Fluid antenna source positioning method and device based on aperture dynamic reconfiguration

Abstract

The invention discloses a fluid antenna source positioning method and device based on aperture dynamic reconfiguration, and relates to the technical field of wireless communication and array signal processing. The scheme is provided for solving the problems of limited positioning precision and insufficient environmental adaptability caused by a fixed array structure and a single source field model. The method comprises the following steps: S1, constructing a fluid antenna system S-FAS with adaptive aperture adjustment capability, and establishing a unified modeling framework based on precise space geometry ESG; s2, switching between aperture compression and aperture expansion according to task requirements, collecting observation data and generating statistics; s3, direction initial estimation is carried out under the compressed aperture, and mutual coupling and amplitude-phase errors are suppressed; and S4, performing joint estimation on the target direction and distance by taking the initial value as an anchor point under the expanded aperture to realize high-precision positioning. The method does not need source type distinguishing or signal isolation, can stably work in near-field, Fresnel region and far-field multi-source mixed environments, has high precision, low complexity and strong adaptability, and is suitable for scenes such as intelligent sensing and low-altitude communication.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wireless communication and array signal processing, and particularly relates to a fluid antenna source positioning method and device based on aperture dynamic reconfiguration. BACKGROUND

[0002] Direction of arrival (DOA) estimation and source positioning are key technologies in array signal processing, and are widely used in radar monitoring, unmanned system navigation, electromagnetic spectrum sensing and other scenarios. With the advent of new demands such as the sixth generation of mobile communication and integrated sensing, the system has higher requirements for high resolution, strong robustness and low latency positioning capability in complex environments.

[0003] Existing technologies mostly rely on fixed structure arrays and use far-field plane wave or near-field spherical wave models. However, in actual communication and sensing tasks, mixed field domains with multiple sources frequently occur, and the target distance distribution spans the near field, Fresnel zone and far field, resulting in significant wave front mismatch and estimation deviation. At the same time, fixed aperture structures are difficult to flexibly regulate array geometry, and it is difficult to balance angle resolution, mutual coupling suppression and modeling accuracy, which limits the performance upper limit of the system in dynamic complex scenarios.

[0004] To solve the problems of limited source positioning accuracy and insufficient rigidity of array structure, a new type of array paradigm with reconfigurable capability has been proposed in recent years. By encapsulating conductive liquid in a controllable medium and combining electromagnetic driving mechanism, the system realizes the rapid migration and dynamic reconfiguration of array elements in physical space, breaks through the geometric rigidity limitation of traditional arrays, and has the advantages of high structural flexibility, fast response speed and strong adaptability. This provides a new physical basis for building an adaptive array positioning system for complex field domains, and also puts forward higher requirements for array architecture and spatial modeling methods. Therefore, it is urgent to propose a new reconfigurable array system that integrates fluid antenna architecture, and to establish a unified spatial modeling and source positioning method to realize high-precision modeling, flexible structure regulation and engineering realizability in mixed field domains such as near field, Fresnel zone and far field, and to break through the performance bottleneck and adaptability limitation of traditional arrays in complex scene source positioning. SUMMARY

[0005] The present application aims to provide a fluid antenna source positioning method and device based on aperture dynamic reconfiguration to solve the problems of the prior art.

