RIS-assisted single-carrier frequency 3D imaging method based on multi-view image correlation

Abstract

The application discloses a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation, and belongs to the field of wireless communication and imaging. A UE sends a known pilot signal at each position, a RIS synchronously generates a random phase configuration, and an AP extracts channel state information of a twice-scattering path of ROI and RIS from a received signal; secondly, a linear mapping relationship between a channel state information measurement vector of the twice-scattering path of ROI and RIS and a ROI scattering coefficient image is established; then, according to an occlusion effect and an anisotropic scattering characteristic, multi-view image correlation characteristics of the ROI under observation of different UE positions are established; finally, according to the multi-view image correlation characteristics, the scattering coefficient image of the ROI is recovered from the channel state information measurement vector. The application has the capabilities of single-carrier 3D imaging, multi-view joint imaging and real-time imaging, and simultaneously significantly improves image estimation precision.

Description

Technical Field

[0001] This invention belongs to the field of wireless communication and imaging, and particularly relates to a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation. Background Technology

[0002] Future wireless communication systems are expected to possess ubiquitous communication and sensing capabilities, with environmental perception occurring unconsciously during communication. Real-time microwave imaging, capable of supporting functions such as surveillance, augmented reality, and environmental reconstruction, is considered a crucial component of environmental perception. Compared to visible light imaging, microwave imaging acquires only coarse-grained scattering characteristics of the target, offering better protection of personal privacy and being unaffected by visible light intensity.

[0003] In recent years, the development of intelligent metamaterials has provided numerous new approaches to microwave imaging, including the intelligent manipulation of the electromagnetic propagation environment by metamaterials to capture the scattering characteristics of targets. Reconfigurable intelligent surfaces (RIS), composed of subwavelength tunable cell arrays, are considered an important component of future mobile communication systems. RIS can be used for auxiliary communication, positioning, imaging, and other functions.

[0004] Rapid phase reconfiguration of RIS provides the foundation for real-time imaging. However, traditional RIS-assisted imaging techniques use only a single carrier frequency for two-dimensional (2D) imaging, failing to achieve comprehensive perception of three-dimensional (3D) space. Furthermore, the limited bandwidth of the communication system results in low range resolution in traditional imaging systems. Moreover, traditional multi-view joint imaging techniques have not yet been introduced into RIS-assisted imaging systems. How to fully utilize the correlation between multi-view images and achieve high-precision 3D imaging using a single carrier frequency signal is a pressing problem that needs to be solved. Summary of the Invention

[0005] Technical problem: Existing RIS-assisted imaging systems only achieve 2D imaging, have limited range resolution, and do not consider multi-view joint imaging. The purpose of this invention is to provide a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation, which fully utilizes the correlation between multi-view images to achieve high-precision single-carrier 3D imaging.

[0006] Technical solution: This invention proposes a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation, characterized by the following steps:

[0007] Step 1: The User Equipment (UE) moves along a continuous trajectory around the Region of Interest (ROI). At each location, the UE transmits a known pilot signal. This signal is scattered by the ROI and the Reconfigurable Smart Metasurface (RIS) before reaching the Access Point (AP). The RIS synchronously generates a random phase configuration in each signal time slot. The AP extracts the channel state information of each UE location and the two scattering paths through the ROI and the RIS in each time slot through channel estimation.

[0008] Step 2: Discretize the region of interest (ROI) into several cubes, and establish a linear mapping relationship between the extracted channel state information measurement vector and the scattering coefficient image of the cubes within the ROI.

[0009] Step 3: Based on the occlusion effect and anisotropic scattering characteristics, the support of the scattering coefficient image is constructed as a binary Markov chain as the user equipment (UE) observation position changes, and the amplitude of the scattering coefficient image is constructed as a Gaussian Markov process as the user equipment (UE) observation position changes, thus establishing the multi-view image correlation features of the region of interest (ROI) under different user equipment (UE) observation positions.

[0010] Step 4: Based on the correlation characteristics of multi-view images, recover the scattering coefficients of the cube within the region of interest (ROI) under different UE locations from the channel state information measurement vectors of the two scattering paths of the region of interest (ROI) and the reconfigurable smart metasurface (RIS), thereby achieving multi-view joint imaging of the ROI.

[0011] Furthermore, in step one, the user equipment (UE) moves along a continuous trajectory. Each change in UE position by a distance d0 is considered a step in UE movement, resulting in a total of T0 consecutive UE positions. At each position, the UE continuously transmits pilot signals. Within each signal time slot, the reconfigurable smart metasurface (RIS) synchronously and randomly generates a set of phase configurations to change the channel state of the two scattering paths through the region of interest (ROI) and the RIS. The access point (AP) extracts the channel state information of the two scattering paths through the ROI and the RIS using the Newton-Orthogonal Matched Pursuit (NOMP) channel estimation algorithm based on the received signal and the known pilot signals.

