ISAC method based on near-field semi-passive RIS assistance
By establishing a semi-passive RIS-assisted ISAC system model under the near-field spherical wave channel, and using the semi-determinal relaxation optimization method to alternately optimize the beamforming matrix, the problem of limited near-field communication and perceptual performance is solved, and the near-field communication is achieved Perceptual performance improvement and channel performance optimization in field ISAC systems.
Patent Information
- Application Number
- CN202410705141.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-03
- Publication Date
- 2024-11-22
AI Technical Summary
Existing research has failed to achieve the performance improvement of semi-passive RIS-assisted ISAC system under near-field spherical wave channels. The traditional channel modeling method is not applicable, resulting in limited near-field communication and perceptual performance.
建立近场半无源RIS辅助的ISAC系统模型,通过半定松弛优化方法交替迭代优化主动和被动波束赋形矩阵,获取费舍尔信息矩阵,考虑角度先验信息进行局部最优解求解,降低雷达信道系数、距离和角度参数的克拉美罗界。
Under the user communication conditions, the perceived performance of the near-field ISAC system is improved, the signal-to-noise ratio and perception accuracy are improved, and the channel performance between the semi-passive RIS and the target is enhanced.
Smart Images

Figure CN119010971A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of near field communication and communication perception integration, and in particular relates to an ISAC method based on near field semi-passive RIS assistance. Background Art
[0002] With the development of wireless mobile communication technology, the number of wireless mobile terminal devices has also experienced explosive growth, resulting in a greater demand for mobile services, which has brought huge challenges to future 6G mobile communications. This has also promoted the continuous development and progress of various communication technologies, such as Integrated Sensing and Communication (ISAC) technology.
[0003] As one of the key technologies of 6G, ISAC technology has received extensive attention from academia and industry. ISAC has better performance than existing frequency division communication and sensing technologies, but the improvement brought by a single technology is ultimately limited, so Reconfigurable Intelligent Surface (RIS) has entered the field of vision of researchers as a complementary technology.
[0004] The combination of RIS and ISAC has greatly improved the communication system. However, existing research work is based on far-field communication. In traditional communication scenarios, the antenna array is small in size, resulting in a small range of near-field communication. However, with the development of RIS and MIMO, the array aperture and wavelength are constantly decreasing, resulting in a rapid increase in the near-field area of the BS, increasing the possibility of sensing and communication in the near field.
[0005] For near-field environments, the plane wave assumption in the far field is invalid, and the propagation form of electromagnetic waves (EM) in near-field channels is more like spherical waves. This makes traditional channel modeling methods no longer applicable, which poses a challenge to existing work. Due to the uniqueness of the electromagnetic wave propagation model in the near field, the near-field channel contains both the user's angle and distance information, so both perception and communication will have greatly improved performance in near-field situations. In particular, the rise of near-field communications has also brought more opportunities for the development of ISAC.
[0006] Compared with passive RIS, using semi-passive RIS to assist ISAC system can achieve more performance improvement. Through the advantage of semi-passive RIS in reducing path loss, the accuracy of sensing results can be greatly improved. However, how to achieve the performance improvement of ISAC system assisted by semi-passive RIS under the channel assumption of near-field spherical wave is a part that has not been studied in the existing technology. Summary of the invention
[0007] The purpose of the present invention is to propose an ISAC method based on near-field semi-passive RIS assistance, so as to improve the perception performance in the near-field ISAC system under the premise of ensuring the communication between the base station and the user.
[0008] The present invention is achieved through the following technical solutions:
[0009] An ISAC method based on near-field semi-passive RIS assistance comprises the following steps:
[0010] Step S1, establishing a semi-passive RIS-assisted ISAC system model in the near field, in which the base station serves multiple users through the semi-passive RIS assistance, and obtaining the covariance matrix of active beamforming according to the system model. When the prior information of the angle is not considered, the Fisher information matrix of the target location is obtained by using the echo signal received by the base station, and the process proceeds to step S2. When the prior information of the angle is considered, the Fisher information matrix of the target location with the prior information is obtained by using the prior information and the echo signal received by the base station, and the process proceeds to step S3.
