A low-complexity method for signal recovery and sensing positioning estimation in multi-user communication

Abstract

The present invention discloses a low-complexity multi-user communication signal recovery and perception positioning estimation method, which generates and transmits an integrated waveform of superimposed pilot and data at the transmitting end, so that the communication signal and the perception signal are linearly superimposed in the symbol domain using a power allocation method in the same time-frequency resource block, and the receiving end uses a designed low-complexity algorithm to process the received signal to realize the function of simultaneous communication and positioning of multiple users; including: using SCMA technology to design an encoder for superimposed symbols at the transmitting end and a decoder for detection and demodulation at the receiving end, and using the UAMP-SBL algorithm at the receiving end to process the received signal to separate the positioning coordinates of the communication signals of multiple users. The present invention is based on the original RIS-assisted communication system, and can use a low-complexity algorithm to simultaneously obtain the communication data of multiple users and accurate user positioning coordinates without adding additional hardware equipment.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to a low-complexity multi-user communication signal recovery and perception positioning estimation method. Background Art

[0002] In order to support emerging applications that require both low-latency perception and high-speed communication, the next generation of wireless networks proposes integrated communication and perception technology to use the same wireless signals and hardware infrastructure to achieve wireless perception and communication at the same time. In communication systems, large-bandwidth millimeter-wave bands have been proven to be able to achieve high-speed communication to meet real-time massive data transmission. At the same time, in existing perception systems, millimeter-wave radars have been widely used to provide accurate perception information. Because of the excellent performance of millimeter-wave bands in communication and perception systems, the use of millimeter-wave signals in integrated communication and perception systems is considered to have great application prospects and has also been a hot topic of research in academia and industry in recent years.

[0003] However, millimeter wave signals can cause severe propagation losses, which often block the LoS path between the receiver and the transmitter, hindering the reliable connection service of integrated communication and perception. RIS can adjust the phase shift of the reflective element by digital control to create a virtual LoS path between the base station and the user, and is regarded as an effective solution to solve the stable connection of the integrated communication and perception system in the millimeter wave frequency band. Specifically, RIS is a completely passive two-dimensional array with many sub-wavelength controllable elements, which can flexibly change the physical properties of the incident electromagnetic wave by adjusting the phase. In addition, RIS is generally made of cheap hardware components and has no RF link, which greatly reduces hardware costs and energy consumption. At present, a large number of studies have proved that the integrated communication and perception system assisted by RIS has significantly improved both communication and perception performance.

[0004] In the design of RIS-assisted millimeter wave band communication and perception integrated system, a practical and efficient method is to use the existing communication modulation method, protocol architecture and hardware infrastructure combined with perception signal processing technology to achieve simultaneous communication and perception functions. This solution realizes the perception of targets in the environment while transmitting data, avoiding the hardware cost of additional deployment of perception networks. In existing technologies, the realization of perception functions will allocate a part of additional time or frequency resources so that the perception pilot uses this part of independent resources to realize the perception function, so there are some methods to study the performance tradeoffs between communication signals and perception pilots in time and frequency resources. However, there are some limitations: the perception pilot consumes additional time or frequency resources, which will damage the spectrum efficiency of the original communication system; and the simultaneous realization of communication and perception functions in the RIS-assisted system has high requirements for computing resources and relies on channel state information. In high-speed mobile scenarios such as the Internet of Vehicles, it may be difficult to obtain channel state information in a timely and accurate manner, thereby damaging the accuracy of communication and perception.

[0005] At the same time, as the number of connections through integrated sensing devices increases, traditional orthogonal multiple access technologies such as time division multiple access and frequency division multiple access can no longer meet the needs of future wireless communication development. As a new multiple access technology, sparse code division multiple access (SCMA) technology combines low-density signal technology with code division multiple access technology, which can effectively solve the resource allocation problem of multiple devices accessing at the same time, that is, using limited resources to achieve high-speed communication and high-precision positioning functions.

[0006] Based on the above problems, in order to improve the performance of RIS-assisted communication and perception, and at the same time improve the spectrum efficiency of the communication-perception integrated system, a design method for communication-perception integrated waveform based on superimposed symbols is proposed to simultaneously realize the functions of communication and user positioning. At the same time, in order to meet the low latency requirements of emerging communication-perception integrated applications, it is necessary to design low-complexity algorithms to effectively complete the communication and positioning functions. Summary of the invention

[0007] The present invention aims to solve the problems of low spectrum efficiency, poor accuracy and high complexity of existing algorithms for realizing communication and positioning at the same time. By transmitting a communication-perception integrated waveform formed by superimposing communication symbols and pilot symbols, the communication signal and the perception signal are linearly superimposed in the symbol domain through a power allocation coefficient. At the same time, a low-complexity signal processing method at the receiving end is proposed, thereby increasing the accurate positioning function with almost no loss of communication performance.