[0012] Furthermore, in step one, the user equipment (UE) and access point (AP) use a single antenna; the reconfigurable smart metasurface (RIS) uses a UPA array, comprising M electromagnetic elements with an element size of ξ. s ×ξ s , where ξ sThe side length of the reconfigurable smart metasurface RIS unit is defined; the scattering phase of the RIS unit is randomly selected within the interval [0, 2π]; the region of interest (ROI) is fixed in position and uniformly divided into N = N x ×N y ×N z N cubes, where N x N y and N z These represent the number of cubes along the x, y, and z axes, respectively, and the cube size is ξ. vx ×ξ vy ×ξ vz , where ξ vx ξ vy and ξ vz These are the dimensions of the cube along the x, y, and z axes, respectively; at the t-th user equipment (UE) location, when the reconfigurable smart metasurface (RIS) uses the k-th phase configuration, the channel state information extracted by the channel estimation through the region of interest (ROI) and the two scattering paths of the reconfigurable smart metasurface (RIS) is determined by the following formula:

[0013]

[0014] in, Let be the free-space channels from the location of the t-th user equipment (UE) to all cubes of the region of interest (ROI), where Let t be the free space channel from the t-th UE location to the n-th cube, where n = 1, 2, ..., N; Let x be the ROI scattering coefficient image observed by the user equipment (UE) at location t, where x t,n Let x be the scattering coefficient of the nth cube observed by the UE at position t; diag(x) t ) as x t A diagonal matrix generated for diagonal elements; Let h be the free-space channel from all cubes of the region of interest (ROI) to all cells of the reconfigurable intelligent metasurface (RIS). v,s,n,m Let m be the free space channel from the nth cube within the region of interest (ROI) to the mth cell of the reconfigurable intelligent metasurface (RIS). Let ω be the k-th phase configuration vector of the reconfigurable intelligent metasurface RIS, where ω k,m The phase of the m-th RIS unit in the RIS phase configuration of the k-th reconfigurable smart metasurface; For the free-space channels from all units of the reconfigurable smart metasurface RIS to the access point AP, where h s,a,m Let z be the free-space channel from the m-th reconfigurable smart metasurface RIS unit to the access point AP, where n = 1, 2, ..., N, m = 1, 2, ..., M;t,k This is the additive white Gaussian noise at the receiving end.

[0015] Furthermore, in step two, the region of interest (ROI) is discretized into several cubes. At the t-th user equipment (UE) location, when the reconfigurable intelligent metasurface (RIS) uses the k-th phase configuration, the extracted channel state information y obtained through the two scattering paths of the ROI and RIS is measured. t,k The scattering coefficient image of the cube within the region of interest (ROI) x t A linear mapping relationship is established using the following formula:

[0016]

[0017] in, It is the sensing vector corresponding to the reconfigurable smart metasurface RIS using the k-th phase configuration at the t-th user equipment (UE) location. There are K different reconfigurable smart metasurface RIS phase configurations at each UE location, and each RIS phase configuration can generate a channel state information measurement y. t,k By combining all channel state information measurements acquired at the location of the t-th user equipment (UE), the channel state information measurement vector obtained through the two scattering paths of ROI and RIS can be obtained. Channel state information measurement vector y t The scattering coefficient image of the cube within the region of interest (ROI) x t The linear mapping relationship is specifically determined by the following formula:

[0018] y t =A t x t +z t t=1,2,...,T0

[0019] in, Let be the sensing matrix at the location of the t-th user equipment (UE). Let be the additive white Gaussian noise vector of the receiver when the UE is at the t-th position.

[0020] Furthermore, in step three, the scattering coefficient of the nth cube within the ROI observed at the t-th user equipment (UE) location is determined by the following formula:

[0021] x t,n =s t,n a t,n

[0022] Among them, s t,n ∈{0,1} is the support symbol of the nth cube within the ROI observed at the location of the t-th user equipment (UE), s t,n=0 indicates that the nth cube within the ROI cannot be observed at the location of the t-th user equipment (UE) due to occlusion, or the observed nth cube within the ROI does not contain the target. t,n =1 indicates that the nth cube within the ROI can be observed from the t-th UE location, and that cube contains the target; a t,n Let s be the scattering coefficient amplitude observed at the location of the t-th user equipment (UE) in the n-th cube within the ROI, and let s be the support of the multi-view image. t,n Features that change with the user equipment (UE) observation location are constructed as support correlation features of multi-view images. These correlation features are described using a binary Markov chain containing two states, 0 and 1. The transition probability between the two states is determined by the following formula:

[0023]

[0024] Where, p 01 ∈[0,1] is the probability that the support symbol of the nth cube changes from 1 to 0, p 10 ∈[0,1] represents the probability that the support symbol of the nth cube changes from 0 to 1; when the binary Markov chain is in steady state, i.e., the scattering coefficient image x of the ROI. t When the probabilities of the support being 0 and 1 become stable, the transition probability p 01 and p 10 The relationship between them is specifically determined by the following formula:

[0025]