[0006] The fluid antenna source positioning method based on aperture dynamic reconfiguration in the present application comprises the following steps: S1. Construct a fluid antenna system with array scalability and dynamic reconfiguration capability, realize adaptive switching between compressed aperture and extended aperture configurations through scaling of array element spacing, and establish a unified modeling and processing framework to support near-far field mixed signal source positioning; accurately model the spatial coherence characteristics under different array configurations and compensate the response; S2. The fluid antenna system switches between compressed aperture configuration and extended aperture configuration according to real-time task requirements; the fluid antenna system integrates a software controllable mechanical scaling execution mechanism, dynamically adjusts the spacing between adjacent array elements, and realizes rapid reconfiguration and switching of the array aperture structure; at each configuration switching, the fluid antenna system automatically records the current configuration identifier and the corresponding time stamp to ensure the cross-configuration coherence of the collected data within the same task cycle; In each configuration mode, the fluid antenna system takes a snapshot for continuous observation, and constructs an observation data matrix under compressed aperture and an observation data matrix under extended aperture, respectively; wherein, represents the array receiving vector collected at the th time snapshot under compressed configuration, represents the array receiving vector collected at the th time snapshot under extended configuration. The fluid antenna system calculates the corresponding sample covariance matrices and based on the two groups of data matrices, respectively, to characterize the spatial statistical characteristics under compressed aperture and extended aperture configurations; S3. Under compressed aperture configuration, the fluid antenna system performs initial estimation of the signal source incident direction DOA based on the multi-snapshot observation data collected in step S2. The mutual coupling matrix describes the mutual coupling behavior between the array elements with translational invariance, and the receiving model under the th time snapshot is constructed as: ; wherein, is the far-field array manifold under compressed configuration, is the incident angle set of the signal source to be estimated, represents the far-field direction vector, is the observation noise vector, is the effective source signal vector; The selection matrix is constructed to extract the central subarray observation , and the sample covariance matrix And perform feature decomposition to extract the signal subspace and noise subspace: ; in, This represents the eigenvector matrix of the signal subspace under the compressed configuration. This represents the eigenvector matrix of the noise subspace under the compressed configuration. and This is the corresponding eigenvalue diagonal matrix; Constructing the directional spectrum function under compressed configuration: ; in, Far-field direction vector center OK; Select the function with the largest amplitude from the directional spectrum function. The angles corresponding to each peak constitute the initial DOA estimation set under compressed aperture. ; S4. In the expanded aperture configuration Under the condition of data collected in step S2 Calculate the sample covariance matrix: ; in, This represents the eigenvector matrix of the signal subspace under the extended aperture configuration. To configure the eigenvector matrix of the noise subspace under the extended aperture, Let be the diagonal matrix of eigenvalues ​​of the corresponding signal subspace. This is the diagonal matrix of eigenvalues ​​for the corresponding noise subspace; An optimization estimation is performed using a two-stage local search strategy guided by directional priors: In the first stage, the coarse orientation estimate obtained under the compressed configuration in step S3 is... Using angular anchor points, after fixing the angle for each direction, the distance spectrum function is constructed along the distance dimension: ; By performing a one-dimensional distance spectrum peak search on each angle using the aforementioned distance spectrum function, the initial distance estimate for the corresponding direction is obtained. ; Based on this, we proceed to the second stage, which involves preliminary parameter estimation. Centered on, in a local window of two-dimensional polar coordinates , Within, construct a two-dimensional spatial spectral function that combines direction and distance: ; Finally, the peak position was selected. for the final fine estimation result.

[0007] The fluid antenna device in the application is provided with the fluid antenna system, and the signal source positioning is performed by using the source positioning method.

[0008] The fluid antenna source positioning method and device based on aperture dynamic reconfiguration have the advantages that adaptive reconstruction of an array element space configuration is taken as the core, the reconfigurable potential of a fluid antenna unit in a physical layout is fully released, a dynamic aperture array architecture with double-state switching capability of compression and expansion is constructed, in the initial stage, the system controls the liquid antenna elements to converge in a local area to form a high-density compressed aperture mode, a standard MUSIC algorithm is used to realize fast coarse estimation of a target direction under low sample and low complexity conditions, and direction positioning in the compressed aperture stage is completed. Then, the system drives the element space to expand, and combines ESG to enter the expanded aperture stage, higher angle resolution is obtained through a spectrum peak refinement mechanism, and then the target distance and angle are solved in linkage, and high-precision reconstruction of the spatial position is realized. The mechanism skillfully combines the near-field and far-field propagation characteristics, has signal independence and frequency band adaptability, can adapt to different modulation waveforms and electromagnetic environments, and has good positioning consistency in various propagation regions such as near field, Fresnel region and far field. Experimental results show that the scheme has superior robustness and resolution ability under typical low-altitude complex environment, multi-source interference background and low sample conditions, is especially suitable for urban low-altitude security, electromagnetic spectrum intelligent sensing, illegal signal source monitoring, dynamic frequency scheduling and other high-dynamic communication sensing fusion scenes, has good system compatibility, and has good engineering application and industrialization prospect. BRIEF DESCRIPTION OF DRAWINGS

[0009] Figure 1 It is a general flowchart of the fluid antenna source positioning method in the application.

