A channel estimation and user positioning method based on uniform circular array intelligent metasurface

Abstract

The present invention relates to a channel estimation and user positioning method based on a uniform circular array intelligent metasurface. Specifically, it includes: arranging reconfigurable reflection units uniformly according to a circular topological structure, designing a uniform circular array intelligent metasurface RIS; constructing a beam training signal model based on standard multivariate decomposition according to the multipath propagation characteristics of the cascade channel, and restoring the path angle parameters on the base station side through a tensor signal processing algorithm; designing a beamforming method for RIS training reflection coefficients based on the phase mode excitation principle, and restoring the path angle parameters on the RIS side through a real-valued spatial spectrum estimation algorithm; combining broadband multi-carrier signals and free space propagation characteristics, restoring the path delay parameters on the user side, and realizing user positioning in combination with direction and distance information. The present invention can realize unambiguous parameter decoupling and channel estimation that traditional linear RIS cannot support with lower hardware cost and training overhead, thereby realizing high-precision user positioning applications.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication and sensing technology, and in particular to a channel estimation and user positioning method based on a uniform circular array intelligent super surface. Background Art

[0002] The connotation of the next generation mobile communication network (6G) will extend far beyond the scope of communication to the fields of perception and computing. The deep integration of mobile communication technology, information technology, big data and artificial intelligence will drive the expansion of 6G technical elements from information transmission to information collection and computing. New Internet and vertical industry businesses with information perception as the core are also gradually emerging. Facing the needs of social and economic informatization, communication perception convergence (ISAC) will surely become the dominant trend of 6G technology and business.

[0003] The demand for spectrum and antennas in ISAC systems drives the expansion of wireless air interface technologies such as millimeter wave communications and massive multiple-input multiple-output (MIMO) to enhanced technologies such as terahertz communications and ultra-large-scale MIMO. However, the traditional evolution route of "trading scale for efficiency" by using higher frequency bands and deploying more antennas faces severe challenges of "three highs" in terms of cost, energy consumption and complexity. In order to develop innovative, efficient and resource-friendly enabling solutions, the industry is seeking positive changes in information interaction modes. Relying on its unique advantages of low overhead, programmability and easy deployment, intelligent metasurface (RIS) technology has become one of the most promising key technologies for the construction of 6G interawareness integration.

[0004] The physical essence of RIS is a reconfigurable metasurface array composed of a large number of sub-wavelength electromagnetic units. Through digital coding, electromagnetic waves can be dynamically controlled to form an electromagnetic field with controllable amplitude, phase, polarization and frequency. Different from traditional wireless systems, the RIS-ISAC system based on intelligent metasurface focuses on RIS deployment and control strategies. By enriching scattering channels, superimposing signals in phase and controlling beam directions, it can achieve goals such as improving multiplexing gain, combating multipath fading and enhancing system coverage. RIS-ISAC technology is expected to break through the inherent shackles of uncontrollable traditional wireless channels, build a programmable intelligent wireless environment, and introduce a new paradigm of future 6G synaesthesia fusion.

[0005] In order to fully realize the potential of RIS to optimize system environment and enhance interaceptive applications, it is crucial to obtain accurate channel state information (CSI). However, RIS usually does not have the ability to receive and process signals. The pilot length required by traditional algorithms such as least squares and least mean square error is proportional to the number of RIS units, which adds a lot of training overhead to large-scale RIS-assisted wireless systems. In addition, the conventional RIS linear array topology will lead to the loss of channel information in the physical space dimension, which introduces inherent channel estimation ambiguity to RIS-assisted wireless systems, thereby hindering the realization of wireless perception applications such as user positioning and environmental mapping. Summary of the invention

[0006] Technical problem: The purpose of this invention is to provide a channel estimation and user positioning method based on a uniform circular array intelligent metasurface, which integrates RIS-assisted communication and perception into a system, greatly expanding the application scenarios of RIS in wireless systems. It can achieve unambiguous parameter decoupling and channel estimation that traditional linear RIS cannot support with lower hardware cost and training overhead, thereby realizing high-precision user positioning applications.

[0007] Technical solution: A channel estimation and user positioning method based on uniform circular array RIS of the present invention comprises the following steps:

[0008] Step S1, RIS system and channel modeling: design the topology and unit layout of the uniform circular array RIS, and establish the RIS-assisted communication and perception system and multipath channel model;

[0009] Step S2, beam training method design: establish a multi-carrier pilot training signal model, and design a RIS reflection coefficient training pattern assignment method based on manifold equivalent transformation;

[0010] Step S3, channel angle parameter estimation: designing a signal real number processing based on equivalent manifold characteristics, and designing an equivalent angle parameter estimation method based on a spatial spectrum estimation algorithm;

[0011] Step S4, parameter decoupling and user positioning: design an angle, delay and fading decoupling method based on RIS topology and free space propagation characteristics to realize user positioning application.