[0026] Where α is the scattering coefficient image x t The sparsity of the scattering coefficient image x t The ratio of the number of elements with a support sign of 1 to the total number of elements, and the amplitude 'a' of the scattering coefficient image. t,n The feature that varies with the observation location of the user equipment (UE) is constructed as the amplitude correlation feature of the multi-view image. This correlation feature is described using a Gaussian Markov process. The change in the amplitude of the scattering coefficient of the nth cube within the ROI between the (t-1)th and tth UE locations is determined by the following formula:

[0027] a t,n =ρa t-1,n +(1-ρ)e t,n

[0028] Where ρ is the time correlation parameter of the scattering coefficient amplitude, e t,n It is a Gaussian random perturbation variable; e t,n The mean is determined by the following formula:

[0029] η e =η

[0030] Where η is the scattering coefficient image of the ROI. t The mean of the non-zero elements, e t,n The variance is determined by the following formula:

[0031]

[0032] in, For the scattering coefficient image of the region of interest (ROI) x t The variance of non-zero elements, and the multi-view image correlation features of the region of interest (ROI) under different user equipment (UE) locations are jointly described by the support correlation feature and the amplitude correlation feature. The correlation feature parameters are:

[0033] Furthermore, in step four, based on Bayes' theorem, the scattering coefficient image of the region of interest (ROI) is recovered from the channel state information measurement vectors of the two scattering paths through the region of interest (ROI) and the reconfigurable smart metasurface (RIS), using the minimum mean square error (MMSE) estimator, and determined by the following formula:

[0034]

[0035] in, For x t,n The MMSE estimation results, For x t,n The estimated value, For x t,n The corresponding random variable, This is a vector composed of the measured and stacked channel state information from all extracted paths through the ROI and RIS scattering paths. for The corresponding random vector, Represents the numerical expectation, with respect to x. t,n The MMSE estimation is transformed into a given under the condition of x t,n Expectations Solving this problem is equivalently transformed into solving the problem with a given... under the condition of x t,n probability density function The solution is based on the scattering coefficient x. t,n With branches t,n and amplitude a t,n The relationship between the probability density functions and the correlation features of multi-view images. The solution can be written as the joint probability density function The solution is determined using the following formula:

[0036]

[0037] in, This is a vector composed of stacked scattering coefficient images observed at all UE locations. This is a vector formed by stacking the supports of the scattering coefficient images observed at all UE locations. Let be the support vector of the scattering coefficient image observed at the t-th UE location. This is a vector composed of the amplitudes of the scattering coefficient images observed at all UE locations. Let be the amplitude vector of the scattering coefficient image observed at the t-th UE location, where ∝ denotes a direct proportion. The parameter is Given x t Under the condition of y t,k The prior probability density function, based on y t,k The additive noise form is obtained. The parameter is Given s t,n and a t,n under the condition of x t,n The probability density function, The parameter is Given s t-1,n under the conditions of s t,n The probability density function, and The parameter is Given a t-1,n Under the condition a t,n The probability density function, and

[0038] Furthermore, in step four, the Expectation Maximization-Turbo-Generalized Approximation Message Passing (EM-Turbo-GAMP) algorithm is used to complete the joint probability density function. Solution: Initialize based on multi-view image correlation features Afterwards, according to and the channel state information measurement vector y at all UE locations t Image of scattering coefficients of a cube within the ROI x t The mapping relationship is obtained using the generalized approximate message-passing GAMP algorithm. and The intermediate solution; based on the output of the GAMP algorithm, the Forward-Backward algorithm is used to update the multi-view image support correlation features. Updating based on multi-view image amplitude correlation features Estimating correlation feature parameters using the Expectation-Maximization (EM) algorithm. Repeat the above steps until the maximum number of iterations is reached, and obtain the joint probability density function based on the calculation result of the last iteration. The solution results yield multi-view images x. t The MMSE estimation results are used to complete the multi-view joint imaging of the ROI, t = 1, 2, ..., T0.

[0039] Beneficial effects: Compared with the prior art, the technical solution of the present invention has the following beneficial technical effects:

[0040] This invention provides a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation. By fully utilizing the correlation of multi-view images, imaging accuracy can be improved, and 3D imaging can be achieved using only a single carrier frequency, achieving a significant performance gain compared to existing methods. The imaging process simultaneously considers the correlation between the support and amplitude of multi-view images, matching the physical signal propagation model, thus improving image estimation accuracy compared to existing methods. Measurements are obtained through rapid changes in the RIS phase, significantly accelerating the acquisition of imaging data compared to existing methods, enabling real-time imaging. Attached Figure Description

[0041] Figure 1 This is a schematic diagram illustrating an application scenario of a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation in an embodiment of the present invention.

[0042] Figure 2 This is a schematic diagram of ROI discretization in a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation in an embodiment of the present invention.

[0043] Figure 3 This is a flowchart illustrating the steps of a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation in an embodiment of the present invention.

[0044] Figure 4 This is a schematic diagram of multi-view image modeling for a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation in an embodiment of the present invention.

[0045] Figure 5 This is a schematic diagram of the EM-turbo-GAMP algorithm for a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation in an embodiment of the present invention.