[0010] Figure 2 It is an initial direction estimation processing flowchart under a compressed aperture configuration.

[0011] Figure 3 It is a joint angle-distance fine estimation flowchart under an expanded aperture configuration.

[0012] Figure 4 It is an initial angle spectrum curve of the compressed aperture MUSIC algorithm.

[0013] Figure 5 It is an angle spectrum curve of a traditional far-field MUSIC algorithm.

[0014] Figure 6 It is a one-dimensional distance spectrum curve under the initial estimated angle .

[0015] Figure 7 To estimate the angle in the initial stage The following is a one-dimensional distance spectrum curve.

[0016] Figure 8 To estimate the angle in the initial stage The following is a one-dimensional distance spectrum curve.

[0017] Figure 9 To estimate the angle in the initial stage The following is a one-dimensional distance spectrum curve.

[0018] Figure 10 For source 1 Two-dimensional angle-distance joint spectral distribution map.

[0019] Figure 11 Source 2 Two-dimensional angle-distance joint spectral distribution map.

[0020] Figure 12 Source 3 Two-dimensional angle-distance joint spectral distribution map.

[0021] Figure 13 Source 4 Two-dimensional angle-distance joint spectral distribution map.

[0022] Figure 14 The graph shows the root mean square error of DOA estimation for mixed field sources as a function of signal-to-noise ratio.

[0023] Figure 15 The graph shows the root mean square error of the distance estimation for mixed field sources as a function of the signal-to-noise ratio.

[0024] Figure 16 This is a graph showing the root mean square error of DOA estimation for mixed field sources as a function of the number of snapshots.

[0025] Figure 17 This is a graph showing the root mean square error of the distance estimation for mixed field sources as a function of the signal-to-noise ratio.

[0026] Figure 18 This is a graph showing the root mean square error of the distance estimation for the mixed field source as a function of the number of snapshots. Detailed Implementation

[0027] The fluid antenna source localization method based on aperture dynamic reconfiguration described in this invention is as follows: Figures 1 to 3 As shown, it includes the following steps: S1. Construct a fluid antenna system (S-FAS) with array scalability and dynamic reconfiguration capabilities. By controlling the scaling of array element spacing, adaptive switching between compressed and expanded aperture configurations is achieved. A unified modeling and processing framework supporting the localization of near-field and far-field mixed signal sources is established. Through precise spatial geometric modeling and response compensation of the spatial coherence characteristics under different array configurations, the system's multi-source resolution and localization capabilities in mixed fields are improved.

[0028] The array geometry of the fluid antenna system consists of M antenna elements arranged in a uniform linear array (ULA) along a linear array direction. The spacing between adjacent antenna elements... Subject to scaling factor control, The initial baseline spacing of the array. This indicates a compressed configuration. This indicates an extended configuration. The fluid antenna system is configured... The total array aperture is expressed as: ; This structure provides a controllable physical basis for subsequent spatial resolution adjustment and modeling.

[0029] Construction of an accurate spatial geometry model: To avoid the fragmented modeling problem of pre-dividing far-field / near-field sources required in traditional methods, the fluid antenna system introduces an accurate spatial geometry (ESG) modeling method to achieve unified response modeling of multiple signal sources in a mixed field. Here, accurate spatial geometry (ESG) refers to exact spatial geometry.

[0030] Any number The positions of the signal sources are represented in polar coordinates. , and the first in the array The propagation distance between each array element is denoted as . Constructing spatial response vectors under an accurate spatial geometric model. , is used to represent the phase response characteristics of a signal from the source point to each unit of the array, and its elements are determined by the corresponding propagation distance and signal wavelength parameters.