[0008] The specific technical solution adopted by the present invention to solve its technical problem is: a low-complexity multi-user communication signal recovery and perception positioning estimation algorithm based on the communication perception integrated waveform of the SCMA technology superimposed symbols, which can be applied to the RIS-assisted millimeter wave communication perception integrated scenario;

[0009] The method provided by the present invention realizes joint communication data detection and multi-user positioning parameter extraction by performing signal processing on the receiving end symbols, and realizes high-speed communication rate and high-precision perception function at the same time. In specific applications, the communication perception integrated waveform based on superimposed symbols generated by the present invention can support multiple modulation methods and parameter settings in the existing wireless communication system to realize efficient uplink communication perception link.

[0010] The present invention proposes a low-complexity multi-user communication signal recovery and perception positioning estimation method, which comprises the following steps:

[0011] S1. The user transmitting end maps the communication data and pilot data through the SCMA codebook to obtain communication symbols and pilot symbols and linearly superimposes them according to the power allocation coefficient to obtain superimposed symbols and transmit them to the RIS end;

[0012] S2, the base station receiving end receives the signal reflected by the RIS end, and stacks the signals received in multiple time slots to build a receiving matrix;

[0013] S3, converting the extraction of the environmental angle parameter into an estimation of the angle grid value (θ, φ) and the sparse vector ψ; obtaining an initial estimation value of the equivalent channel between the base station and the user through the estimation value of the sparse vector;

[0014] S4, according to the initial estimated value of the equivalent channel, subtract the received part corresponding to the sensing pilot from the received matrix, and obtain the initial estimated value of the communication data using the linear minimum mean square error method and the SCAM decoder;

[0015] S5. Iterative calculation of the initial estimate of the communication data: The estimated value of the angle grid (θ (j) , ψ (j) and the sparse vector estimate ψ (j) , reduce the dimension of the sensing matrix and the sparse vector, obtain the corresponding channel matrix and subtract the receiving part corresponding to the sensing pilot from the receiving matrix T to determine the estimated value of the communication data;

[0016] After the iteration is completed, the final estimated values ​​of the sparse vector, angle grid and communication data are obtained; finally, the communication data, angle parameters in the environmental channel and user coordinates are restored.

[0017] Furthermore, the power allocation coefficient ξ kis the power coefficient of the communication symbol allocated by the kth user to the entire symbol; satisfying 0<ξ k <1.

[0018] Furthermore, the extraction of environmental angle parameters is converted into the estimation of angle grid values ​​(θ, φ) and sparse vector ψ as follows: the millimeter wave channel between the RIS and each user is converted into the form of a dictionary matrix Ω(θ, φ) multiplied by a sparse vector ψ, wherein the dictionary matrix is ​​composed of array responses corresponding to (θ, φ) after the azimuth and elevation angles are uniformly quantized in the angle domain.

[0019] Furthermore, the initial estimated value of the equivalent channel between the base station and the user is obtained by using the estimated value of the sparse vector: according to the known information of the pilot symbol, the initial estimated value of the sparse vector ψ is obtained by using the least squares estimation. (0) , where the superscript (j) represents the result obtained at the jth iteration;

[0020] According to the obtained ψ (0) , the equivalent channel initial estimation value H between the base station and the user is obtained through affine transformation (0) ;

[0021] According to H (0) , subtract the received part corresponding to the sensing pilot from the received matrix Y, and use the linear minimum mean square error method and SCAM decoder to obtain the initial estimate of the communication data Will After demodulation, re-modulation is performed to obtain the processed

[0022] Furthermore, the SCAM decoder uses a message passing algorithm to update the probability corresponding to each codeword of each user, and c ∈{1,...,N c The codeword with the maximum a posteriori probability is selected from the codewords as the decoded data; in the process of calculating the initial estimate of the equivalent channel, all codewords are set to the same prior probability.

[0023] Furthermore, the iterative calculation is specifically as follows:

[0024] Based on the estimated data value in the previous iteration Recover the superimposed symbol estimate X (j-1) , the estimated value of the grid (θw) is obtained by the off-grid unitary approximate message passing sparse Bayesian learning method j) ,φ (j) ) and the sparse vector estimate ψ (j) , where j = 1 during the initial value estimation process;

[0025] Select the obtained sparse vector estimate ψ (j)Greater than ∈ ψ ||ψ (j) || the number of units, where ∈ ψ is to set the threshold and discard the corresponding grid value (θ (j) ,φ (j) ), reducing the dimensions of the perception matrix and sparse vectors for updating grid values ​​and sparse vectors with low complexity;

[0026] After reducing the dimensions of the perception matrix and sparse vector, the corresponding channel matrix H is obtained (j) , subtract the received part corresponding to the sensing pilot from the received matrix Y, and calculate the estimated value of the communication data After demodulation, it is re-modulated to obtain the processed

[0027] Repeat the above steps until the number of iterations j reaches the preset maximum value j max or the log-likelihood function reaches stability;

[0028] Through the final (θ (J) ,φ (J) ) and ψ (J) , where J is the final number of iterations, the angle parameters in the communication data and the environmental channel can be restored at the receiving end.