[0046] Figure 6 This is a simulation result diagram of the estimation accuracy using individual imaging from each viewpoint and multi-view joint imaging in an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0048] Before introducing the technical solution of this invention, the applicable application scenarios of this invention will be described first. The solution of this invention is for a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation. Please refer to... Figure 1 , Figure 1 This is a schematic diagram illustrating an application scenario of a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation in an embodiment of the present invention. For example... Figure 1 As shown, the system operates in 3D space. The system includes an access point (AP), a signal recovery array (RIS), an imaging target located within the area of ​​interest (ROI), and a user equipment (UE). The AP's received signal is processed in a processing unit, and the RIS phase is controlled by the AP via a control unit. The operating frequency is f = 3 GHz, and the wavelength is λ = 10 cm. In this application scenario, both the AP and UE use a single antenna, and the RIS uses a uniform planar array. The RIS center coincides with the origin, and the RIS elements are arranged along the y-axis and z-axis. The RIS has a total of M = 48 × 48 adjustable elements, each with dimensions of [missing information]. The ROI is uniformly divided into N = 1000 cubes, each with a side length of ξ. vx =ξ vy =ξ vz A cube with a value of 2λ, such as Figure 2 As shown; the UE's continuous movement trajectory in space Randomly generated in space; the center of the ROI is also located in space. The ROI is randomly distributed within the RIS aperture, with z = 0; the ROI is located in the near field of the RIS aperture, and there is no LOS path between the ROI and the AP. The UE transmits a known pilot signal while moving along a continuous trajectory around the ROI, and this signal is received by the AP. The UE position, AP position, RIS position, and RIS phase configuration are all known at the processing unit.

[0049] The technical solution of this invention will be described next. Please refer to [link / reference]. Figure 3 , Figure 3 This is a flowchart illustrating the steps of a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation in an embodiment of the present invention. In this embodiment, the specific steps of the present invention include:

[0050] Step 1: The User Equipment (UE) moves along a continuous trajectory around the Region of Interest (ROI). At each location, the UE transmits a known pilot signal. This signal is scattered by the ROI and the Reconfigurable Smart Metasurface (RIS) before reaching the Access Point (AP). The RIS synchronously generates a random phase configuration in each signal time slot. The AP extracts the channel state information of each UE location and each time slot through the two scattering paths via the ROI and the RIS using channel estimation.

[0051] In this embodiment, the user equipment (UE) moves along a continuous trajectory. Each change in UE position by a distance d0 is considered a step for the UE, resulting in a total of T0 = 10 consecutive UE positions. At each position, the UE continuously transmits pilot signals. Within each signal time slot, the reconfigurable smart metasurface (RIS) synchronously and randomly generates a set of phase configurations to change the channel state of the two scattering paths through the region of interest (ROI) and the RIS. The access point (AP), based on the received signal and the known pilot signals, extracts the channel state information of the two scattering paths through the ROI and the RIS using the Newton-Orthogonal Matched Pursuit (NOMP) channel estimation algorithm, which is then used for imaging the ROI.

[0052] In this embodiment, at the t-th user equipment (UE) location, when the reconfigurable smart metasurface (RIS) uses the k-th phase configuration, the channel state information extracted by the channel estimation through the region of interest (ROI) and the two scattering paths of the reconfigurable smart metasurface (RIS) is determined by the following formula:

[0053]

[0054] in, Let be the free-space channels from the location of the t-th user equipment (UE) to all cubes of the region of interest (ROI), where Let t be the free space channel from the t-th UE location to the n-th cube, where n = 1, 2, ..., 1000; Let x be the ROI scattering coefficient image observed by the user equipment (UE) at location t, where x t,n Let x be the scattering coefficient of the nth cube observed by the UE at position t; diag(x) t ) as x t A diagonal matrix generated for diagonal elements; Let h be the free-space channel from all cubes of the region of interest (ROI) to all cells of the reconfigurable intelligent metasurface (RIS). v,s,n,mLet m be the free space channel from the nth cube within the region of interest (ROI) to the mth cell of the reconfigurable intelligent metasurface (RIS). Let ω be the k-th phase configuration vector of the reconfigurable intelligent metasurface RIS, where ω k,m The phase of the m-th RIS unit in the RIS phase configuration of the k-th reconfigurable smart metasurface; For the free-space channels from all units of the reconfigurable smart metasurface RIS to the access point AP, where h s,a,m Let z be the free-space channel from the m-th reconfigurable smart metasurface RIS unit to the access point AP, where n = 1, 2, ..., 1000 and m = 1, 2, ..., 2304; t,k This is the additive white Gaussian noise at the receiving end.

[0055] Step 2: Discretize the region of interest (ROI) into several cubes, and establish a linear mapping relationship between the extracted channel state information measurement vector and the scattering coefficient image of the cubes within the ROI.