[0012] The step S1, designing the topological structure and unit layout of the uniform circular array RIS, establishing the RIS-assisted communication and perception system and multipath channel model, specifically includes:

[0013] Consider a uniform circular array RIS, composed of N R Reflection units are arranged around a circle with a radius of r, where the azimuth angle of element n is β n =2π(n-1) / N R , position p n = r[cosβ n ,sinβ n ] T ; Assume that the incident angle of the far-field plane electromagnetic wave is θ and the unit direction is Then the uniform circular array RIS manifold is

[0014]

[0015] Where k0 = 2π / λ is the wave number of the electromagnetic wave with wavelength λ, the subscript R refers to the RIS related parameters, and exp(x) = e xis an exponential function with the natural constant e as the base, (·) T represents transpose, is an imaginary unit;

[0016] Consider a RIS-assisted wireless communication and sensing system, where a single base station and U users are equipped with a single antenna, and the broadband system base frequency is f c , the transmission bandwidth is f s , divided into K subcarriers; the RIS is deployed near the user side, the user-RIS channel is dominated by the line-of-sight path, and the line-of-sight path of the base station-RIS channel is blocked and is composed of non-line-of-sight scattering paths; the uplink channel h from the base station and user u to the RIS on subcarrier k BR,k and h RU,u,k They are modeled as:

[0017]

[0018] Among them, the subscripts B, U, D, and A refer to the base station BS, user User, departure angle AoD, and arrival angle AoA related parameters respectively, and L is h BR,k The number of component paths, f k =f c +kf s / K is the subcarrier k frequency, θ D,l ,τ BR,l ,α BR,l h BR,k The departure angle, delay and fading of the RIS side of the scattering path l, θ A,u ,τ RU,u ,α RU,u h RU,u,k The arrival angle, delay and fading of the RIS side of the line-of-sight path are modeled as α, BR,l ∝(4πf c τ BR,l ) -1 and α RU,u =(4πf c τ RU,u ) -1 , where the proportional symbol ∝ indicates that the former is affected by the unknown loss of the scatterer.

[0019] The step S2, establishing a multi-carrier pilot training signal model, and designing a RIS reflection coefficient training pattern assignment method based on manifold equivalent transformation, specifically includes:

[0020] Channel estimation beam training uses T tr Timeframes and K tr frequency domain subcarriers, configure the reflection coefficient array vector of the RIS array in training frame t User u sends pilot symbol xu , the uplink training signal model is

[0021]

[0022] in, is the equivalent cascade channel, n k,t is the channel Gaussian white noise, ⊙ represents the Khatri-Rao product, Diag(x) represents the matrix with x as the diagonal, and eq refers to the equivalent channel related parameters; according to formula (2), it can be rewritten as

[0023]

[0024] in, represents the Hadamard product, and the pilot symbol is Substituting equation (4) into equation (3), we get

[0025]

[0026] in, is the number of equivalent paths, and the two-dimensional index is mapped to the one-dimensional index is the equivalent attenuation of path p, is the equivalent delay of path p, is the RIS equivalent array manifold corresponding to path p, where θ eq,p ={θ D,l ,θ A,u} contains the angle parameters of the two paths; K tr T on subcarriers tr The received training signals in the training frames are combined into a data matrix

[0027] Y=Ψ T A R,eq G eq +N (6)

[0028] in, is the RIS reflection coefficient sequence in the training cycle, A R,eq =[a R (θ eq,1 ),...,a R (θ eq,P )] collects the RIS array manifold of all equivalent paths, The element at (p, k) is the equivalent complex gain α of path p on subcarrier k eq,p exp(-j2πf k τ eq,p ), They represent the received data matrix and the noise matrix respectively, and their elements at (t, k) are respectively y of the formula t,k ,nt,k ;

[0029] According to the topological structure of the uniform circular array RIS, the phase of the electromagnetic response of the equivalent path p at the RIS array element n is:

[0030]

[0031] in, Returns the phase value of the input complex number.