[0029] Furthermore, the estimated value of the grid (θ (j) ,φ (j) ) and the sparse vector estimate ψ (j) Specifically: For the following received signal y U =D ψ +n U To recover the sparse vector ψ, where D is the perception matrix, n u is noise, including the following processes:

[0030] S501: Perform singular value decomposition on the perception matrix D to obtain Multiply the two ends of the received signal by the matrix get in

[0031] S502: Set sparse vector Each element The prior distribution of where γ g Satisfy the parameter a ψ and b ψ The gamma distribution, p(γ ψ,g α ψ ,bψ )=I(γ ψ,g ; 1+a ψ ,b ψ );a ψ and b ψ is an adjustable parameter, and the noise parameter is modeled as β ψ , after introducing the hard constraint factor κ=Λψ, we have p(κ q |ψ)=δ(κ q -Λ(q,:)ψ); the noise parameter is modeled as β ψ , after introducing the hard constraint factor κ=Λψ, we have p(κ q |ψ)=δ(κ q -Λ(q,:)ψ);

[0032] S503: Joint distribution based on conditions

[0033] p(ψ, k, γ ψ , β ψ |r)∝p(r|k,β ψ )p(k|ψ)p(ψ|γ ψ )p(γ ψ |a ψ , b ψ )

[0034] ·

[0035] ·

[0036] The factor graph of the message passing method is used to iteratively update and finally obtain the sparse vector ψ;

[0037] Repeat the process below:

[0038] renew renew renew renew renew renew Update s = τ s (rp); update renew renew renew renew renew Update u←u+1; iterate the update process until or u>u max Finally, we get the sparse vector ψ (1) ;

[0039] in and They are vectors and The gth element of ; G1, G2, G3 and G4 are the number of quantized copies in the angle domain; the above τ p ,p,v κ , τ s ,s, q, and a ∈ are intermediate variables in the update process, u is the number of iterations, δ ψ and u max The convergence threshold and maximum number of iterations are set respectively; M is the number of base station antennas, and T is the number of stacked time slots.

[0040] Furthermore, the updating method after reducing the dimension of the operation matrix / vector in the signal processing at the receiving end solves the problem of X in (j) , and The recovery process includes the following:

[0041] Recover data using the SCMA decoder

[0042] Using low-complexity message passing methods, we can recover

[0043] Use subgradient descent to update

[0044] Repeat the process until two iterations The difference converges.

[0045] Furthermore, the position coordinates of each user are estimated and determined by estimating and restoring the azimuth and elevation angles from one RIS to the user. Specifically, the position coordinates and relative distance of the user's partial axis are determined by the formula of the arrival angle and coordinates between the RIS and the user, and then the full position coordinates of the user are obtained according to the relationship between the base station and the user.

[0046] On the other hand, the specification also provides a system for low-complexity multi-user communication signal recovery and perception positioning estimation method, including the following modules:

[0047] Transmit signal module: used to map the communication data and pilot data through the SCMA codebook, obtain the communication symbols and pilot symbols, and linearly superimpose them according to the power allocation coefficient to obtain the superimposed symbols and transmit them to the RIS end;

[0048] Receiving signal module: used to receive the signal reflected by the RIS end, and stack the signals received in multiple time slots to build a receiving matrix;

[0049] Initialization module: used to obtain the initial estimation value of communication data using the linear minimum mean square error method and SCAM decoder according to the information of the pilot symbol;

[0050] Parameter estimation module: configured to obtain a grid estimate and a sparse vector estimate through a unitary approximate message passing sparse Bayesian learning method off the grid based on the initial estimate obtained by the initialization module;

[0051] A user positioning module, used to obtain the position coordinates of all users according to the angle domain grid values ​​and the estimated values ​​of the sparse vector obtained by the parameter estimation module;

[0052] Data detection module: used to calculate and restore communication data values ​​based on the estimated values ​​obtained by the parameter estimation module, and includes modulation and demodulation functions.

[0053] Through the above steps, the efficiency and accuracy of joint communication and positioning in RIS-assisted multi-user systems can be effectively improved, and the communication efficiency and positioning accuracy of the communication-aware integrated waveform based on low-complexity algorithms and superimposed symbols can be improved.

[0054] Compared with the existing invention, the beneficial effects of the present invention are:

[0055] The present invention provides a new waveform design scheme in a communication perception integrated system and a low-complexity multi-user communication signal recovery and perception positioning estimation algorithm. First, the communication signal and the perception pilot are superimposed in the symbol domain in the form of power allocation, and the positioning function can be added on the basis of the original communication function; then the UAMP-SBL method is used to solve the compressed sensing problem with low complexity, and the dimensions of the perception matrix and sparse vector are significantly reduced after the first iteration, thereby reducing the computational complexity in the matrix / vector operation process, and realizing low-complexity and high-precision positioning perception. In addition, the existing scheme has little change to the existing communication protocol and hardware equipment, and the positioning function can be added to the original communication system flexibly and at low cost.