[0031] The method of constructing a precise spatial geometric model is applicable to any source distance and incident angle, can accurately reflect the geometric characteristics during signal propagation, and has a unified adaptability to near-field, far-field and mixed scenarios, eliminating the near-field and far-field modeling discontinuity problem existing in traditional methods.

[0032] To establish an auxiliary index for field determination, and in order to achieve dynamic adaptive switching between modeling methods and estimation algorithms for the fluid antenna system under different propagation conditions, a Rayleigh distance criterion based on array aperture configuration is defined. : ; wherein, denotes the Rayleigh distance of baseline aperture, denotes the working wavelength. The Rayleigh distance criterion is used as the core basis of modeling selection, to judge the spatial propagation mechanism between the signal source and the array in real time, spherical wave or plane wave, thus serving as the physical boundary of near-field / far-field modeling method switching. It fully embodies the real-time regulation of array aperture adjustment on the near / far-field boundary: when decreases, i.e. the aperture is compressed, the effective aperture of the array is reduced, resulting in decreases, and the near-field region shrinks accordingly; when increases, i.e. the aperture is expanded, the aperture increases, resulting in significantly increases, and the far-field boundary retreats, and part of the mid-distance signal source may cross into the near-field.

[0033] In the actual operation process of the fluid antenna system, after each array structure switching, the device will automatically call the Rayleigh distance criterion to compare the distance estimation value of the signal source with the current criterion in real time: if , it is judged that the signal source is in the near-field, and the ESG modeling and joint angle-distance estimation strategy should be enabled; if , it can be approximated as a far-field condition, and the MUSIC type angle estimation method is used to reduce the computational complexity.

[0034] The Rayleigh distance criterion has a closed-form analytical expression and a monotonic dependence on , so it has real-time calculation capability. In various operating scenarios, it can assist in automatically switching modeling methods, adjusting estimation algorithms, and evaluating resolution performance, becoming an important support module for the entire positioning system to realize field recognition and strategy scheduling closed-loop control.

[0035] S2. The fluid antenna system switches between the compressed aperture configuration and the expanded aperture configuration according to real-time task requirements. The fluid antenna system integrates a software-controllable mechanical scaling execution mechanism, which can dynamically adjust the distance between adjacent units of the array, realize the rapid reconstruction and switching of the array aperture structure, and record the current configuration identifier and the corresponding time stamp at each configuration switching, ensuring the cross-configuration coherence of the collected data within the same task period.

[0036] In each configuration mode, the fluid antenna system continuously observes with a uniform snapshot number , and constructs the observation data matrix under the compressed aperture and the observation data matrix under the expanded aperture, respectively. Wherein, This indicates the first [number] under the compressed configuration. The array receive vector acquired in a snapshot at time. Indicates the first under extended configuration The array receive vectors are acquired via snapshots. To ensure amplitude and phase consistency, unified time-frequency synchronization and gain calibration are performed before data acquisition to ensure the amplitude-phase continuity and dynamic range stability of the fluid antenna system before and after configuration switching.

[0037] The fluid antenna system calculates the corresponding sample covariance matrix based on two sets of data matrices. and This method characterizes the spatial statistical features under compressed and extended aperture configurations, providing a unified input basis for subsequent cross-aperture fusion positioning algorithms. By comparing and fusing the covariance characteristics under different configurations, the fluid antenna system can improve the accuracy, robustness, and multi-task adaptability of source parameter estimation while maintaining resolution and robustness.

[0038] S3. In the compressed aperture configuration, the fluid antenna system performs an initial estimation of the incident direction of the signal source (DOA) based on the multi-snapshot observation data acquired in step S2. Since compact arrays are susceptible to interference from physical effects such as element mutual coupling and amplitude-phase mismatch in practical applications, directly using an ideal array model will lead to deviations or distortions in the direction estimation. To balance modeling accuracy and computational efficiency, this invention introduces a lightweight modeling strategy, combining a Toeplitz structure mutual coupling matrix with a far-field array manifold approximation, and incorporating symmetric central subarray extraction and amplitude-phase renormalization techniques to improve the stability and robustness of the direction estimation.