[0011] Step S2, constructing a passive beamforming matrix that satisfies the constant modulus constraint, and performing alternating iterative optimization on the active beamforming matrix and the passive beamforming matrix based on the system model and the Fisher information matrix and the semidefinite relaxation optimization method to obtain local optimal solutions of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound of the radar channel coefficient, the distance parameter and the angle parameter between the semi-passive RIS and the target;
[0012] Step S3, constructing a passive beamforming matrix that satisfies the constant modulus constraint, and performing alternating iterative optimization on the active beamforming matrix and the passive beamforming matrix based on the system model and the Fisher information matrix with prior information, so as to obtain local optimal solutions of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound of the radar channel coefficient, the distance parameter and the angle parameter between the semi-passive RIS and the target.
[0013] Furthermore, in step S1, the base station of the ISAC system model has N BS antennas, each user is a single-antenna user, and the semi-passive RIS is based on L = L y L z The elements on the rectangular plane form L y Line and L z Column, N elements in L elements are reflection units, M elements are perception units, and the covariance matrix of the active beamforming is in, is the transmission signal of the base station, w k(t) represents the transmission beamforming vector of the information flow transmitted by user k at time t, represents the transmission beamforming vector for user k to transmit information stream, s(t) represents the dedicated sensing signal for achieving full freedom of target sensing, c k (t) represents the communication signal, which is a Gaussian random signal with zero mean and unit variance.
[0014] Furthermore, in step S1, when the prior information of the angle cannot be obtained, the Fisher information matrix of the target location is obtained using the echo signal received by the base station, which specifically includes the following steps:
[0015] Step S11: According to the system model, the signal stream received by the kth user is obtained as follows: in, represents the near-field channel vector between the semi-passive RIS and the kth user, represents the phase of the reflection unit, H represents the far-field Rice channel between the base station and the semi-passive RIS, and n k (t) represents the additive Gaussian white noise of the kth user that follows the complex Gaussian distribution;
[0016] Step S12: According to the signal stream received at the kth user, the transmission rate at the kth user is obtained as follows: in, represents the variance of the noise;
[0017] Step S13: According to the transmission signal of the base station, the echo signal received by the base station is represented as y s (t) = Gθ N Hx k (t)+n s (t), where represents the radar echo channel between the semi-passive RIS and the target, β s represents the radar channel coefficient between the semi-passive RIS and the target, a sm (r s ,θ s ) represents the perceived near-field array response vector, a sn (r s ,θ s ) represents the reflected near-field array response vector, r s represents the distance from the semi-passive RIS to the target, θ s represents the angle of the target relative to the semi-passive RIS, n s (t) has a mean of 0 M , with variance σ 2 I M The complex Gaussian distribution of .
[0018] Step S14: according to the echo signal, set the unknown quantity including the radar channel coefficient, the distance parameter and the angle parameter as Then the Fisher information matrix of the target location is expressed as in, Represents β s The real part of Represents β s The imaginary part of
[0019]
[0020] q∈{r s ,θ s}, and Indicates a sm (r s ,θ s ) and a sn (r s ,θ s ) About r s The derivative of and Indicates a sm (r s ,θ s ) and a sn (r s ,θ s ) about θ s The derivative of , T represents the number of samples, Represents the variance of the noise generated during perception.