[0056] This method comprehensively considers the performance of communication and perception, and linearly superimposes communication symbols and pilot symbols in the same symbol domain according to the power allocation coefficient, which can not only realize the positioning function of multiple users, but also ensure that the communication capability is not affected, and truly realize the integrated function of communication and perception;

[0057] This method uses SCMA technology to achieve efficient use of spectrum resources for multiple users, and improve communication efficiency and positioning accuracy under limited spectrum resources;

[0058] This method uses UAMP-SBL technology to solve the recovery of sparse vectors in compressed sensing problems. Compared with the traditional SBL method, the designed UAMP-SBL avoids the matrix inversion operation, which greatly reduces the complexity while ensuring that the accuracy is basically unaffected.

[0059] This method significantly reduces the dimensions of matrices and vectors according to the form of sparse vectors after the first iteration, thereby greatly reducing the computational complexity of matrix / vector operations in subsequent iterations;

[0060] Based on the design of the transmitting synaesthesia integrated waveform and the receiving end low-complexity recovery method, efficient communication and perception functions can be achieved simultaneously, and the existing communication system can be fully utilized to achieve centimeter-level high-precision positioning function. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 is a block diagram of the method of the present invention;

[0062] Figure 2 It is a simulation diagram of the signal-to-noise ratio and positioning error under different transmission strategies according to the method of the present invention;

[0063] Figure 3 It is a simulation diagram of the signal-to-noise ratio and average spectrum efficiency under different transmission strategies according to the method of the present invention;

[0064] Figure 4 It is a simulation diagram of the number of transmission symbols and positioning errors of the method of the present invention compared with different design methods;

[0065] Figure 5 It is a simulation diagram of the number of transmission symbols and throughput of the method of the present invention compared with different design methods. DETAILED DESCRIPTION

[0066] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0067] The present invention studies the uplink of K users to the base station assisted by I RIS (I ≥ 2). Because the millimeter wave signal is easily blocked, it is assumed that the direct path from the base station to the user is blocked by obstacles. The base station is composed of M uniform linear arrays placed along the Y axis, and the i-th RIS is composed of N i =N i，y ×N i，z The base station is composed of a uniform rectangular array in the YoZ plane. The base station reflects the user's uplink signal from the RIS, and at the same time detects the user's communication signal and senses the environmental parameters between the base station and the RIS.

[0068] Communication perception integrated system modeling: This paper studies the uplink of multiple users to the base station assisted by I RIS (I ≥ 2). Because the millimeter wave signal is easily blocked, it is assumed that the direct path from the base station to the user is blocked by obstacles. The base station is composed of M uniform linear arrays placed along the Y axis, and the i-th RIS is composed of N i =N i，y ×N i，z The base station is composed of a uniform rectangular array in the YoZ plane. The base station reflects the user's uplink signal from the RIS, and at the same time detects the user's communication signal and senses the environmental parameters between the base station and the RIS.

[0069] In the tth time slot, the base station s All user uplink signals reflected by RIS are received on subcarriers

[0070]

[0071] Among them, H r，i is the channel from the ith RIS to the base station, θ i，t is the phase of the i-th RIS at the t-th time slot, h b，ik is the channel from the kth user to the i-th RIS, is the nth time slot s The noise of each subcarrier.

[0072] Superposition symbol modeling: Communication symbols and pilot symbols are linearly superimposed in the same symbol domain according to the power allocation coefficient, that is, in and Denote communication and pilot symbols respectively, ξ k represents the power allocation factor for communication data of the kth user, satisfying 0<ξ k <1.

[0073] Sparse modeling of millimeter wave channels in the angle domain: In the millimeter wave frequency band, the wireless channel follows a geometric channel model. Specifically, assuming that there is only a LoS path between the i-th RIS and the k-th user, the corresponding millimeter wave channel model is

[0074]

[0075] where α b，ik is the path gain of the corresponding channel, and the array responses are The corresponding effective arrival angle and departure angle are and in and are the azimuth and elevation angles of arrival from the i-th RIS to the k-th user, respectively, and d BS and are the intervals between the base station and the i-th RIS array unit, and λ is the carrier wavelength.

[0076] The millimeter wave angle domain By selecting the grid points and The millimeter wave channel can be converted into

[0077]

[0078] in

[0079]

[0080] is about the angle domain grid points and The dictionary matrix composed of array responses, ψ i，k is a sparse vector, and the positions of non-zero elements in the vector correspond to the positions of elements selected from the grid points. Specifically, set and The elements in the angular domain are uniformly quantized into G1 and G2 equal parts, where G1 and G2 are both configurable parameters.