[0039] Considering the small array aperture under compressed configuration, most signal sources meet the far-field condition. Therefore, the spherical wave direction vector can be approximated as a far-field manifold. Based on this, a mutual coupling matrix is ​​introduced. Describe the translation-invariant mutual coupling behavior between array elements in the array, and construct the first... The receiving model under each time snapshot is: ; in, For the far-field array manifold under compressed configuration, Let be the set of incident angles of the signal source to be estimated. Represents the far-field direction vector. To observe the noise vector, This is the effective source signal vector.

[0040] To improve modeling robustness and suppress uncertainty in edge elements, a selection matrix is ​​constructed. Extracting central subarray observations , calculate the sample covariance matrix , and perform eigen-decomposition to extract the signal subspace and noise subspace: ; wherein, represents the eigenvector matrix of the signal subspace under the compressed configuration, represents the eigenvector matrix of the noise subspace under the compressed configuration, and are the corresponding eigenvalue diagonal matrices.

[0041] On this basis, the classic MUSIC method is used to construct the direction spectrum function under the compressed configuration: ; wherein, is the center row of the far-field direction vector .

[0042] Finally, the angles corresponding to the largest amplitude peaks in the direction spectrum function are selected to form the initial DOA estimation set under the compressed aperture, and the direction estimation result will be used as the initial value of the direction search in the extended aperture stage in step S4, and is further optimized in combination with the distance dimension estimation to realize high-precision two-dimensional source positioning.

[0043] S4. Under the condition of the extended aperture configuration , the physical size of the array is increased and the inter-element spacing is pulled apart, and the mutual coupling effect between the elements is significantly weakened, so that the mutual coupling modeling correction term is no longer introduced in this step. At this time, the Rayleigh distance of the array is increased, and the spherical wave characteristics are obvious. Therefore, the fluid antenna system adopts a near-far-field unified modeling method based on the ESG model in this stage to jointly and finely estimate the incident angle and distance of the target source. Based on the data collected in step S2, the sample covariance matrix is calculated: ; wherein, represents the eigenvector matrix of the signal subspace under the extended aperture configuration, is the eigenvector matrix of the noise subspace under the extended aperture configuration, is the eigenvalue diagonal matrix of the corresponding signal subspace, is the eigenvalue diagonal matrix of the corresponding noise subspace. To avoid high computational complexity caused by directly performing global search in the two-dimensional polar coordinate parameter space , the application proposes a two-stage local search strategy based on direction prior guidance for optimization estimation.

[0044] In the first stage, the rough direction estimation obtained in step S3 under the compression configuration is refined After fixing the angle for each direction as the angle anchor point, a distance spectrum function is constructed along the distance dimension: ; A one-dimensional distance spectrum peak search is performed on each angle through the distance spectrum function to obtain an initial distance estimation value under the corresponding direction On this basis, the second stage is entered, and the initial estimation parameter is taken as the center, a two-dimensional polar coordinate local window , is constructed, and a two-dimensional spatial spectrum function of joint direction-distance is constructed: ; Finally, the peak position is selected, that is, the final fine estimation result of the joint direction and distance of the kth target source.

[0045] The fluid antenna device in the application is provided with the fluid antenna system, and the signal source positioning is performed by using the source positioning method.

[0046] The effectiveness of the method proposed in the application is proved by numerical results.

[0047] In order to comprehensively verify the performance of the technical scheme of the application under different observation conditions, the experiment adopts a unified array structure and channel parameter configuration. Specifically, the number of array antennas is set to a fixed value , and the carrier wavelength is kept constant , so as to eliminate the interference of frequency difference on the estimation performance. The number of snapshots is fixed to , which is used to suppress random noise disturbance and ensure the stability and repeatability of the results. In the snapshot number dependence analysis, the signal-to-noise ratio is kept constant and fixed to , and only the value of the number of snapshots is changed, so as to quantify the influence of the amount of data on the estimation accuracy and convergence speed. The double-axis evaluation strategy effectively avoids the cross coupling between multiple parameters, and can comprehensively reflect the robustness and stability of the method of the application under various noise environments and snapshot settings.