[0021] Furthermore, the step S2 specifically includes the following steps:
[0022] Step S21, the active beamforming matrix optimization sub-problem is expressed as Among them, P max represents the transmitter power, γ min represents the lower bound of the rate guaranteeing user communication, F k represents the covariance matrix of the kth user communication beam;
[0023] Step S22: According to the rate expression at the kth user and the Fisher information matrix F, the optimization subproblem in step S1 is re-expressed as Solving this reformulated problem yields the optimal R x ,in, Constraint C 1 Expressed as Tr(R x )≤Pmax , C 2 Expressed as
[0024] Q is the passive beamforming matrix,
[0025] Step S23: Based on the optimal R obtained in step S23 x , the passive beamforming matrix optimization subproblem is expressed as By ignoring the rank-one constraint, the optimization subproblem is transformed into a convex problem and solved;
[0026] Step S24: perform iterative solution to step S22 and step S23 until convergence, and obtain R x , Q is the local optimal solution. For the passive beamforming matrix Q, it does not necessarily satisfy the rank-one constraint, so the singular value decomposition method is used to restore a rank-one solution.
[0027] Furthermore, in step S1, the Fisher information matrix with prior information is expressed as in,
[0028]
[0029] I 2 Represents a unit diagonal matrix of dimension 2
[0030] p(θ s ) is the prior information of the angle The probability density function of , which obeys the Gaussian mixture distribution and is expressed as represents the mean of the j-th Gaussian probability density function, represents the corresponding variance, p j ∈[0,1] represents the weight of the jth Gaussian probability density function and satisfies For the unknown quantity ζ, we can get
[0031] in, Indicates the jth 1 The weights of the Gaussian probability density functions, Indicates the jth 2 The weights of the Gaussian probability density functions, Indicates the jth1 The mean of the Gaussian probability density functions, represents the corresponding variance, Indicates the jth 2 The mean of the Gaussian probability density functions, represents the corresponding variance, J represents the number of Gaussian probability density functions, p k ∈[0,1] represents the weight of the kth Gaussian probability density function and satisfies
[0032] Furthermore, the step S3 specifically includes the following steps:
[0033] Step S31, the active beamforming matrix optimization sub-problem is expressed as Solving this optimization subproblem, we can get the optimal R x , where U -1 is the introduced variable, Q is the passive beamforming matrix, Constraint C 1,p Each element of the matrix can be re-expressed as:
[0034] in,
[0035]
[0036] Step S32: Based on the optimal R obtained in step S31 x , the passive beamforming matrix optimization sub-problem is expressed as By ignoring the rank-one constraint, the problem is transformed into a convex problem form and can be solved;
[0037] Step S33, iterate and solve step S31 and step S32 until convergence, and obtain R x , Q is the local optimal solution. For the passive beamforming matrix Q, it does not necessarily satisfy the rank-one constraint, so the singular value decomposition method is used to restore a rank-one solution.
[0038] The present invention has the following beneficial effects:
[0039] 1. The present invention first establishes a semi-passive RIS-assisted ISAC system model in the near field, obtains a Fisher information matrix based on the model, and performs alternating iterative optimization on the active beamforming matrix and the passive beamforming matrix based on a semidefinite relaxation optimization method to obtain local optimal solutions of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound of the radar channel coefficient, distance parameter and angle parameter between the semi-passive RIS and the target, thereby achieving the base station under the condition of satisfying the service user communication, that is, under the condition of satisfying the user signal-to-noise ratio constraint and the transmitter power constraint, improving the perception performance of the near-field ISAC system.
[0040] 2. Considering the prior information of angle, the present invention obtains the Fisher information matrix with prior information based on the ISAC system model assisted by the semi-passive RIS in the near field, and then obtains the local optimal solution of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound of the radar channel coefficient, distance parameter and angle parameter between the semi-passive RIS and the target. By adding the prior information of angle, the perception performance of the near-field ISAC system is further improved. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] The present invention is further described in detail below with reference to the accompanying drawings.
[0042] Figure 1 The flowchart of the present invention is to optimize active and passive beamforming when the prior information of the angle is not considered.
[0043] Figure 2 This is a diagram of the RMMSE simulation results when the present invention does not consider the prior information of the angle.
[0044] Figure 3 This is a simulation result diagram comparing the present invention with the maximum likelihood estimation when the prior information of the angle is not considered.