[0056] In this embodiment, the Region of Interest (ROI) is discretized into several cubes. At the t-th User Equipment (UE) location, when the Reconfigurable Smart Metasurface (RIS) uses the k-th phase configuration, the extracted channel state information y obtained through the two scattering paths of the ROI and RIS is measured. t,k The scattering coefficient image of the cube within the region of interest (ROI) x t A linear mapping relationship is established using the following formula:

[0057]

[0058] in, It is the sensing vector corresponding to the reconfigurable smart metasurface RIS using the k-th phase configuration at the t-th user equipment (UE) location. There are K = 80 different reconfigurable smart metasurface RIS phase configurations at each UE location, and each RIS phase configuration can generate a channel state information measurement y. t,k By combining all channel state information measurements acquired at the location of the t-th user equipment (UE), the channel state information measurement vector obtained through the two scattering paths of ROI and RIS can be obtained. Channel state information measurement vector y t The scattering coefficient image of the cube within the region of interest (ROI) x t The linear mapping relationship is specifically determined by the following formula:

[0059] y t =A t x t +z t t=1,2,...,T0

[0060] in, Let be the sensing matrix at the location of the t-th user equipment (UE). Let be the additive white Gaussian noise vector of the receiver when the UE is at the t-th position.

[0061] Step 3: Based on the occlusion effect and anisotropic scattering characteristics, the support of the scattering coefficient image is constructed as a binary Markov chain as the user equipment (UE) observation position changes, and the amplitude of the scattering coefficient image is constructed as a Gaussian Markov process as the user equipment (UE) observation position changes. Multi-view image correlation features of the region of interest (ROI) under different user equipment (UE) observation positions are established.

[0062] In this embodiment, the scattering coefficient of the nth cube within the ROI observed at the t-th user equipment (UE) location is determined by the following formula:

[0063] x t,n =s t,n a t,n

[0064] Among them, s t,n ∈{0,1} is the support symbol of the nth cube within the ROI observed at the location of the t-th user equipment (UE), s t,n =0 indicates that the nth cube within the ROI cannot be observed at the location of the t-th user equipment (UE) due to occlusion, or the observed nth cube within the ROI does not contain the target. t,n =1 indicates that the nth cube within the ROI can be observed from the t-th UE location, and that cube contains the target; a t,n Let s be the scattering coefficient amplitude observed at the location of the t-th user equipment (UE) in the n-th cube within the ROI, and let s be the support of the multi-view image. t,n Features that change with the user equipment (UE) observation location are constructed as support correlation features of multi-view images. These correlation features are described using a binary Markov chain containing two states, 0 and 1. The transition probability between the two states is determined by the following formula:

[0065]

[0066] Where, p 01 =0.1 represents the probability that the support symbol of the nth cube changes from 1 to 0. 10 =0.002 represents the probability that the support symbol of the nth cube changes from 0 to 1; when the binary Markov chain is in steady state, i.e., the scattering coefficient image x of the ROI. t When the probabilities of the support being 0 and 1 become stable, the transition probability p 01 and p10 The relationship between them is specifically determined by the following formula:

[0067]

[0068] Where α = 0.02 is the scattering coefficient image x t The sparsity of the scattering coefficient image x t The ratio of the number of elements with a support sign of 1 to the total number of elements, and the amplitude 'a' of the scattering coefficient image. t,n The feature that varies with the observation location of the user equipment (UE) is constructed as the amplitude correlation feature of the multi-view image. This correlation feature is described using a Gaussian Markov process. The change in the amplitude of the scattering coefficient of the nth cube within the ROI between the (t-1)th and tth UE locations is determined by the following formula:

[0069] a t,n =ρa t-1,n +(1-ρ)e t,n

[0070] Where ρ = 0.9 is the time correlation parameter of the scattering coefficient amplitude, e t,n It is a Gaussian random perturbation variable; e t,n The mean is determined by the following formula:

[0071] η e =η

[0072] Where η = 1 is the ROI scattering coefficient image x t The mean of the non-zero elements, e t,n The variance is determined by the following formula:

[0073]

[0074] in, For the scattering coefficient image of the region of interest (ROI) x t The variance of non-zero elements, and the multi-view image correlation characteristics of the region of interest (ROI) under different user equipment (UE) locations are jointly described by the support correlation characteristics and the magnitude correlation characteristics, such as... Figure 4 As shown, the correlation feature parameters are

[0075] Step 4: Based on the correlation characteristics of multi-view images, recover the scattering coefficient of the cube within the region of interest (ROI) under different UE positions from the channel state information measurement vectors of the two scattering paths of the region of interest (ROI) and the reconfigurable smart metasurface (RIS), thereby realizing multi-view joint imaging of the ROI.