[0032] The activation function ω(β) with any period of 2π can be expressed in Fourier series as The spatial harmonics ω excited by the array m (β) = exp(jmβ) matching coefficient c m ; Using phase mode m to excite the RIS array aperture, the standardized far-field diagram can be obtained

[0033]

[0034] Among them, J m (ζ) is a type of m-order Bessel function, ζ = 2k0rcosφ, ζ is temporarily omitted for simplicity eq,p ,φ eq,p , Subscript of; From equation (8), the visual domain φ∈[-π,π] is transformed to ζ∈[-2k0r,2k0r]. In order to make the RIS uniform circular array aperture only excite a limited number of modes, let M be the highest order of the excitable modes. According to the empirical rule, the value of M is determined as

[0035] M≈2k0r (9)

[0036] When |m|≥M, The amplitude is small and can be ignored in the entire visual domain. Based on the above phase mode excitation principle, the beamforming weight vector is defined as

[0037]

[0038] Let T tr =2M+1, the training pattern for designing RIS reflection coefficient is

[0039]

[0040] According to formula (8) and the properties of Bessel function J -m (ζ)=(-1) m J m (ζ), the RIS beam space array manifold after the reflection coefficient training pattern Ψ beamforming is equivalent to

[0041]

[0042] in, The angle is The manifold of electromagnetic waves on a uniform linear array, J(ζ)=Diag(J M (ζ),...,J1(ζ),J0(ζ),J1(ζ),...,J M (ζ)).

[0043] The step S3, designing a signal real number processing based on equivalent manifold characteristics, and designing an equivalent angle parameter estimation method based on a spatial spectrum estimation algorithm, specifically includes:

[0044] Formula (12) conforms to the central Hermite structure. In order to enable the beam space manifold signal processing to be performed in the real domain, the orthogonal linear transformation is defined as where γ m =2πi,i∈[-M,M], after being processed by the central Hermite matrix W, the uniform circular array RIS manifold is transformed into real numbers:

[0045]

[0046] in, (·) H represents conjugate transpose;

[0047] Definition C = Diag([(-1) M ,...,(-1) 1 ,1,1,...,1]), according to formula (13), we can get

[0048]

[0049] in, extract Front, middle, back (T tr -2) elements but Contains the same phase value, according to the recursive relation of Bessel function J m-1 (ζ)+J m+1 (ζ)=(2m / ζ)J m (ζ) can be obtained

[0050]

[0051] Where, Γ=2Diag(-(M-1),...,-1,0,1,...,M-1); Formula (15) can be rewritten as in Contains vector Φ=Diag(ω eq,1,...,ω eq,P ),

[0052] The processed signal is recorded as Z = W H Y, calculate the sampling autocorrelation matrix And perform real-valued eigendecomposition to obtain the eigenvectors corresponding to the first P eigenvalues There exists a non-singular matrix Satisfy S=B R T, where Contains real-valued beam space manifolds Defining the selection matrix Extraction of CWS before, during and after (T tr -2) Get the sampling subspace Combination The structure is easy to prove Where D = Diag((-1) M-2 ,...,(-1) 1 ,(-1) 0 ,(-1) 1 ,...,(-1) M ), is a reverse permutation matrix with the anti-diagonal line set to 1 and the rest set to 0, then the subspace matrix form of equation (15) is

[0053]

[0054] Where Ξ=T -1 ΦT, When T tr When ≥2P, equation (16) is an overdetermined system of equations with a unique least squares solution Compute eigendecomposition Get eigenvalues Automatic pairing and recovery

[0055] The step S4, designing an angle, delay and fading decoupling method based on RIS topology and free space propagation characteristics to realize user positioning application, specifically includes:

[0056] According to formula (7), The definition of

[0057]

[0058] Among them, mod 2π (·) represents the modulo operation with 2π as the modulus; Substituting the channel path angle parameter estimate back into the RIS array manifold of formula (12), we can get G eq The least squares solution of Channel estimation training uses continuous Ktr When ≤K adjacent subcarriers, the equivalent delay has a least squares solution

[0059]

[0060] The equivalent path fading can be recovered by

[0061]

[0062] Select the same multipath component l corresponding to the equivalent path According to the free space loss model, we can get The user-RIS channel propagation delay can be recovered by the following formula:

[0063]

[0064] In addition, there are Thus, the uniform circular array RIS-assisted channel estimation method has achieved the unique decoupling of the angle, delay and fading parameters of the channel path from the base station and the user to the RIS.

[0065] Based on the above line-of-sight path parameter information, user positioning application can be directly implemented through the azimuth-distance positioning method.

[0066]

[0067] in, is the prior physical location of RIS, v c is the propagation speed of electromagnetic waves.

[0068] The channel estimation and user positioning method based on uniform circular array RIS of the present invention is summarized as the following algorithm. This algorithm basically adopts algebraic operations and does not involve operations such as random initialization and loop iteration. Therefore, it can obtain high-precision channel estimation results with low computational complexity and operation robustness.

[0069]

[0070] Beneficial effects: Compared with the prior art, the advantages of the present invention are:

[0071] 1) The present invention designs a new topological paradigm of uniform circular array RIS, which supports electromagnetic wave signal reflection in 360° azimuth domain, greatly expanding the 180° azimuth domain supported by linear array RIS.