[0045] Figure 4 This is a diagram of RMMSE simulation results when the present invention considers prior information of angles.
[0046] Figure 5 This is a simulation result diagram comparing the present invention with the maximum likelihood estimation when the prior information of the angle is considered. DETAILED DESCRIPTION
[0047] The ISAC method based on near-field semi-passive RIS assistance includes the following steps:
[0048] Step S1, establishing a semi-passive RIS-assisted ISAC system model in the near field, in which the base station serves multiple users through the semi-passive RIS assistance, and obtaining the covariance matrix of active beamforming according to the system model. When the prior information of the angle is not considered, the Fisher information matrix of the target location is obtained by using the echo signal received by the base station, and the process proceeds to step S2. When the prior information of the angle is considered, the Fisher information matrix of the target location with the prior information is obtained by using the prior information and the echo signal received by the base station, and the process proceeds to step S3.
[0049] Specifically, the base station of the ISAC system model has N BS = 5 antennas, K = 2 single-antenna users, a sensing target, and a semi-passive RIS with L = 40 elements. The semi-passive RIS is composed of L = L y L z The elements on the rectangular plane form a horizontal L y = 5 rows and vertical L z =8 columns, N=35 elements out of L elements are reflection units, M=5 elements are perception units, L=M+N.
[0050] According to the above ISAC system model, the base station's transmission signal is expressed as Among them, w k (t) represents the transmission beamforming vector of the information flow transmitted by user k at time t, s(t) represents the dedicated sensor signal for achieving full freedom of target sensing, c k (t) represents the communication signal, which is a Gaussian random signal with zero mean and unit variance, such as k = i, k≠i, s(t) represents the dedicated sensor signal for achieving full freedom of target sensing, and its covariance matrix is expressed as Therefore, the covariance matrix of active beamforming (also the covariance matrix of the transmitted signal) is in, represents a transmission beamforming vector for user transmission information stream. In this embodiment, the sampling number T=128.
[0051] Regarding “when the prior information of the angle cannot be obtained, the Fisher information matrix of the target location is obtained by using the echo signal received by the base station”, the specific steps include:
[0052] Step S11: According to the system model, the signal stream received by the kth user is obtained as follows: in, For useful signals, is the inter-user interference, To perceive interference, represents the additive Gaussian white noise at the kth user that follows a complex Gaussian distribution, represents the phase of the reflection unit, n k (t) represents the additive Gaussian white noise of the kth user that follows the complex Gaussian distribution;
[0053] represents the near-field channel vector between the semi-passive RIS and the kth user. The near-field channel between the kth user and the nth reflection unit of the semi-passive RIS can be expressed as [h k ] n =β k exp(-j2πr kn / λ), represents the near-field channel gain, λ represents the wavelength, and r kn =||u k -i n || 2 ,u k =(x k ,y k ,0) represents the position of the kth user in the three-dimensional Cartesian coordinate system, u I =(0,y i ,z i ) represents the position of the semi-passive RIS in the three-dimensional Cartesian coordinate system, i n =(0,i y (n)d R +y i ,[i z (n)+1]d R +z i ) represents the position of each element used for communication, and i y (n) = mod (n-1, L y )and are the row and column indices respectively, M z =M / N y , d R =λ is the interval between reflection elements. By using the user's angle information and distance information, r kn Represented as r kn (θ k ,r k )=||[r k cosθ,r k sinθ,0]-i n || 2 , where θ = arctan((y k -y i ) / x k ), u k =(x k ,yk ,0) represents the location information of the kth user. H represents the far-field Rice channel between the base station and the semi-passive RIS, which can be modeled as Among them, β BS is the channel gain at a reference distance of 1 m, α is the path loss exponent, K h is the Rice factor, Obeys a complex Gaussian distribution (0,1),
[0054] and represent the azimuth arrival angle and elevation arrival angle of the semi-passive RIS, θ j =π-θ i is the departure angle at the base station, and They represent the coordinates of the semi-passive RIS and the base station in the xoz plane respectively.
