims of the application and the drawings are used to distinguish different objects, and are not used to describe a particular order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion.

[0177] In embodiments of the present application, the term "indicate" can be direct indication or indirect indication, or can represent an associated relationship. For example, A indicates B, which can mean that B can be obtained through A; or A indirectly indicates B, for example, A indicates C, and B can be obtained through C; or A and B have an associated relationship.

[0178] In embodiments of the present application, the term "corresponding" can represent a direct or indirect relationship between the two, or an associated relationship between the two, or an indication and the indicated, configuration and configuration, etc.

[0179] In embodiments of the present application, the term "and / or" is only used to describe the associated relationship between the associated objects, which means that there can be three relationships, for example, A and / or B, which can mean that A exists alone, A and B exist together, and B exists alone. In addition, the character " / " in this paper generally represents an "or" relationship between the associated objects before and after.

[0180] In various embodiments of the present application, the size of the serial number of the above processes does not mean the order of execution, and the execution order of the processes should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0181] In several embodiments provided by the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division, and actual implementation can be in another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the shown or discussed mutually can be indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other form.

[0182] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e., may be located in one place, or may be distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment of the present application according to actual needs.

[0183] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can be physically present alone, or two or more units can be integrated into one unit.

[0184] In the above embodiments, all or part can be realized by software, hardware, firmware or any combination thereof. When realized by software, all or part can be realized in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed by a computer, the computer instructions produce the processes or functions described in the embodiments of the present application. The computer can be a general-purpose computer, a special-purpose computer, a computer network or other programmable device. The computer instructions can be stored in a computer readable storage medium or transferred from one computer readable storage medium to another, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer readable storage medium can be any available medium readable by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (such as floppy disk, hard disk, magnetic tape), optical media (such as digital video disc (DVD)) or semiconductor media (such as solid state disk (SSD)) and the like.

[0185] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for wireless communication, comprising: The method comprises: obtaining a first measurement matrix of a multipath communication system and an observation vector, wherein the first measurement matrix is used to represent measurement quantities corresponding to different time delays in a delay spectrum of the multipath communication system, and the observation vector is used to indicate observation values of channel frequency responses (CFRs) or channel impulse responses (CIRs) corresponding to multiple paths in the multipath communication system; sparsely recovering the delay spectrum of the multipath communication system according to the first measurement matrix and the observation vector under the constraint of a first constraint condition, wherein the first constraint condition is used to constrain the relationship between the number of non-zero elements in the delay spectrum and the number of paths in the multipath communication system.

2. The method of claim 1, wherein, The sparsely recovering the delay spectrum of the multipath communication system according to the first measurement matrix and the observation vector under the constraint of the first constraint condition comprises: Under the constraint of the first constraint condition, according to the first measurement matrix A and the observation vector h, the first constraint condition is solved by formula sparsely recovering the delay spectrum u of the multipath communication system, the first constraint condition is represented as ‖u‖0<k, wherein k represents the maximum value of the number of allowed paths in the multipath communication system.

3. The method of claim 1 or 2, wherein, The method further comprises: adjusting the first measurement matrix based on a first parameter and / or a second parameter to obtain an adjusted second measurement matrix, wherein the first parameter represents antenna gains corresponding to subcarriers in the multipath communication system, and the second parameter represents a sampling clock deviation between a receiving end and a transmitting end in the multipath communication system.

4. The method of claim 3, wherein, The second parameter is used to indicate that a channel impulse response (CIR) of a channel in the multipath communication system is increased by a first time offset in the time domain; and / or The second parameter is used to indicate that a phase of a channel frequency response (CFR) of a channel in the multipath communication system is increased by a first rotation phase in the frequency domain.

5. The method of claim 3 or 4, wherein, The second measurement matrix is expressed as where g i represents the antenna gain corresponding to the i-th subcarrier, δ f represents the sampling clock deviation, f i represents the i-th subcarrier, T j represents the j-th value of the time delay spectrum, F i represents the i-th value of F, f(nT s ) represents the time domain convolution of the transceiver filter at time nT s , T S is the sampling period, and n represents the n-th sampling point in the time domain.

6. The method of any one of claims 3-5, wherein, The method further comprises: inputting the observation vector into a first neural network model for model prediction to obtain the total number of paths in the multipath communication system and a first time delay estimation value corresponding to each path in the multipath communication system.

7. The method of claim 6, wherein, The first neural network model is trained based on a loss function, wherein the loss function is used to calculate the difference between a reconstructed delay spectrum and an original delay spectrum of the multipath communication system; and / or the loss function is used to calculate an asymmetric error between the reconstructed delay spectrum and the original delay spectrum.