[0072] 2) The present invention designs a channel estimation method based on uniform circular array RIS, which supports the accurate recovery and unique decoupling of parameters such as two-dimensional spatial path angle, delay and fading. Compared with the linear array RIS that does not support unambiguous angle parameter decoupling, it achieves a qualitative breakthrough.

[0073] 3) The present invention designs a RIS training pattern based on the phase mode excitation principle, which can significantly reduce the training pilot overhead and is applicable to constraints such as the constant envelope of the RIS reflection coefficient and the quantized phase.

[0074] 4) The present invention designs a user positioning method based on deep channel estimation of RIS, which integrates RIS-assisted communication and perception into a system, greatly expanding the application scenarios of RIS in wireless systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 The present invention is a flowchart of a channel estimation and user positioning method based on a uniform circular array (RIS) according to an embodiment of the present invention.

[0076] Figure 2 Schematic diagram of a wireless communication and perception system assisted by a uniform circular array RIS according to an embodiment of the present invention.

[0077] Figure 3 This is a relationship diagram between the channel angle parameter decoupling accuracy and the signal-to-noise ratio based on the uniform circular array RIS according to an embodiment of the present invention.

[0078] Figure 4 This is a relationship diagram between the channel delay parameter decoupling accuracy and the signal-to-noise ratio based on the uniform circular array RIS according to an embodiment of the present invention.

[0079] Figure 5 This is a relationship diagram between the channel fading parameter decoupling accuracy and the signal-to-noise ratio based on the uniform circular array RIS according to an embodiment of the present invention.

[0080] Figure 6 This is a relationship diagram between the user positioning application accuracy and the signal-to-noise ratio based on the uniform circular array RIS according to an embodiment of the present invention. DETAILED DESCRIPTION

[0081] The following will be described in detail with reference to the relevant drawings of the present invention and in combination with the embodiments, so as to more clearly and in detail illustrate the purpose, technical solutions and outstanding advantages of the present invention.

[0082] like Figure 1 As shown, the channel estimation and user positioning method based on uniform circular array RIS provided in the embodiment of the present invention includes the following design steps. It should be emphasized that, under the premise of meeting the operating requirements of the method of the present invention, the specific numerical values ​​of the embodiment do not limit the scope of application of the present invention.

[0083] S1.RIS system and channel modeling

[0084] The embodiment of the present invention considers a uniform circular array RIS, which consists of N R = 256 reflection units, with a radius of r = λ c The circular distribution of c=0.01m is the basic wavelength of the system. The azimuth angle of array element n is β n =2π(n-1) / N R , position p n = r[cosβ n ,sinβ n ] T To avoid spatial aliasing, the spacing Δ between adjacent reflectors in the RIS must satisfy That is, the RIS array size N R and the radius r must satisfy the above constraints. Assume that the incident angle of the far-field plane electromagnetic wave is θ and the unit direction is Then the uniform circular array RIS manifold is

[0085]

[0086] Where k0 = 2π / λ c =200πrad / m is the electromagnetic wave number, the subscript R refers to the RIS related parameters, exp(x)=e x is an exponential function with the natural constant e as base, (·) T represents transpose, Represents an imaginary unit.

[0087] like Figure 2 As shown, the embodiment of the present invention considers a RIS-assisted wireless communication and perception system, in which a single base station serves U = 2 users, the broadband system adopts OFDM mode, and the basic frequency is f c =30GHz, the transmission bandwidth is f s =0.32GHz, divided into K = 128e subcarriers. The base stations and users are randomly distributed in a circle 7.5m to 15m away from the RIS. The line-of-sight path of the base station-RIS channel is blocked, including L = 3 non-line-of-sight paths, and the user-RIS channel is dominated by 1 line-of-sight path. The uplink channel h from the base station and user u to the RIS on subcarrier k is BR,k and h RU,u,k They are modeled as:

[0088]

[0089] Among them, the subscripts B, U, D, and A refer to the base station BS, user User, departure angle AoD, and arrival angle AoA related parameters respectively, and L is h BR,k The number of component paths, f k =f c +kf s / K is the frequency of subcarrier k. θ D,l ,τ BR,l ,α BR,l h BR,kThe departure angle, delay and fading of the RIS side of the scattering path l, θ A,u ,τ RU,u ,α RU,u h RU,u,k Arrival angle, delay and fading on the RIS side of the line-of-sight path; path fading obeys the free space loss model α BR,l ∝(4πf c τ BR,l ) -1 and α RU,u =(4πf c τ RU,u ) -1 , where the proportional symbol ∝ indicates that the former is affected by the unknown loss of the scatterer. According to the device distribution of the embodiment, the path angle {θ D,l ,θ A,u} is randomly generated within [0,2π] radians, and the propagation delay {τ BR,l ,τ RU,u} is randomly generated in [25ns, 50ns] (ns: nanoseconds), and the unknown loss of the scatterer is randomly generated in (0.5, 1.0).