[0055] In this implementation, β BS =-30dB, α = 2.2, K h =3, f c =10GHz,λ=0.03m;
[0056] Step S12: According to the signal stream received at the kth user, the transmission rate at the kth user is obtained as follows: in, represents the transmission beamforming vector for user k to transmit information stream, represents the variance of the noise at the kth user;
[0057] Step S13: According to the transmission signal of the base station, the echo signal received by the base station is represented as y s (t) = Gθ N Hx k (t)+n s (t), where represents the radar echo channel between the semi-passive RIS and the target, β s represents the radar channel coefficient between the semi-passive RIS and the target (specifically, it represents the signal attenuation caused by the transmission from the semi-passive RIS to the target and then from the target to the RIS sensor), u s =(x s ,y s ,0) represents the position of the perceived target, κ represents the energy attenuation caused by the radar cross section, represents the distance from the semi-passive RIS to the target, θ s =arctan((y s -y i ) / xs ) represents the angle of the target relative to the semi-passive RIS, a sm (r s ,θ s ) represents the perceived near-field array response vector, a sn (r s ,θ s ) represents the reflected near-field array response vector, which are respectively expressed as Among them, a sm =exp(-j2πr sm / λ), a sn =exp(-j2πr sn / λ), r sm (θ s ,r s )=||[r s cosθ,r s sinθ,0]-i m || 2 , r sn (θ s ,r s )=||[r s cosθ,r s sinθ,0]-i n || 2 ,i n =(0,i y (m)d R +y i ,[i z (m)+1]d R +z i ) represents the position of each element used for perception, i y (n) = mod(m-1, L y )and are the row and column indices, respectively, x k (t) represents the transmission signal of the base station, n s (t) has a mean of 0 M , the variance is The complex Gaussian distribution, I M represents the unit diagonal matrix of dimension M;
[0058] Step S14: according to the echo signal, set the unknown quantity including the radar channel coefficient, the distance parameter and the angle parameter as Then the Fisher information matrix of the target location is expressed as in, Represents β s The real part of Represents β sFor the imaginary part of the matrix F, each element can be expressed as:
[0059]
[0060] The symbols in the above formula represent the following meanings:
[0061] q∈{r s ,θ s}, and Indicates a sm (r s ,θ s ) and a sn (r s ,θ s ) About r s The derivative of and Indicates a sm (r s ,θ s ) and a sn (r s ,θ s ) about θ s The derivatives of these four items can be expressed as,
[0062] The symbols in these four formulas represent the following meanings:
[0063] Then the CRB of ζ can be obtained and expressed as Through some mathematical transformations, this CRB can be further expressed as in,
[0064] According to the formula CRB can be further simplified as in,
[0065] T represents the number of samples, Represents the variance of the noise generated during perception.
[0066] Step S2, constructing a passive beamforming matrix that satisfies the constant modulus constraint, and performing alternating iterative optimization on the active beamforming matrix and the passive beamforming matrix based on the system model and the Fisher information matrix and the semidefinite relaxation optimization method to obtain local optimal solutions of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound CRB(ζ) of the radar channel coefficient, distance parameter and angle parameter between the semi-passive RIS and the target;
[0067] The specific steps include:
[0068] Step S21: Under the conditions of satisfying SINR constraint and maximum transmit power constraint, the active beamforming matrix optimization subproblem can be expressed as: Among them, P max represents the transmitter power, γ min represents the lower bound of the rate guaranteeing user communication, F k represents the covariance matrix of the kth user communication beam;
[0069] Step S22: According to the rate expression at the kth user and the Fisher information matrix F, the optimization subproblem in step S1 is re-expressed as Solving this reformulated problem yields the optimal R x ,in, Constraint C 1 Expressed as Tr(R x )≤P max , C 2 Expressed as in, Q is the passive beamforming matrix,
[0070] Step S23: Based on the optimal R obtained in step S23 x , the passive beamforming matrix optimization subproblem is expressed as By ignoring the rank-one constraint, the optimization subproblem is transformed into a convex problem and solved;
[0071] Step S24: perform iterative looping on steps S22 and S23 until the target value converges. x , Q is the local optimal solution. For the passive beamforming matrix Q, it does not necessarily satisfy the rank-one constraint, so the singular value decomposition method is used to restore a rank-one solution. The above optimization process is as follows Figure 1 shown.