[0076] In this embodiment, based on Bayes' theorem, the minimum mean square error (MMSE) estimator is used to recover the scattering coefficient image of the region of interest (ROI) from the channel state information measurement vectors of the two scattering paths through the region of interest (ROI) and the reconfigurable smart metasurface (RIS), determined by the following formula:

[0077]

[0078] in, For x t,n The MMSE estimation results, For x t,n The estimated value, For x t,n The corresponding random variable, This is a vector composed of the measured and stacked channel state information from all extracted paths through the ROI and RIS scattering paths. for The corresponding random vector, Represents the numerical expectation, with respect to x. t,n The MMSE estimation is transformed into a given under the condition of x t,n Expectations Solving this problem is equivalently transformed into solving the problem with a given... under the condition of x t,n probability density function The solution is based on the scattering coefficient x. t,n With branches t,n and amplitude a t,n The relationship between the probability density functions and the correlation features of multi-view images. The solution can be written as the joint probability density function The solution is determined using the following formula:

[0079]

[0080] in, This is a vector composed of stacked scattering coefficient images observed at all UE locations. This is a vector formed by stacking the supports of the scattering coefficient images observed at all UE locations. Let be the support vector of the scattering coefficient image observed at the t-th UE location. This is a vector composed of the amplitudes of the scattering coefficient images observed at all UE locations. Let be the amplitude vector of the scattering coefficient image observed at the t-th UE location, where ∝ denotes a direct proportion. The parameter is Given x t Under the condition of y t,k The prior probability density function, based on yt,k The additive noise form is obtained. The parameter is Given s t,n and a t,n under the condition of x t,n The probability density function, The parameter is Given s t-1,n under the conditions of s t,n The probability density function, and The parameter is Given a t-1,n Under the condition a t,n The probability density function, and

[0081] In this implementation, the Expectation Maximization-Turbo-Generalized Approximation Message Passing (EM-Turbo-GAMP) algorithm is used to complete the calculation of the joint probability density function. The solution, such as Figure 5 As shown: Initialization based on multi-view image correlation features Afterwards, according to and the channel state information measurement vector y at all UE locations t Image of scattering coefficients of a cube within the ROI x t The mapping relationship is obtained using the generalized approximate message-passing GAMP algorithm. and The intermediate solution; based on the output of the GAMP algorithm, the Forward-Backward algorithm is used to update the multi-view image support correlation features. Updating based on multi-view image amplitude correlation features Estimating correlation feature parameters using the Expectation-Maximization (EM) algorithm. Repeat the above steps until the maximum number of iterations is reached, and obtain the joint probability density function based on the calculation result of the last iteration. The solution results yield multi-view images x. t The MMSE estimation results were used to complete the multi-view joint imaging of the ROI, t=1,2,...,10.

[0082] In a specific simulation example, the simulation parameters are as follows:

[0083] Simulation platform: MATLAB, Monte Carlo simulation count: 1000.

[0084] System workspace:

[0085] Received signal signal-to-noise ratio: 20dB.

[0086] Operating frequency: f = 3 GHz, wavelength: λ = 10 cm.

[0087] RIS: Center position coordinates [0,0,0] T Number of units M = 48 × 48, unit size The number of phase configurations at each UE location is K = 80.

[0088] ROI: Total volume 20λ×20λ×20λ, cube side length ξ vx =ξ vy =ξ vz =2λ, number of cubes N=1000, center position coordinates in space It is internally distributed and z = 0.

[0089] UE: in space Continuous trajectories are randomly generated within the system, with the spacing between consecutive UE locations d0 = 5λ, and the number of consecutive UE locations in the system T0 = 10.

[0090] Multi-view image correlation feature parameters The noise variance is determined by the signal-to-noise ratio, with sparsity α = 0.02, mean η = 1, and variance... transition probability p 01 =0.1, correlation coefficient ρ=0.9.

[0091] To evaluate the imaging accuracy of the RIS-assisted single-carrier 3D imaging based on multi-view image correlation proposed in this invention: the center position of the ROI is located in space. The simulation was performed at 5λ intervals, and the imaging accuracy of multi-view joint imaging and individual view imaging was evaluated using normalized mean square error (NMSE) as the metric. The simulation results are as follows: Figure 6 As shown, the leftmost rectangle represents the top view of the RIS, and each square grid represents a ROI location. The color intensity of the grid represents the NMSE of the image when the ROI is located there. From Figure 6 As can be seen, the NMSE of multi-view joint imaging is significantly reduced compared to imaging from each view individually, which can achieve higher precision single-carrier 3D imaging and allows imaging at greater distances.

[0092] Therefore, the simulation results confirm that when using the RIS-assisted single-carrier 3D imaging method based on multi-view image correlation proposed in this invention, the correlation of multi-view images can be effectively utilized to improve 3D imaging accuracy and expand the imaging range.