[0081] refer to Figure 1 As shown, the specific implementation steps of the present invention are:

[0082] Step 1: For the nth s subcarriers, in the tth time slot, the transmitter generates and transmits the superimposed symbol in and They are the communication symbol and pilot symbol generated in the tth time slot, and they are all codewords mapped through the SCMA codebook. k is the power coefficient of the communication symbol allocated by the kth user to the entire symbol, satisfying 0<ξ k <1;

[0083] Step 2: For the nth s subcarriers, in the tth time slot, the base station receives the symbol Among them, H r，i is the channel from the ith RIS to the base station, θ i，t is the phase of the i-th RIS at the t-th time slot, h b，ik is the channel from the kth user to the i-th RIS, is the noise of the tth time slot. By stacking the nth s The base station receives symbols of T time slots of subcarriers and constructs the receiving matrix:

[0084]

[0085] The corresponding RIS matrix is ​​Θ i =[θ i，1 , ..., θ i，T ], the transmit signal matrix is The noise matrix is

[0086] Step 3: By using the LSE method through the sensing pilot, the initial channel estimate from the kth user to the i-th RIS can be obtained. The superscript (0) represents the initialization process, and further obtains the cascade channel estimation value from the base station to the kth user The received signal minus the estimated pilot part can be written as in is in the nth s subcarriers, the pilot matrix transmitted in the tth time slot is used as the estimated value of the data symbol, that is, in is in the nth s subcarriers, the communication symbol matrix transmitted in the tth time slot, is in the nth s The set of users transmitting data simultaneously in the subcarriers, is equivalent to noise, assuming the variance is N0. Using the low-complexity message passing method to recover the data symbols in step 3, it can be written as in and h k，t The matrices are and The element at row m and column t2, is the corresponding noise, m = mod(t, M), t2 = [t / M];

[0087] Step 4: Use the message passing algorithm to update the probability of each codeword of each user. c codewords, so that the nth c ∈{1,...,N c} Select the maximum a posteriori probability codeword from the codewords as the decoded data;

[0088] For the SCMA decoder process, let and Represents the function nodes and variable nodes in the factor graph, then from the function nodes Pass to variable node The information can be written

[0089]

[0090] The likelihood function is written as is the kth user in the nth sThe nth subcarrier c Code words, It is a slave node arrive External information can be updated through the following methods

[0091]

[0092] Among them G k represents the subcarrier index of the kth user, The iterative process is initialized by assuming that all codewords have the same prior probability, such as Repeat the above process until the posterior probability value converges. Assuming that the number of iterations at convergence is S, the estimated data symbol It can be expressed by the selected transmission codeword as

[0093] Step 5: According to the kth user obtained in step 4, s Data symbol estimation value under subcarrier The received signal is vectorized and written as in It is nth s The noise vector of the subcarriers is substituted into h b，ik The millimeter wave angular domain sparse channel model is obtained

[0094] in N1 and N2 are the number of units in the first and second RIS blocks, respectively. Next, we use off-grid unitary approximate message passing (UAMP) sparse Bayesian learning (SBL) to obtain the sparse vector ψ (1) The superscript (1) represents the parameter value obtained in the first iteration.

[0095] The received signal y is trained using the off-grid Unitary Approximate Message Passing (UAMP) sparse Bayesian Learning (SBL) method. U =Dψ+n U To recover the sparse vector ψ, first perform singular value decomposition on the matrix D to obtain Multiply the two ends of the received signal by the matrix get in Furthermore, assuming that the sparse vector Each element The prior distribution of where γ g Satisfies the gamma distribution, such as p(γ ψ，g |a ψ , bψ )=Γ(γ ψ，g ; 1+a ψ , b ψ ), where a ψ and b ψ is an adjustable parameter, and the noise parameter is modeled as β ψ , after introducing the hard constraint factor κ=Λψ, we have p(k q |ψ)=δ(k q -Λ(q,:)ψ);

[0096] According to the following joint distribution

[0097]

[0098] The factor graph of the message passing method is used to iteratively update the final sparse vector ψ and repeat the following process: Update renew renew renew renew renew Update s = τ s (rp); update renew renew renew renew renew Update u←u+1; iterate the update process until or u>u max Finally, we get the sparse vector ψ (1) .in and They are vectors and The gth element of . G3 and G4 are the quantized shares; the above τ p ,p,v κ , τ s ,s, and a ∈ are intermediate variables in the update process, u is the number of iterations, δ ψ and u max The convergence threshold and maximum number of iterations are set respectively. M is the number of base station antennas, and T is the number of stacked time slots.

[0099] Step 6: Based on the data obtained in steps 4 and 5 and ψ (1) , using sub-gradient descent to obtain the grid value

[0100] The update of grid value solves the following offline problem arg minη ||y U -V2B U2 (η)ψ (1) || 2 To perform the LSE method, Since the problem is a nonlinear function of η, it is difficult to obtain a closed-form expression for updating η.

[0101] Therefore, the subgradient descent method with adjustable step size is used, and the update formula of η is The superscript (j) is the jth iteration update value. is the gradient direction of the objective function, The step size, η, is adjustable according to the number of iterations. max and η min are the maximum and minimum values ​​of the selected grid values, c2 and d2 are the learning rate and decay factor respectively.