[0072] exist Figure 2In the figure, the horizontal axis is the number of iterations, and the vertical axis is the RMMSE. It can be clearly seen that the present invention not only meets the communication requirements but also increases the perception accuracy.
[0073] Figure 3 In the figure, the horizontal axis is the radar SNR, which is [-30:10:30], and the vertical axis is the RMMSE. It can be seen from the figure that the angle CRB can be well estimated, but there is a gap between the distance CRB and the estimated CRB. This is because the distance between the target and the semi-passive RIS is in the near field and the distance is very close, so the CRB distance can reach a lower theoretical value, and the maximum likelihood estimation method reaches the limit of its estimation accuracy, so this phenomenon occurs.
[0074] Step S3, constructing a passive beamforming matrix that satisfies the constant modulus constraint, and performing alternating iterative optimization on the active beamforming matrix and the passive beamforming matrix based on the system model and the Fisher information matrix with prior information, so as to obtain local optimal solutions of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound of the radar channel coefficient, the distance parameter and the angle parameter between the semi-passive RIS and the target;
[0075] Assume that the probability density function (PDF) of the prior information of the angle obeys a Gaussian mixture distribution and is expressed as follows in, represents the mean of the j-th Gaussian probability density function, represents the variance of the j-th Gaussian probability density function, p j ∈[0,1] represents the weight of the jth Gaussian probability density function and satisfies J is an integer, 1≤j≤J;
[0076] For the unknown quantity ζ, its joint distribution is Among them, f(Y s |ζ) represents the conditional PDF for a given ζ, Represents the marginal distribution of ζ, so the Fisher information matrix with prior information can be obtained and expressed as follows in, ln(f(Y s ζ)) can be expressed as So F Y can be expressed as in, Each element in the above matrix can be expressed as
[0077]
[0078] I 2 represents a unit diagonal matrix with a dimension of 2. The symbols in the above formula represent the following meanings:
[0079] In addition, for the unknown quantity ζ, we can get F p for in, Indicates the jth 1 The weights of the Gaussian probability density functions, Indicates the jth 2 The weights of the Gaussian probability density functions, Indicates the jth 1 The mean of the Gaussian probability density functions, represents the corresponding variance, Indicates the jth 2 The mean of the Gaussian probability density functions, Represents the corresponding variance, J represents the number of Gaussian probability density functions, then the Fisher information matrix containing prior information can be further expressed as in, p k ∈[0,1] represents the weight of the kth Gaussian probability density function and satisfies
[0080] Then the CRB of the unknown quantity ζ can be expressed as
[0081] Then step S3 specifically includes the following steps:
[0082] Step S31: When the SINR constraint and the maximum transmit power constraint of the user are satisfied, the active beamforming matrix optimization subproblem is expressed as: Solving this optimization subproblem, we can get the optimal R x , where U -1 represents the introduced intermediate variable, Q is the passive beamforming matrix, Constraint C 1,p Each element of the matrix can be re-expressed as: in,
[0083] Step S32: Based on the optimal R obtained in step S31 x, the passive beamforming matrix optimization sub-problem is expressed as By ignoring the rank-one constraint, the problem is transformed into a convex problem form and can be solved;
[0084] Step S33, iterate and solve step S31 and step S32 until the target value converges, and the obtained R x , Q is the local optimal solution. For the passive beamforming matrix Q, it does not necessarily satisfy the rank-one constraint, so the singular value decomposition method is used to restore a rank-one solution.