[0048] In order to ensure the objectivity of the comparison conclusion, multiple representative comparison algorithms are set in the experiment as the performance baseline, including MILE, SDM and other classic methods. All algorithms are tested under the same array structure, noise environment and initialization condition, and the inter-element spacing is uniformly set to , to build a fair comparison platform. In terms of result presentation, for the unified estimation stage, the following key variables are defined: ACC represents the initial angle estimation result obtained in the compressed aperture stage; AAR represents the angle-distance joint fine estimation result in the extended aperture stage. The two-stage cooperation represents the multi-scale estimation mechanism in the "coarse capture-fine" process of the method.

[0049] In addition, to evaluate the approximation ability of the algorithm to the theoretical upper limit of performance, the Cramer-Rao bound (CRB) is introduced as the theoretical lower bound standard in the experiment. Among them, represents the lower limit of the estimation variance under the standard array structure , corresponds to the theoretical optimal estimation performance under the extended array structure , which is used to measure the theoretical upper limit of the long-distance perception accuracy. These two types of lower bounds provide a quantitative reference for the approximation ability of the algorithm under different array configurations, which helps to comprehensively verify the effectiveness of the algorithm from both theoretical and experimental dimensions.

[0050] To further verify the universality and adaptability of the present application in the mixed scene of near field-Fresnel zone-far field, a representative non-uniform spatial test scene is specially constructed. Four signal sources are set, and their incident angles are , and the corresponding radial distances are , respectively, covering typical near field, Fresnel transition zone and far field region, forming a composite field distribution. To ensure the fairness of the lateral comparison, the array structure and channel parameters are kept consistent during the test to ensure the repeatability of the experimental results and the reliability of the method conclusion.

[0051] In the first stage of compressed aperture configuration , the subspace spectrum function required for direction estimation is constructed according to step S3. As shown in Figure 4 , the method proposed in the present application can effectively obtain the coarse incident angle estimation results of the four signal sources. Compared with the traditional far field MUSIC algorithm as shown in Figure 5 , which does not consider mutual coupling and mixed field modeling error, it presents significant spectral peak drift, false pseudo-peak and spectral leakage problems in the near field and Fresnel zone, resulting in serious degradation of recognition performance. This stage significantly improves the direction distinguishability and spectral peak stability in the mixed field region through the "mutual coupling absorption into incident source covariance + manifold far field approximation modeling" mechanism, providing reliable initial values for subsequent estimation.

[0052] In the second stage of extended aperture configuration , the algorithm takes the incident angle initial value obtained in the first stage as the center, performs one-dimensional search along the radial distance dimension for each signal source, constructs the distance spectrum function to extract the initial distance estimation. As shown in Figure 6 andFigure 7 As shown, for the two sources in near-field and Fresnel region, the range profiles appear sharp main peaks near and respectively, which indicates that the proposed method has good near-mid-field range resolution. Figure 8 and Figure 9 The range profiles of far-field source and extreme far-field source are shown, and the spectrum peaks are still focused, without main lobe broadening or side lobe shift, which fully verifies that the ESG modeling framework constructed by the present application still has good main lobe control ability and estimation accuracy even in far-field conditions. Finally, the four spectrum peak positions obtained constitute the initial range estimates of the four sources, which are used as the search center of the subsequent two-dimensional joint refinement estimation stage, effectively reducing the search space and improving the estimation efficiency.