[0093] The present invention has been described in detail above with reference to the accompanying drawings, providing a RIS-assisted single-carrier 3D imaging method based on multi-view image correlation. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are illustrative and not restrictive. The embodiments are only for illustrating the technical ideas of the present invention and are not intended to limit the scope of protection of the present invention. Those skilled in the art will make changes to the specific embodiments and application scope under the guidance of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A RIS-assisted single-carrier 3D imaging method based on multi-view image correlation, characterized in that, The method includes the following steps: Step 1: The User Equipment (UE) moves along a continuous trajectory around the Region of Interest (ROI). At each location, the UE transmits a known pilot signal. This signal is scattered by the ROI and the Reconfigurable Smart Metasurface (RIS) before reaching the Access Point (AP). The RIS synchronously generates a random phase configuration in each signal time slot. The AP extracts the channel state information of each UE location and the two scattering paths through the ROI and the RIS in each time slot through channel estimation. Step 2: Discretize the region of interest (ROI) into several cubes, and establish a linear mapping relationship between the extracted channel state information measurement vector and the scattering coefficient image of the cubes within the ROI. Step 3: Based on the occlusion effect and anisotropic scattering characteristics, the support of the scattering coefficient image is constructed as a binary Markov chain as the user equipment (UE) observation position changes, and the amplitude of the scattering coefficient image is constructed as a Gaussian Markov process as the user equipment (UE) observation position changes, thus establishing the multi-view image correlation features of the region of interest (ROI) under different user equipment (UE) observation positions. Step 4: Based on the correlation characteristics of multi-view images, recover the scattering coefficient of the cube in the region of interest (ROI) under different UE positions from the channel state information measurement vector of the two scattering paths through the region of interest (ROI) and the reconfigurable smart metasurface (RIS), thereby realizing multi-view joint imaging of the region of interest (ROI). In step three, the first The ROI observed at the location of the user equipment (UE) within the _th The scattering coefficient of each cube is determined by the following formula: ； in, For the first The ROI observed at the location of the user equipment (UE) within the _th Support symbols for a cube, Indicates the first Due to occlusion, the location of the first user equipment (UE) within the ROI cannot be observed. The cube or the observed ROI within the . The cube does not contain the target. Indicates the first At the location of the UE, the ROI within the first UE location can be observed. There are cubes, and each cube contains the target; For the ROI The cube in the... The amplitude of the scattering coefficient observed at the location of each user equipment (UE), and the support of the multi-view image. Features that change with the user equipment (UE) observation location are constructed as support correlation features of multi-view images. These correlation features are described using a binary Markov chain containing two states, 0 and 1. The transition probability between the two states is determined by the following formula: ； in, For the first The probability that the support sign of a cube changes from 1 to 0. For the first The probability of the support sign of a cube changing from 0 to 1; the scattering coefficient image of the ROI when the binary Markov chain is in steady state. When the probabilities of the support being 0 and 1 become stable, the transition probability... and The relationship between them is specifically determined by the following formula: ； in, For scattering coefficient image The sparsity of the scattering coefficient image The ratio of the number of elements with a support sign of 1 to the total number of elements, and the amplitude of the scattering coefficient image. Features that vary with the user equipment (UE) observation location are constructed as amplitude correlation features of multi-view images. These correlation features are described using a Gaussian Markov process, with the first value within the ROI being... The scattering coefficient amplitude of the cube in the th... The location of the user equipment (UE) and the first The changes in the location of each User Equipment (UE) are determined by the following formula: ； in, The time correlation parameter of the scattering coefficient amplitude. It is a Gaussian random perturbation variable; The mean is determined by the following formula: ； in, Image of ROI scattering coefficients The mean of the non-zero elements in the middle. The variance is determined by the following formula: ； in, Scattering coefficient image of the region of interest (ROI) The variance of non-zero elements, and the multi-view image correlation features of the region of interest (ROI) under different user equipment (UE) locations are jointly described by the support correlation feature and the amplitude correlation feature. The correlation feature parameters are: ; In step four, based on Bayes' theorem, the minimum mean square error (MMSE) estimator is used to recover the scattering coefficient image of the region of interest (ROI) from the channel state information measurement vectors of the two scattering paths through the region of interest (ROI) and the reconfigurable smart metasurface (RIS), determined by the following formula: ； in, for The MMSE estimation results, for The estimated value, for The corresponding random variable, This is a vector composed of the measured and stacked channel state information from all extracted ROI and RIS scattering paths. for The corresponding random vector, Represents the numerical expectation, for The MMSE estimation is transformed into a given Under the conditions Expectations Solving this problem is equivalently transformed into solving the problem with a given... Under the conditions probability density function The solution is based on the scattering coefficient. With branches and amplitude The relationship between the probability density functions and the correlation features of multi-view images. The solution can be written as the joint probability density function The solution is determined using the following formula: ； in, This is a vector composed of stacked scattering coefficient images observed at all UE locations. This is a vector formed by stacking the supports of the scattering coefficient images observed at all UE locations. For the first The support vector of the scattering coefficient image observed at each UE location. This is a vector composed of the amplitudes of the scattering coefficient images observed at all UE locations. For the first The amplitude vector of the scattering coefficient image observed at each UE location. Indicates direct proportion. The parameter is In the given Under the conditions The prior probability density function, according to The additive noise form is obtained. The parameter is In the given and Under the conditions The probability density function, The parameter is In the given Under the conditions The probability density function, and , The parameter is In the given Under the conditions The probability density function, and .