[0085] exist Figure 4 In the figure, the horizontal axis is the number of iterations and the vertical axis is the RMMSE. It can be clearly seen that the present invention not only meets the communication requirements but also increases the perception accuracy. At the same time, the reduction of CRB for angles also proves the effectiveness of adding angle prior information.
[0086] exist Figure 5 In the figure, the horizontal axis is the radar signal-to-noise ratio SNR, which is [-20:10:30], and the vertical axis is RMMSE. The reduction of CRB with respect to angles indicates that the additional prior information improves the positioning performance with respect to angles.
[0087] The above description is only a preferred embodiment of the present invention, and therefore cannot be used to limit the scope of implementation of the present invention. That is, equivalent changes and modifications made according to the scope of the patent application of the present invention and the contents of the specification should still fall within the scope covered by the patent of the present invention.
Claims
1. An ISAC method based on near-field semi-passive RIS assistance, characterized in that: The steps include: Step S1, establishing a semi-passive RIS-assisted ISAC system model in the near field, in which the base station serves multiple users through the semi-passive RIS assistance, and obtaining the covariance matrix of active beamforming according to the system model. When the prior information of the angle is not considered, the Fisher information matrix of the target location is obtained by using the echo signal received by the base station, and the process proceeds to step S2. When the prior information of the angle is considered, the Fisher information matrix of the target location with the prior information is obtained by using the prior information and the echo signal received by the base station, and the process proceeds to step S3. Step S2, constructing a passive beamforming matrix that satisfies the constant modulus constraint, and performing alternating iterative optimization on the active beamforming matrix and the passive beamforming matrix based on the system model and the Fisher information matrix and the semidefinite relaxation optimization method to obtain local optimal solutions of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound of the radar channel coefficient, the distance parameter and the angle parameter between the semi-passive RIS and the target; Step S3, constructing a passive beamforming matrix that satisfies the constant modulus constraint, and performing alternating iterative optimization on the active beamforming matrix and the passive beamforming matrix based on the system model and the Fisher information matrix with prior information, so as to obtain local optimal solutions of the active beamforming matrix and the passive beamforming matrix, thereby minimizing the Cramer-Rao bound of the radar channel coefficient, the distance parameter and the angle parameter between the semi-passive RIS and the target.
2. The ISAC method based on near-field semi-passive RIS assistance according to claim 1, characterized in that: In step S1, the base station of the ISAC system model has N BS antennas, each user is a single-antenna user, and the semi-passive RIS is based on L = L y L z The elements on the rectangular plane form L y Line and L z Column, N elements in L elements are reflection units, M elements are perception units, and the covariance matrix of the active beamforming is in, is the transmission signal of the base station, w k (t) represents the transmission beamforming vector of the information flow transmitted by user k at time t, represents the transmission beamforming vector for user k to transmit information stream, s(t) represents the dedicated sensing signal for achieving full freedom of target sensing, c k (t) represents the communication signal, which is a Gaussian random signal with zero mean and unit variance.
3. The ISAC method based on near-field semi-passive RIS assistance according to claim 2, characterized in that: In step S1, when the prior information of the angle cannot be obtained, the Fisher information matrix of the target location is obtained by using the echo signal received by the base station, which specifically includes the following steps: Step S11: According to the system model, the signal stream received by the kth user is obtained as follows: in, represents the near-field channel vector between the semi-passive RIS and the kth user, represents the phase of the reflection unit, H represents the far-field Rice channel between the base station and the semi-passive RIS, and n k (t) represents the additive Gaussian white noise of the kth user that follows the complex Gaussian distribution; Step S12: According to the signal stream received at the kth user, the transmission rate at the kth user is obtained as follows: in, represents the variance of the noise; Step S13: According to the transmission signal of the base station, the echo signal received by the base station is represented as y s (t) = Gθ N Hx k (t)+n s (t), where represents the radar echo channel between the semi-passive RIS and the target, β s represents the radar channel coefficient between the semi-passive RIS and the target, asm(rs,θs) represents the perceived near-field array response vector, asn(rs,θs) represents the reflected near-field array response vector, and r s represents the distance from the semi-passive RIS to the target, θ s represents the angle of the target relative to the semi-passive RIS, n s (t) has a mean of 0 M , with variance σ 2 I M The complex Gaussian distribution of . Step S14: according to the echo signal, set the unknown quantity including the radar channel coefficient, the distance parameter and the angle parameter as Then the Fisher information matrix of the target location is expressed as in, Represents β s The real part of Represents β s The imaginary part of q∈{r s ,θ s }, and Indicates a sm (r s ,θ s ) and a sn (r s ,θ s ) with respect to rs, and Indicates a sm (r s ,θ s ) and a sn (r s ,θ s ) about θ s The derivative of , T represents the number of samples, Represents the variance of the noise generated during perception.