[0053] Figure 10 , Figure 11 , Figure 12 and Figure 13 The two-dimensional joint estimation performance of the unified ESG modeling and estimation algorithm proposed by the present application in different electromagnetic propagation regions is fully shown, which verifies its wide adaptability and accurate positioning ability in near-field, Fresnel region and far-field conditions. In the near-field scenario shown in Figure 10 , i.e. the source position is , the two-dimensional spectrum shows sharp point-like peaks in both angle and range dimensions, verifying its excellent two-dimensional resolution; Figure 11 The estimation results of the Fresnel region source are shown, when the source position is , the spectrum peak is slightly broadened in the range dimension, but the angle dimension still maintains good focusing, indicating that the fluid antenna system has natural adaptability to the intermediate region source. Figure 12 and Figure 13 correspond to the far-field source position and the extreme far-field source position respectively, the spectrum shows a horizontal stretching trend in the range direction, while the focusing characteristics in the angle direction are still clear, fully meeting the physical characteristics of the distance being indistinguishable and the angle being dominant in far-field propagation. The evolution of the spectrum peak shape from point to line clearly depicts the essential difference between near-field and far-field source space characteristics, and also reflects the regional adaptive estimation ability of the proposed method under the unified modeling framework.

[0054] Figure 14 and Figure 15 and Figure 17Further, the estimation performance of different algorithms under various SNR conditions is compared in a typical mixed source scenario. The comparative algorithms such as MILE and SDM generally have performance collapse under low SNR conditions, the estimation error deviates significantly from the theoretical bound, and it is also difficult to effectively process mixed field sources in the medium and high SNR interval. The proposed algorithm does not need to pre-classify the source near-field and far-field types, and realizes global robust estimation based on a unified ESG modeling strategy. It maintains stable and reliable estimation accuracy in the full SNR interval. Especially under medium and high SNR conditions, the estimation error quickly approaches the CRB lower bound, showing outstanding noise resistance and estimation algorithm robustness.

[0055] Figure 16 and Figure 18 The error convergence trend of each algorithm under different source conditions is presented with the increase of snapshots under fixed SNR conditions. The results show that the traditional method generally fails to converge when the sample size is insufficient or the source type is unknown. However, the proposed method AAR can achieve positioning accuracy close to the theoretical lower bound with a small number of snapshots, and has consistent and robust performance in near-field, Fresnel zone and far-field, which is significantly better than the existing comparative algorithms. The results show that the proposed method has good sample utilization efficiency and generalization ability, and is suitable for practical deployment in complex dynamic scenarios, and has high engineering practical value and promotion potential.

[0056] For those skilled in the art, various corresponding changes and modifications can be made to the above-described technical solutions and concepts, and all these changes and modifications should belong to the protection scope of the present application claims.