[0090] S2. Beam training method design

[0091] Channel estimation beam training uses T tr = {25,49} time frames and K tr = 128 frequency domain subcarriers. The reflection coefficient vector of the RIS array is configured in the training frame t. User u sends pilot symbol x u , then the uplink training signal model is

[0092]

[0093] in, is the equivalent cascade channel, n k,t is the channel Gaussian white noise, ⊙ represents the Khatri-Rao product, Diag(x) represents the matrix with x as the diagonal, and eq refers to the equivalent channel related parameters. According to formula (2), it can be rewritten as

[0094]

[0095] in, represents the Hadamard product. Let the pilot symbol be Substituting equation (4) into equation (3), we get

[0096]

[0097] in, is the number of equivalent paths, and the two-dimensional index is mapped to the one-dimensional index is the equivalent attenuation of path p, is the equivalent delay of path p. is the RIS equivalent array manifold corresponding to path p, where θ eq,p ={θ D,l ,θ A,u} contains the angle parameters of the two paths. tr T on subcarriers tr The received training signals in the training frames are combined into a data matrix

[0098] Y=Ψ T A R,eq G eq +N (6)

[0099] in, is the RIS reflection coefficient sequence in the training cycle, A R,eq =[a R (θ eq,1 ),...,a R (θ eq,P )] collects the RIS array manifold of all equivalent paths, The element at (p, k) is the equivalent complex gain α of path p on subcarrier k eq,p exp(-j2πf k τ eq,p ), They represent the received data matrix and the noise matrix respectively, and their (t, k)th elements are y of the formula t,k ,n t,k .

[0100] According to the topological structure of the uniform circular array RIS, the phase model of the electromagnetic response of the equivalent path p at the RIS array element n is:

[0101]

[0102] in, Returns the phase value of the input complex number. According to the embodiment equipment distribution, the equivalent angle range is φ eq,p ∈[-π,π],

[0103] The activation function ω(β) with any period of 2π can be expressed in Fourier series as The spatial harmonics ω excited by the array m (β) = exp(jmβ) matching coefficient c m , the normalized far-field diagram is obtained by exciting the RIS array aperture using the phase mode m

[0104]

[0105] Among them, J m (ζ) is a type of m-order Bessel function, ζ = 2k0r cosφ, ζ is temporarily omitted for simplicity eq,p ,φ eq,p , Assume that M is the highest order of modes that can be excited by the aperture of the RIS uniform circular array. The embodiment of the present invention determines that the value of M should satisfy the empirical criteria:

[0106]

[0107] When |m|≥M, The amplitude is small and can be ignored in the entire visual domain. The embodiment of the present invention adopts the mode parameter M={12,24}, corresponding to the channel beam training frame number T tr =2M+1={25,49}. Based on the above phase mode excitation principle, the beamforming weight vector is defined as

[0108]

[0109] The training pattern for designing the RIS reflection coefficient is

[0110]

[0111] According to formula (8) and the Bessel function property J -m (ζ)=(-1) m J m (ζ), the RIS beam space array manifold after the reflection coefficient training pattern Ψ beamforming is equivalent to

[0112]

[0113] in, The angle is The manifold of electromagnetic waves on a uniform linear array, J(ζ)=Diag(J M (ζ),...,J1(ζ),J0(ζ),J1(ζ),...,J M (ζ)).

[0114] S3. Channel angle parameter estimation

[0115] Define orthogonal linear transformation where γ m =2πi,i∈[-M,M], after being processed by the central Hermite matrix W, the uniform circular array RIS manifold is transformed into real numbers:

[0116]

[0117] in, (·) H represents the conjugate transpose.

[0118] Definition C = Diag([(-1) M ,...,(-1) 1 ,1,1,...,1]), according to formula (13), we can get

[0119]

[0120] in, extract Front, middle, back (T tr -2) elements but Contains the same phase value. According to the recursive relation of Bessel function J m-1 (ζ)+J m+1 (ζ)=(2m / ζ)J m (ζ) can be obtained

[0121]

[0122] Where Γ=2Diag(-(M-1),...,-1,0,1,...,M-1). Formula (15) can be rewritten as in Contains vector Φ=Diag(ω eq,1 ,...,ω eq,P ),