[0102] Step 7: Based on the sparse vector ψ obtained in step 5 (1) In specific form, select the original perception matrix D(X (j) ,η) and the sparse vector ψ (1) The corresponding rows and columns produce the perception matrix after the reduced dimension and Further iterative updates through matrix and vector operations with significantly reduced dimensions and The specific dimension reduction is: select the sparse vector estimate ψ (j) Greater than ∈ ψ ,||ψ (j) || the number of units, where ∈ ψ is to set the threshold and discard the corresponding grid value (θ (j) ,φ (j) ), reducing the dimensions of the perception matrix and sparse vectors for updating grid values ​​and sparse vectors with low complexity;

[0103] The updated method after reducing the dimension of the operation matrix / vector solves the j-th iteration In and The update problem is to repeat the following process: according to the designed SCMA decoder, the corresponding data matrix is ​​restored. The corresponding sparse vector can be recovered using a low-complexity message passing method Use subgradient descent to update Until two iterations The difference converges.

[0104] Step 8: By obtaining the effective arrival angles of I RIS to K users, the location coordinates of K users can be obtained through the conversion relationship.

[0105] Furthermore, the updated method after reducing the dimension of the operation matrix / vector solves the problem of X in (j+1) , and The update problem is to repeat the following process: according to the designed SCMA decoder, the corresponding data matrix is ​​restored. The corresponding sparse vector can be recovered using a low-complexity message passing method Use subgradient descent to update Until two iterations The difference converges.

[0106] Furthermore, assuming that the location coordinates of the i-th RIS and the k-th user are q Ri =[q Ri，x ,q Ri，y ,q Ri，z ] and p Uk =[p Uk，x , p Uk，y , p Uk，z ], the distance between them is defined as d U2R，ik =||p Uk -q Ri ||. When I = 2, the effective arrival angles of the two RIS blocks and the kth user are and The y-axis and z-axis coordinates of the k-th user's position can be obtained through these equations. and the distance from the two RIS to the kth user Because the base station and the user are located at the same end of the RIS, the x-axis coordinate of the kth user's position is and In this way, the location coordinates of the kth user (p Uk，x , p Uk，y , p Uk，z ).

[0107] On the other hand, an embodiment of the present invention further provides a system for a low-complexity multi-user communication signal recovery and perception positioning estimation method, comprising the following modules:

[0108] Transmit signal module: used to map the communication data and pilot data through the SCMA codebook, obtain the communication symbols and pilot symbols, and linearly superimpose them according to the power allocation coefficient to obtain the superimposed symbols and transmit them to the RIS end;

[0109] Receiving signal module: used to receive the signal reflected by the RIS end, and stack the signals received in multiple time slots to build a receiving matrix;

[0110] Initialization module: used to obtain the initial estimation value of communication data using the linear minimum mean square error method and SCAM decoder according to the information of the pilot symbol;

[0111] Parameter estimation module: configured to obtain a grid estimate and a sparse vector estimate through a unitary approximate message passing sparse Bayesian learning method off the grid based on the initial estimate obtained by the initialization module;

[0112] A user positioning module, used to obtain the position coordinates of all users according to the angle domain grid values ​​and the estimated values ​​of the sparse vector obtained by the parameter estimation module;

[0113] Data detection module: used to calculate and restore communication data values ​​based on the estimated values ​​obtained by the parameter estimation module, and includes modulation and demodulation functions.

[0114] The functions and effects of the present invention are further illustrated by the following simulation experiments:

[0115] (1) Simulation conditions

[0116] Assume that the number of RIS is I = 2, the number of RIS configuration units is N1 = N2 = 12 × 12, and the locations of base stations and RIS are q B =[0,0,0],q R1 =[-30, 28, 21], q R2 = [-20, 30, 20], meters (m). The number of users is K = 6, and they are randomly distributed in a cubic space p Uk ∈([10, 30], [5, 25], [0, 20]). Set the number of base station antennas M = 16, the number of transmission symbols T = 256, and the other parameters are set as follows: max 100, δ ψ =δ η =10 -3 , G1=G2=16 and ξ k =0.5.

[0117] In the simulation, we compared the performance of the method of the present invention with orthogonal frequency division multiple access (OFDMA) and superposition symbol schemes and the pure pilot and semi-RIS schemes in the prior art in terms of positioning error and spectrum efficiency. The positioning error is The signal-to-noise ratio corresponding to the spectrum efficiency is

[0118]

[0119] (2) Simulation results

[0120] Figure 2The influence of different access modes on the integrated performance of multi-user communication perception is explained. s =6 OFDMA scheme and K = 6 and N s =3 for the case of no interference suppression. Figure 2 As shown in Figure 2, OFDMA modulation has the best positioning performance, but it has high requirements for frequency resources, especially when there are a large number of users. As the preferred solution that considers both resource and communication perception integration performance, the proposed SCMA solution can achieve 10 at 10dB. -2 The positioning accuracy of the proposed method can reach 0.1 m, and it improves significantly with the increase of transmission power. However, without proper inter-user interference suppression strategy, the positioning accuracy can only reach 0.1 m.