Claims

1. A method for locating a fluid antenna source based on dynamic aperture reconfiguration, characterized in that, Includes the following steps: S1. Construct a fluid antenna system with array scalability and dynamic reconfiguration capabilities. By controlling the scaling of array element spacing, achieve adaptive switching between compressed aperture and expanded aperture configurations. Establish a unified modeling and processing framework to support the localization of near-field and far-field mixed signal sources. Accurate spatial geometric modeling and response compensation are performed on the spatial coherence characteristics under different array configurations. S2. The fluid antenna system switches between compressed aperture configuration and expanded aperture configuration according to real-time task requirements; the fluid antenna system integrates a software-controllable mechanical scaling execution mechanism to dynamically adjust the spacing between adjacent array elements, thereby achieving rapid reconstruction and switching of the array aperture structure; during each configuration switch, the fluid antenna system automatically records the current configuration identifier and corresponding timestamp to ensure cross-configuration coherence of data collected within the same task cycle; In each configuration mode, the fluid antenna system uses a uniform number of snapshots. Continuous observations were conducted, and observation data matrices were constructed for each compressed aperture. Observation data matrix under extended aperture ;in, This indicates the first [number] under the compressed configuration. The array receive vector acquired in a snapshot at time. Indicates the first under extended configuration The array receive vector acquired in a snapshot at time; The fluid antenna system calculates the corresponding sample covariance matrix based on two sets of data matrices. and To characterize the spatial statistical features under the configurations of compressed aperture and expanded aperture; S3. In the compressed aperture configuration, the fluid antenna system performs an initial estimation of the incident direction of the signal source (DOA) based on the multiple snapshot observation data acquired in step S2; Mutual coupling matrix Describe the translation-invariant mutual coupling behavior between array elements in the array, and construct the first... The receiving model under each time snapshot is: ； in, For the far-field array manifold under compressed configuration, Let be the set of incident angles of the signal source to be estimated. Represents the far-field direction vector. To observe the noise vector, The effective source signal vector; Construct the selection matrix Extracting central subarray observations Calculate the sample covariance matrix And perform feature decomposition to extract the signal subspace and noise subspace: ； in, This represents the eigenvector matrix of the signal subspace under the compressed configuration. This represents the eigenvector matrix of the noise subspace under the compressed configuration. and This is the corresponding eigenvalue diagonal matrix; Constructing the directional spectrum function under compressed configuration: ； in, Far-field direction vector center OK; Select the function with the largest amplitude from the directional spectrum function. The angles corresponding to each peak constitute the initial DOA estimation set under compressed aperture. ; S4. In the expanded aperture configuration Under the condition of data collected in step S2 Calculate the sample covariance matrix: ； in, This represents the eigenvector matrix of the signal subspace under the extended aperture configuration. To configure the eigenvector matrix of the noise subspace under the extended aperture, Let be the diagonal matrix of eigenvalues ​​of the corresponding signal subspace. This is the diagonal matrix of eigenvalues ​​for the corresponding noise subspace; An optimization estimation is performed using a two-stage local search strategy guided by directional priors: In the first stage, the coarse orientation estimate obtained under the compressed configuration in step S3 is... Using angular anchor points, after fixing the angle for each direction, the distance spectrum function is constructed along the distance dimension: ； By performing a one-dimensional distance spectrum peak search on each angle using the aforementioned distance spectrum function, the initial distance estimate for the corresponding direction is obtained. ; Based on this, we proceed to the second stage, which involves preliminary parameter estimation. Centered on a local window in two-dimensional polar coordinates , Within, construct a two-dimensional spatial spectral function that combines direction and distance: ； Finally, the peak position was selected. This is the final, detailed estimation result.

2. The fluid antenna source localization method based on aperture dynamic reconfiguration according to claim 1, characterized in that, In step S1, the array geometry of the fluid antenna system consists of M antenna elements arranged in a uniform linear array along a linear array direction; the spacing between adjacent antenna elements... Subject to scaling factor control, The initial baseline spacing of the array. This indicates a compressed configuration. Indicates an extended configuration; the fluid antenna system is configured... The total array aperture is expressed as follows: 。 3. The fluid antenna source localization method based on aperture dynamic reconfiguration according to claim 2, characterized in that, In step S1, precise spatial geometric modeling specifically involves: any first The positions of the signal sources are represented in polar coordinates. , and the first in the array The propagation distance between each array element is denoted as . ; Constructing spatial response vectors under an accurate spatial geometric model , is used to represent the phase response characteristics of a signal from the source point to each unit of the array. The elements of the phase response characteristics are determined by the corresponding propagation distance and signal wavelength parameters.

4. The fluid antenna source localization method based on aperture dynamic reconfiguration according to claim 3, characterized in that, In step S1, the Rayleigh distance criterion index is defined. : ； in, Rayleigh distance representing the baseline aperture, The operating wavelength is indicated; the distance criterion index serves as the core basis for model selection, used to determine the spatial propagation mechanism between the signal source and the array in real time; thus serving as the physical boundary for switching between near-field and far-field modeling methods.

5. The fluid antenna source localization method based on aperture dynamic reconfiguration according to claim 4, characterized in that, During operation, the fluid antenna system automatically invokes the distance criterion index to estimate the distance to the signal source after each array structure switch. Compared with current criteria Perform real-time comparison: If If the source is determined to be in the near field, a precise spatial geometry modeling and joint angle-distance estimation strategy is employed; if If the condition is true, it is determined to be a far-field condition, and the angle estimation method of the MUSIC type is used for calculation.

6. A fluid antenna device, characterized in that, The fluid antenna system is provided, and the source localization method as described in any one of claims 1-5 is used to locate the signal source.

Images