[0123] The processed signal is recorded as Z = W H Y, calculate the sampling autocorrelation matrix And perform real-valued eigendecomposition to obtain the eigenvectors corresponding to the first P eigenvalues There exists a non-singular matrix Satisfy S=B R T, where Contains real-valued beam space manifolds Defining the selection matrix Extraction of CWS before, during and after (T tr -2) Get the sampling subspace Combination The structure is easy to prove Where D = Diag((-1) M-2 ,...,(-1) 1 ,(-1) 0 ,(-1) 1 ,...,(-1) M ), is a reverse permutation matrix with the anti-diagonal line set to 1 and the rest set to 0, then the subspace matrix form of equation (15) is

[0124]

[0125] Where Ξ=T -1 ΦT, When T tr When ≥2P, equation (16) is an overdetermined system of equations with a unique least squares solution Compute eigendecomposition Get eigenvalues Automatic pairing and recovery

[0126] S4. Parameter decoupling and user positioning

[0127] According to formula (7), The definition of

[0128]

[0129] Among them, mod 2π (·) represents the modulo operation modulo 2π. Observe that cos in the domain [-π,π] -1 (·) operation will result in {θ D,l ,θ A,u There are two groups of solutions, but only the order of assigning binary variables is exchanged without generating wrong solutions. Physically, it can be explained that the paths from the base station and the user to the RIS are equivalent and symmetrical in the cascade channel. In the embodiment of the present invention, when the number of base station-RIS paths and the number of users are different, that is, L≠U, the 2P group angle estimates are divided into (L+U) group clusters by algorithms such as K-means clustering, where the estimated values ​​that appear U times are regarded as The estimated value repeated L times is regarded as

[0130] Substituting the channel path angle parameter estimate back into the RIS array manifold of equation (12), we can obtain G eq The least squares solution of Channel estimation training uses continuous K tr When ≤K adjacent training subcarriers, the equivalent delay has a least squares solution

[0131]

[0132] The equivalent path fading can be recovered by

[0133]

[0134] Select the same multipath component l corresponding to the equivalent path According to the free space loss model, we can get The user-RIS channel propagation delay can be recovered by the following formula:

[0135]

[0136] In addition, there are The embodiment of the present invention solves the least squares solution by simultaneously solving the LU(U-1) / 2 group of effective constraint equations under the conditions of L>1, U>2, further improving the estimation accuracy. So far, the uniform circular array RIS-assisted channel estimation method has achieved the parameter decoupling of the angle, delay and fading of the base station and user to RIS path.

[0137] Based on the above line-of-sight path parameter information, user positioning application can be directly implemented through the azimuth-distance positioning method.

[0138]

[0139] in, is the a priori RIS physical location, v c is the propagation speed of electromagnetic waves.

[0140] The embodiment of the present invention is based on the above parameter configuration and runs 10 4 The Monte Carlo simulation tests the performance of the channel estimation and user positioning method based on the uniform circular array RIS. The relationship between the decoupling accuracy of the angle, delay and fading parameters obtained by the channel estimation method and the signal-to-noise ratio (SNR) is shown in the figure below: Figure 3 , Figure 4 and Figure 5 As shown in the figure, the relationship between the accuracy of the user positioning method and the change of SNR is as follows Figure 6 As shown, the accuracy index is measured by the root mean square error (RMSE). It can be seen from the simulation results that the channel estimation method provided by the embodiment of the present invention can obtain accurate channel parameter recovery and unique decoupling performance. tr <<N R Under the conditions of 0.1°, it is especially possible to obtain an angle estimation accuracy of the order of 0.1ns and a delay estimation accuracy of the order of 0.1ns, thereby supporting a user positioning performance close to 1.0cm.

[0141] In summary, the channel estimation and user positioning method based on uniform circular array RIS provided by the embodiment of the present invention fully explores the potential of RIS topological structure to "sense" the spatial propagation environment by modeling and analyzing the channel parameter information contained in the RIS array manifold, thereby realizing the unambiguous decoupling of RIS cascade channel path parameters, which is a major problem. In addition, the embodiment of the present invention uses the electromagnetic reconfigurable characteristics of the RIS reflection unit to design a special passive beamforming diagram, which effectively reduces the training overhead while also being applicable to the constraints of RIS power and phase in actual systems.

[0142] In contrast, the traditional linear RIS array manifold is in is the spatial angle determined by the physical angle θ, the array element spacing d and the signal wavelength λ, then the channel equivalent path response between the base station and the user via the RIS cascade is The arrival angle and departure angle Coupling is equivalent to the angle It cannot achieve unique decoupling, nor can it support the recovery of other coupling parameters such as equivalent delay and fading. Although wireless communication applications such as beamforming and capacity optimization rely only on cascaded channels, However, wireless sensing applications such as user positioning and environmental mapping must obtain real and unique propagation path information. The channel estimation and user positioning method based on uniform circular array RIS provided by the embodiment of the present invention breaks through the bottleneck of existing technologies through low-cost, low-energy consumption, programmable and easy-to-deploy RIS technology, and realizes the organic integration of wireless communication and sensing systems.