[0121] exist Figure 3 In , we can see that the SCMA scheme achieves the best spectrum efficiency performance under the medium SNR regime because it requires less resources to achieve reliable communication. Even though the OFDMA scheme has the best spectrum efficiency performance under high SNR, the SCMA scheme of the present invention is suitable for serving more users when resources are limited. In addition, it can be observed that ξ k It will also significantly affect the performance of communication perception integration. Therefore, choosing the best ξ k is an important question, which can be addressed by numerical results.

[0122] Figure 4 The positioning performance of different methods for different T at 15dB is shown. active =M semi-RIS scheme as a benchmark. It can be seen that these schemes can achieve centimeter-level positioning, but the requirements for the transmission symbol length are different. Specifically, the pure pilot scheme requires 160 symbols, while the superimposed symbol scheme of the present invention only requires 192 symbols, which is lower than other schemes. This shows that the scheme proposed by the present invention requires fewer symbol resources to achieve communication capabilities, while also achieving high-precision positioning.

[0123] Figure 5 It is proved that the proposed superimposed symbol scheme can achieve the maximum effective throughput in communication, showing the number of effective bits that can be transmitted while ensuring centimeter-level positioning. It can be seen that the strategy of using a pure pilot scheme to achieve positioning performs poorly in terms of effective throughput because the transmission of pilot symbols occupies a large amount of communication resources. Among them, the present invention has the largest effective throughput because the superimposed symbols contain communication bits, so the communication process is uninterrupted.

[0124] It can be seen from the above results that the SCMA transmission scheme of the present invention can efficiently realize the communication perception function of multiple users under the condition of limited spectrum resources. The superimposed pilot scheme of the present invention has obvious improvements in positioning accuracy and spectrum efficiency compared with the existing communication perception integrated method using separate pilots and semi-RIS.

[0125] The above embodiments are used to illustrate the present invention rather than to limit the present invention. Any modification and change made to the present invention within the spirit of the present invention and the protection scope of the claims shall fall within the protection scope of the present invention.

Claims

1. A low-complexity multi-user communication signal recovery and perception positioning estimation method, characterized in that: The method comprises the following steps: S1. The user transmitting end maps the communication data and pilot data through the SCMA codebook to obtain communication symbols and pilot symbols and linearly superimposes them according to the power allocation coefficient to obtain superimposed symbols and transmit them to the RIS end; S2, the base station receiving end receives the signal reflected by the RIS end, and stacks the signals received in multiple time slots to build a receiving matrix; S3, converting the extraction of the environmental angle parameter into an estimation of the angle grid value (θ, φ) and the sparse vector ψ; obtaining an initial estimation value of the equivalent channel between the base station and the user through the estimation value of the sparse vector; S4, according to the initial estimated value of the equivalent channel, subtract the received part corresponding to the sensing pilot from the received matrix, and obtain the initial estimated value of the communication data using the linear minimum mean square error method and the SCAM decoder; S5. Iterative calculation of the initial estimate of the communication data: The estimated value of the angle grid (θ (j) ,φ (j) ) and the sparse vector estimate ψ (j) , reduce the dimension of the sensing matrix and the sparse vector, obtain the corresponding channel matrix, and subtract the received part corresponding to the sensing pilot from the received matrix Y to determine the estimated value of the communication data; After the iterations are completed, the final estimates of the sparse vectors, angle grids, and communication data are obtained; Finally, the communication data, angle parameters in the environmental channel and user coordinates are recovered.

2. A low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 1, characterized in that: The power distribution coefficient ξ k is the power coefficient of the communication symbol allocated by the kth user to the entire symbol; satisfying 0<ξ k <1.

3. The low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 1, characterized in that: The method of converting the extraction of the environmental angle parameters into the estimation of the angle grid value (θ, φ) and the sparse vector ψ is specifically as follows: the millimeter wave channel between the RIS and each user is converted into the form of a dictionary matrix Ω(θ, φ) multiplied by a sparse vector ψ, wherein the dictionary matrix is ​​composed of array responses corresponding to (θ, φ) after the azimuth and elevation angles are uniformly quantized in the angle domain.

4. The low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 1, characterized in that: The method of obtaining the initial estimated value of the equivalent channel between the base station and the user through the estimated value of the sparse vector is as follows: according to the known information of the pilot symbol, the initial estimated value of the sparse vector ψ is obtained by using the least squares estimation. (0) , where the superscript (j) represents the result obtained at the jth iteration; According to the obtained ψ (0) , the equivalent channel initial estimation value H between the base station and the user is obtained through affine transformation (0) ; According to H (0) , subtract the received part corresponding to the sensing pilot from the received matrix Y, and use the linear minimum mean square error method and SCAM decoder to obtain the initial estimate of the communication data Will After demodulation, re-modulation is performed to obtain the processed 5. The low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 4, characterized in that: The SCAM decoder uses a message passing algorithm to update the probability corresponding to each codeword of each user and c ∈{1,...,N c The codeword with the maximum a posteriori probability is selected from the codewords as the decoded data; in the process of calculating the initial estimate of the equivalent channel, all codewords are set to the same prior probability.