2. The RIS-assisted single-carrier 3D imaging method based on multi-view image correlation according to claim 1, characterized in that, In step one, the movement path of the user equipment (UE) is a continuous trajectory, and the UE's position changes by a certain distance each time. Considering a user equipment (UE) move one step, there are a total of A continuous user equipment (UE) location; The user equipment (UE) continuously transmits pilot signals at each location. Within each signal time slot, the reconfigurable intelligent metasurface (RIS) synchronously and randomly generates a set of phase configurations to change the channel state through the region of interest (ROI) and the two scattering paths of the reconfigurable intelligent metasurface (RIS). The access point (AP) extracts the channel state information through the region of interest (ROI) and the reconfigurable smart metasurface (RIS) double scattering path based on the received signal and the known pilot signal using the Newton Orthogonal Matched Pursuit (NOMP) channel estimation algorithm.

3. The RIS-assisted single-carrier 3D imaging method based on multi-view image correlation according to claim 1, characterized in that, In step one, the user equipment (UE) and access point (AP) use a single antenna; the reconfigurable smart metasurface (RIS) uses a UPA array, which contains a total of One electromagnetic unit, the unit size is ,in, The side length of the reconfigurable smart metasurface RIS unit; the scattering phase of the RIS unit is in the interval... The region of interest (ROI) is randomly selected within the region; the location of the ROI remains fixed and is uniformly divided into... A cube, in which , and These represent the number of cubes along the x, y, and z axes, respectively, and the cube dimensions are... ,in, , and These are the dimensions of the cube along the x, y, and z axes, respectively; in the... At each user equipment (UE) location, the reconfigurable intelligent metasurface RIS uses the first In the case of a phase configuration, the channel state information extracted by the channel estimation through the region of interest (ROI) and the reconfigurable smart metasurface (RIS) double scattering path is determined by the following formula: ； in, For from the first The free-space channels from the location of each User Equipment (UE) to all cubes of the Region of Interest (ROI), where... For the first The UE position to the first Free space channels of a cube, ; For User Equipment (UE) in the The scattering coefficient images of the ROI observed at each location, where, For UE in the The first observation at the location The scattering coefficient of a cube; For A diagonal matrix generated for diagonal elements; For all cubes of the region of interest (ROI) to all cells of the reconfigurable intelligent metasurface (RIS), therein, For the region of interest (ROI) within the 1st From a cube to the reconfigurable intelligent metasurface RIS Free space channels for each unit; The first reconfigurable intelligent metasurface RIS There are phase configuration vectors, where... For the first In the first reconfigurable intelligent metasurface RIS phase configuration The phase of each RIS unit; For the free-space channels from all units of the reconfigurable smart metasurface RIS to the access point AP, where, For the first A free-space channel from a reconfigurable intelligent metasurface RIS unit to the access point (AP). , ; This is the additive white Gaussian noise at the receiving end.

4. The RIS-assisted single-carrier 3D imaging method based on multi-view image correlation according to claim 3, characterized in that, In step two, the region of interest (ROI) is discretized into several cubes, and in the first step... At each user equipment (UE) location, the reconfigurable intelligent metasurface RIS uses the first When configuring each phase, the channel state information extracted from the ROI and RIS double scattering paths will be measured. Scattering coefficient image of a cube within the region of interest (ROI) A linear mapping relationship is established using the following formula: ； in, , it is in the first At each user equipment (UE) location, the reconfigurable intelligent metasurface RIS uses the first The sensing vector corresponding to each phase configuration, and the location of each user equipment (UE) has... Each of the three different reconfigurable smart metasurface RIS phase configurations can generate a channel state information measurement. , will the By combining all channel state information measurements acquired at the location of each user equipment (UE), the channel state information measurement vector obtained through the two scattering paths of ROI and RIS can be obtained. Channel state information measurement vector Scattering coefficient image of a cube within the region of interest (ROI) The linear mapping relationship is specifically determined by the following formula: ； in, For the first The perception matrix at the location of each user equipment (UE) For UE located in the The additive white Gaussian noise vector at the receiver at each position.

5. The RIS-assisted single-carrier 3D imaging method based on multi-view image correlation according to claim 1, characterized in that, In step four, the Expectation Maximization-Turbo-Generalized Approximation Message Passing (EM-Turbo-GAMP) algorithm is used to complete the joint probability density function. Solution: Initialize based on multi-view image correlation features , , , , Afterwards, according to , and channel state information measurement vectors at all UE locations Scattering coefficient image of a cube within the ROI The mapping relationship is obtained using the generalized approximate message-passing GAMP algorithm. and The intermediate solution; based on the output of the GAMP algorithm, the Forward-Backward algorithm is used to update the multi-view image support correlation features. Updating based on multi-view image amplitude correlation features The Expectation-Maximization (EM) algorithm is used to estimate the correlation feature parameters. Repeat the above steps until the maximum number of iterations is reached, and obtain the joint probability density function based on the calculation result of the last iteration. The solution results yield multi-view images. Based on the MMSE estimation results, multi-view joint imaging of the ROI was completed. .

Images