Claims

1. A channel estimation and user positioning method based on uniform circular array intelligent metasurface, characterized in that: The method comprises the following steps: Step S1, RIS system and channel modeling: design the topology and unit layout of the uniform circular array RIS, and establish the RIS-assisted communication and perception system and multipath channel model; Step S2, beam training method design: establish a multi-carrier pilot training signal model, and design a RIS reflection coefficient training pattern assignment method based on manifold equivalent transformation; Step S3, channel angle parameter estimation: designing a signal real number processing based on equivalent manifold characteristics, and designing an equivalent angle parameter estimation method based on a spatial spectrum estimation algorithm; Step S4, parameter decoupling and user positioning: designing an angle, delay and fading decoupling method based on RIS topology and free space propagation characteristics to realize user positioning application; in, The step S1, designing the topological structure and unit layout of the uniform circular array RIS, establishing the RIS-assisted communication and perception system and multipath channel model, specifically includes: Consider a uniform circular array RIS, composed of N R Reflection units are arranged around a circle with a radius of r, where the azimuth angle of element n is β n =2π(n-1) / N R , position is p n = r[cosβ n ,sinβ n ] T ; Assume that the incident angle of the far-field plane electromagnetic wave is θ and the unit direction is Then the uniform circular array RIS manifold is Where k0 = 2π [λ is the wave number of the electromagnetic wave with wavelength λ, the subscript R refers to the RIS related parameters, exp(x) = e x is an exponential function with the natural constant e as the base, (·) T represents transpose, is an imaginary unit; Consider a RIS-assisted wireless communication and sensing system, where a single base station and U users are equipped with a single antenna, and the broadband system base frequency is f c , the transmission bandwidth is f s , divided into K subcarriers; the RIS is deployed near the user side, the user-RIS channel is dominated by the line-of-sight path, and the line-of-sight path of the base station-RIS channel is blocked and is composed of non-line-of-sight scattering paths; the uplink channel h from the base station and user u to the RIS on subcarrier k BR,k and h RU,u,k They are modeled as: Among them, the subscripts B, U, D, and A refer to the base station BS, user User, departure angle AoD, and arrival angle AoA related parameters respectively, and L is h BR,k The number of component paths, f k =f c +kf s [K is the subcarrier k frequency, θ D,l ,τ BR,l ,α BR,l h BR,k The departure angle, delay and fading of the RIS side of the scattering path l, θ A,u ,τ RU,u ,α RU,u h RU,u,k The arrival angle, delay and fading of the RIS side of the line-of-sight path are modeled as α, BR,l ∝(4πf c τ BR,l ) -1 and α RU,u =(4πf c τ RU,u ) -1 , where the proportional symbol ∝ indicates that the former is affected by the unknown loss of the scatterer.

2. The channel estimation and user positioning method based on uniform circular array intelligent metasurface according to claim 1, characterized in that: The step S2, establishing a multi-carrier pilot training signal model, and designing a RIS reflection coefficient training pattern assignment method based on manifold equivalent transformation, specifically includes: Channel estimation beam training uses T tr Timeframes and K tr frequency domain subcarriers, configure the reflection coefficient array vector of the RIS array in training frame t User u sends pilot symbol x u , the uplink training signal model is in, is the equivalent cascade channel, n k,t is the channel Gaussian white noise, ⊙ represents the Khatri-Rao product, Diag(x) represents the matrix with x as the diagonal, and eq refers to the equivalent channel related parameters; according to formula (2), it can be rewritten as in, represents the Hadamard product, and the pilot symbol is Substituting equation (4) into equation (3), we get in, is the number of equivalent paths, and the two-dimensional index is mapped to the one-dimensional index is the equivalent attenuation of path p, is the equivalent delay of path p, is the RIS equivalent array manifold corresponding to path p, where θ eq,p ={θ D,l ,θ A,u } contains the angle parameters of the two paths; K tr T on subcarriers tr The received training signals in the training frames are combined into a data matrix Y=Ψ T A R,eq G eq +N (6) in, is the RIS reflection coefficient sequence in the training cycle, A R,eq =[a R (θ eq,1 ),...,a R (θ eq,P )] collects the RIS array manifold of all equivalent paths, The element at (p, k) is the equivalent complex gain α of path p on subcarrier k eq,p exp(-j2πf k τ eq,p ), They represent the received data matrix and the noise matrix respectively, and their elements at (t, k) are respectively y of formula (5) t,k ,n t,k ; According to the topological structure of the uniform circular array RIS, the phase of the electromagnetic response of the equivalent path p at the RIS array element n is: in, Returns the phase value of the input complex number. The activation function ω(β) with any period of 2π can be expressed in Fourier series as The spatial harmonics ω excited by the array m (β) = exp(jmβ) matching coefficient c m ; Using phase mode m to excite the RIS array aperture, the standardized far-field diagram can be obtained Among them, J m (ζ) is a type of m-order Bessel function, ζ = 2k0rcosφ, ζ is temporarily omitted for simplicity eq,p ,φ eq,p , Subscript; From equation (8), the visual domain φ∈[-π,π] is transformed to ζ∈[-2k0r,2k0r]. In order to make the RIS uniform circular array aperture only excite a limited number of modes, let M be the highest order of the excitable modes. According to the empirical rule, the value of M is determined to be M≈2k0r(9) when |m|≥M, The amplitude is small and can be ignored in the entire visual domain. Based on the above phase mode excitation principle, the beamforming weight vector is defined as Let T tr =2M+1, the training pattern for designing RIS reflection coefficient is According to formula (8) and the properties of Bessel function J -m (ζ)=(-1) m J m (ζ), the RIS beam space array manifold after the reflection coefficient training pattern Ψ beamforming is equivalent to in, The angle is The manifold of electromagnetic waves on a uniform linear array, J(ζ)=Diag(J M (ζ),...,J1(ζ),J0(ζ),J1(ζ),...,J M (ζ)).