6. The low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 1, characterized in that: The iterative calculation is specifically as follows: Based on the estimated data value in the previous iteration Recover the superimposed symbol estimate X (j-1) , the estimated value of the grid (θ) is obtained by sparse Bayesian learning method of unitary approximate message passing off the grid (j) ,φ (j) ) and the sparse vector estimate ψ (j) , where j = 1 during the initial value estimation process; Select the obtained sparse vector estimate ψ (j) Greater than ∈ ψ ||ψ (j) || the number of units, where ∈ ψ is to set the threshold and discard the corresponding grid value (θ (j) ,φ (j) ), reducing the dimensions of the perception matrix and sparse vectors for updating grid values ​​and sparse vectors with low complexity; After reducing the dimensions of the sensing matrix and sparse vector, the corresponding channel matrix H(j) is obtained. The received part corresponding to the sensing pilot is subtracted from the received matrix Y to calculate the estimated value of the communication data. After demodulation, it is re-modulated to obtain the processed Repeat the above steps until the number of iterations j reaches the preset maximum value j max or the log-likelihood function reaches stability; Through the final (θ (J) ,φ (J) ) and ψ (J) , where J is the final number of iterations, the angle parameters in the communication data and the environmental channel can be restored at the receiving end.

7. A low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 6, characterized in that: The estimated value of the grid (θ) is obtained by the off-grid unitary approximate message passing sparse Bayesian learning method (j) ,φ (j) ) and the sparse vector estimate ψ (j) Specifically: For the following received signal y U =Dψ+n U To recover the sparse vector ψ, where D is the perception matrix, n u For noise, The process includes: S501: Perform singular value decomposition on the perception matrix D to obtain Multiply the two ends of the received signal by the matrix get in S502: Set sparse vector Each element The prior distribution of where γ ψ，g Satisfy the parameter a ψ and b ψ The gamma distribution, p(γ ψ，g |a ψ , b ψ )=Γ(γ ψ，g ; 1+a ψ , b ψ );a ψ and b ψ is an adjustable parameter, and the noise parameter is modeled as β ψ , after introducing the hard constraint factor k = Λψ, we have p(κ q |ψ)=δ(κ q -Λ(q,:)ψ); S503: Joint distribution based on conditions The factor graph of the message passing method is used to iteratively update and finally obtain the sparse vector ψ; Repeat the process below: renew renew renew renew renew renew Update s = τ s (rp); update renew renew renew renew renew Update u←u+1; iterate the update process until or u>u max Finally, we get the sparse vector ψ (1) ; in and They are vectors and The gth element of ; G1, G2, G3 and G4 are the number of quantized copies in the angle domain; the above τ p ,p,v k , τ s ,s, q, and a ∈ are intermediate variables in the update process, u is the number of iterations, δ ψ and u max The convergence threshold and maximum number of iterations are set respectively; M is the number of base station antennas, and T is the number of stacked time slots.

8. A low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 7, characterized in that: The update method after reducing the dimension of the operation matrix / vector in the signal processing at the receiving end solves the j-th iteration X in (j) , and The recovery process includes the following: Recover data using the SCMA decoder Using low-complexity message passing methods, we can recover Use subgradient descent to update Repeat the process until two iterations The difference converges.

9. The low-complexity multi-user communication signal recovery and perception positioning estimation method according to claim 1, characterized in that: The position coordinates of each user are estimated and determined by estimating and restoring the azimuth and elevation angles from one RIS to the user. Specifically, the position coordinates and relative distance of the user's partial axis are determined by the formula of the arrival angle and coordinates between the RIS and the user, and then the full position coordinates of the user are obtained based on the relationship between the base station and the user.

10. A system for implementing the low-complexity multi-user communication signal recovery and perception positioning estimation method according to any one of claims 1 to 9, characterized in that: Includes the following modules: Transmit signal module: used to map the communication data and pilot data through the SCMA codebook, obtain the communication symbols and pilot symbols, and linearly superimpose them according to the power allocation coefficient to obtain the superimposed symbols and transmit them to the RIS end; Receiving signal module: used to receive the signal reflected by the RIS end, and stack the signals received in multiple time slots to build a receiving matrix; Initialization module: used to obtain the initial estimation value of communication data using the linear minimum mean square error method and SCAM decoder according to the information of the pilot symbol; Parameter estimation module: configured to obtain a grid estimate and a sparse vector estimate through a unitary approximate message passing sparse Bayesian learning method off the grid based on the initial estimate obtained by the initialization module; A user positioning module, used to obtain the position coordinates of all users according to the angle domain grid values ​​and the estimated values ​​of the sparse vector obtained by the parameter estimation module; Data detection module: used to calculate and restore communication data values ​​based on the estimated values ​​obtained by the parameter estimation module, and includes modulation and demodulation functions.

Images