8. The method of claim 6 or 7, wherein, The method further comprises: obtaining a first prediction set corresponding to the multipath communication system and a second time delay estimation value corresponding to the multipath communication system, wherein the first prediction set comprises a first time delay estimation value corresponding to each path in the multipath communication system, and the second time delay estimation value is a time delay estimation value with the smallest recovery residual; replacing the first time delay estimation value of the corresponding path in the first prediction set with the second time delay estimation value to obtain a second prediction set. The method comprises:

9. An apparatus for wireless communication, the apparatus comprising: ​ An acquisition unit is configured to acquire a first measurement matrix and an observation vector of a multipath communication system, wherein the first measurement matrix is used to represent measurement quantities corresponding to different time delays in a delay spectrum of the multipath communication system, and the observation vector is used to indicate observation values of channel frequency responses (CFRs) or channel impulse responses (CIRs) corresponding to multiple paths in the multipath communication system. A processing unit is configured to perform sparse recovery on the delay spectrum of the multipath communication system under a first constraint condition, according to the first measurement matrix and the observation vector, wherein the first constraint condition is used to constrain a relationship between a number of non-zero elements in the delay spectrum and a number of paths in the multipath communication system.

10. The apparatus of claim 9, wherein, The apparatus is configured to perform sparse recovery on the delay spectrum u of the multipath communication system under a first constraint condition, according to the first measurement matrix and the observation vector, wherein the first constraint condition is represented as ‖u‖0<k, and k represents a maximum value of a number of allowed paths in the multipath communication system. Under the constraint of the first constraint condition, according to the first measurement matrix A and the observation vector h, the first constraint condition is solved by formula The apparatus further includes: The first measurement matrix is adjusted based on a first parameter and / or a second parameter to obtain an adjusted second measurement matrix, wherein the first parameter represents antenna gains corresponding to subcarriers in the multipath communication system, and the second parameter represents a sampling clock deviation between a receiving end and a sending end in the multipath communication system.

11. The apparatus of claim 9 or 10, wherein, The second parameter is used to indicate that a channel impulse response (CIR) of a channel in the multipath communication system is increased by a first time offset in a time domain; and / or The second parameter is used to indicate that a phase of a channel frequency response (CFR) of the channel in the multipath communication system is increased by a first rotation phase in a frequency domain.

12. The apparatus of claim 11, wherein, The apparatus further includes: The observation vector is input into a first neural network model to perform model prediction, so as to obtain a total number of paths in the multipath communication system and a first time delay estimation value corresponding to each path in the multipath communication system.

13. The apparatus of claim 11 or 12, wherein, The second measurement matrix is expressed as Wherein, g i represents the antenna gain corresponding to the i-th subcarrier, δ f represents the sampling clock deviation, f i represents the i-th subcarrier, T j represents the j-th value of the time delay spectrum, F i represents the i-th value of F, f(nT s ) represents the time domain convolution of the transceiver filter at time nT s , T S is the sampling period, and n represents the n-th sampling point in the time domain.

14. The apparatus of any one of claims 11-13, wherein, The first neural network model is trained based on a loss function, wherein the loss function is used to calculate a difference between a reconstructed delay spectrum and an original delay spectrum of the multipath communication system, and / or the loss function is used to calculate an asymmetric error between the reconstructed delay spectrum and the original delay spectrum. The apparatus further includes:

15. The apparatus of claim 14, wherein, A first prediction set corresponding to the multipath communication system and a second time delay estimation value corresponding to the multipath communication system are acquired, wherein the first prediction set includes first time delay estimation values corresponding to each path in the multipath communication system, and the second time delay estimation value is a time delay estimation value with a minimum recovery residual; 16. The apparatus of claim 14 or 15, wherein, The first time delay estimation value corresponding to a path in the first prediction set is replaced by the second time delay estimation value, so as to obtain a second prediction set. The apparatus includes: A memory is configured to store instructions; 17. An apparatus for wireless communication, the apparatus comprising: A processor is configured to execute the instructions stored in the memory, so as to perform the method in any one of claims 1-8. ​ ​ 18. A communication device, characterized by A communication device comprising a transceiver, a memory for storing a program, and a processor for invoking the program in the memory and controlling the transceiver to receive or send signals, so as to make the communication device perform the method according to any one of claims 1-8.

19. An apparatus, comprising: A device comprising a processor for invoking a program from a memory, so as to make the device perform the method according to any one of claims 1-8.

20. A chip, characterized by A chip comprising a processor for invoking a program from a memory, so that the device installed with the chip performs the method according to any one of claims 1-8.

21. A computer-readable storage medium, characterized in that, A computer program stored on a computer readable medium, the program causing a computer to perform the method according to any one of claims 1-8.

22. A computer program product, characterised in that, A computer program product comprising a program, the program causing a computer to perform the method according to any one of claims 1-8.

23. A computer program, characterized in that, The computer program causes a computer to perform the method according to any one of claims 1-8.

Images