4. The ISAC method based on near-field semi-passive RIS assistance according to claim 3 is characterized in that: The step S2 specifically includes the following steps: Step S21, the active beamforming matrix optimization sub-problem is expressed as subject to Tr(R x )≤P max , where Pmax represents the transmitter power and γmin represents Guaranteed lower bound on user communication rate, F k represents the covariance matrix of the kth user communication beam; Step S22: According to the rate expression at the kth user and the Fisher information matrix F, the optimization subproblem in step S1 is re-expressed as Solving this reformulated problem yields the optimal R x ,in, Constraint C1 is expressed as Tr(R x )≤P max , C2 is expressed as Q is the passive beamforming matrix, Step S23: Based on the optimal R obtained in step S23 x , passive beamforming matrix optimization subproblem Expressed as st[Q] nn =1,Q±0,, by ignoring the rank-one constraint, the optimal Transform the subproblem into a convex problem and solve it; Step S24: perform iterative solution to step S22 and step S23 until convergence, and obtain R x , Q is the local optimal solution. For the passive beamforming matrix Q, it does not necessarily satisfy the rank-one constraint, so the singular value decomposition method is used to restore a rank-one solution.
5. The ISAC method based on near-field semi-passive RIS assistance according to claim 2, characterized in that: In step S1, the Fisher information matrix with prior information is expressed as in, I2 represents the unit diagonal matrix with dimension 2 p(θ s ) is the prior information of the angle The probability density function of , which obeys the Gaussian mixture distribution and is expressed as represents the mean of the j-th Gaussian probability density function, represents the corresponding variance, p j ∈[0,1] represents the weight of the jth Gaussian probability density function and satisfies For the unknown quantity ζ, we can get F p for Other elements, Among them, p j1 represents the weight of the j1th Gaussian probability density function, represents the weight of the j2th Gaussian probability density function, represents the mean of the j1th Gaussian probability density function, represents the corresponding variance, represents the mean of the j2th Gaussian probability density function, represents the corresponding variance, J represents the number of Gaussian probability density functions, p k ∈[0,1] represents the weight of the k-th Gaussian probability density function and is full 6. The ISAC method based on near-field semi-passive RIS assistance according to claim 5, characterized in that: The step S3 specifically includes the following steps: Step S31, the active beamforming matrix optimization sub-problem is expressed as Solving this optimization subproblem, we can get the optimal R x , where U -1 is the introduced variable, Q is the passive beamforming matrix, Constraint C 1,p Each element of the matrix can be re-expressed as: in, Step S32: Based on the optimal R obtained in step S31 x , the passive beamforming matrix optimization sub-problem is expressed as By ignoring the rank-one constraint, the problem is transformed into a convex problem form and can be solved; Step S33, iterate and solve step S31 and step S32 until convergence, and obtain R x , Q is the local optimal solution. For the passive beamforming matrix Q, it does not necessarily satisfy the rank-one constraint, so the singular value decomposition method is used to restore a rank-one solution.
Citation Information
Cited By
Robust safety beamforming and target tracking method and equipment of full-duplex sensing integrated system
CN121150763A