3. The channel estimation and user positioning method based on uniform circular array intelligent metasurface according to claim 2, characterized in that: The step S3, designing a signal real number processing based on equivalent manifold characteristics, and designing an equivalent angle parameter estimation method based on a spatial spectrum estimation algorithm, specifically includes: Equation (12) conforms to the central Hermite structure. In order to enable the beam space manifold signal processing to be performed in the real domain, the orthogonal linear transformation is defined as where γ m =2πi,i∈[-M,M], after being processed by the central Hermite matrix W, the uniform circular array RIS manifold is transformed into real numbers: in, (·) H represents conjugate transpose; Definition C = Diag([(-1) M ,...,(-1) 1 ,1,1,...,1]), according to formula (13), we can get in, extract Front, middle, back (T tr -2) elements but Contains the same phase value, according to the recursive relation of Bessel function J m-1 (ζ)+J m+1 (ζ)=(2m / ζ)J m (ζ) can be obtained Where, Γ=2Diag(-(M-1),...,-1,0,1,...,M-1); Formula (15) can be rewritten as in Contains vector Φ=Diag(ω eq,1 ,...,ω eq,P ), The processed signal is recorded as Z = W H Y, calculate the sampling autocorrelation matrix And perform real-valued eigendecomposition to obtain the eigenvectors corresponding to the first P eigenvalues There exists a non-singular matrix Satisfy S=B R T, where Contains real-valued beam space manifolds Defining the selection matrix i=-1,0,1, extract CWS before, during and after (T tr -2) Get the sampling subspace Combination The structure is easy to prove Where D = Diag((-1) M-2 ,...,(-1) 1 ,(-1) 0 ,(-1) 1 ,...,(-1) M ), is a reverse permutation matrix with the anti-diagonal line set to 1 and the rest set to 0, then the subspace matrix form of equation (15) is in, When T tr When ≥2P, equation (16) is an overdetermined system of equations with a unique least squares solution Compute eigendecomposition Get eigenvalues Automatic pairing and recovery 4. The channel estimation and user positioning method based on uniform circular array intelligent metasurface according to claim 2, characterized in that: The step S4, designing an angle, delay and fading decoupling method based on RIS topology and free space propagation characteristics to realize user positioning application, specifically includes: According to formula (7), The definition of Among them, mod 2π (·) represents the modulo operation with 2π as the modulus; Substituting the channel path angle parameter estimate back into the RIS array manifold of formula (12), we can get G eq The least squares solution of Channel estimation training uses continuous K tr When ≤K adjacent subcarriers, the equivalent delay has a least squares solution The equivalent path fading can be recovered by Select the same multipath component l corresponding to the equivalent path According to the free space loss model, we can get The user-RIS channel propagation delay can be recovered by the following formula: In addition, there are Thus, the uniform circular array RIS-assisted channel estimation method has achieved the unique decoupling of the angle, delay and fading parameters of the channel path from the base station and the user to the RIS. Based on the above line-of-sight path parameter information, user positioning application can be directly implemented through the azimuth-distance positioning method. in, is the prior physical location of RIS, v c is the propagation speed of electromagnetic waves.

Images