A fluid antenna source positioning method and device based on aperture dynamic reconfiguration

Abstract

The application 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 application is proposed to solve the problems of limited positioning accuracy and insufficient environmental adaptability caused by fixed array structure and single source field model. The steps are as follows: S1. Constructing a fluid antenna system S-FAS with adaptive aperture adjustment capability, and establishing a unified modeling framework based on accurate spatial geometry ESG; S2. Switching between compressed and expanded apertures according to task requirements, collecting observation data and generating statistics; S3. Estimating the direction under the compressed aperture and suppressing mutual coupling and amplitude and phase errors; S4. Jointly estimating the target direction and distance under the expanded aperture with the initial value as the anchor point to realize high-precision positioning. The application can work stably in a near-field, Fresnel zone and far-field multi-source mixed environment without distinguishing source types or isolating signals, has high precision, low complexity and strong adaptability, and is suitable for intelligent sensing and low-altitude communication scenes.

Description

Technical Field

[0001] This invention relates to the field of wireless communication and array signal processing technology, and in particular to a method and apparatus for locating fluid antenna sources based on aperture dynamic reconfiguration. Background Technology

[0002] Direction of arrival (DOA) estimation and source localization are key technologies in array signal processing, widely used in radar monitoring, unmanned system navigation, and electromagnetic spectrum sensing. With the emergence of new demands such as sixth-generation mobile communication and integrated sensing, systems are placing higher requirements on high-resolution, robust, and low-latency positioning capabilities in complex environments.

[0003] Existing technologies largely rely on fixed-structure arrays, employing far-field plane wave or near-field spherical wave models. However, real-world communication and sensing tasks frequently involve multi-source mixed fields, with target distance distributions spanning the near field, Fresnel zone, and far field, leading to significant wavefront mismatch and estimation bias. Furthermore, fixed-aperture structures make it difficult to flexibly adjust array geometry, hindering the balance between angular resolution, mutual coupling suppression, and modeling accuracy, thus limiting the system's performance ceiling in dynamic and complex scenarios.

[0004] To address the challenges of limited source localization accuracy and insufficient array structural rigidity, emerging fluid antenna systems have proposed a novel array paradigm with reconfigurable capabilities in recent years. This system encapsulates conductive liquid within a controllable medium and combines it with an electromagnetic drive mechanism to achieve rapid migration and dynamic reconfiguration of array elements in physical space. This overcomes the geometric rigidity limitations of traditional arrays, offering advantages such as high structural flexibility, fast response speed, and strong adaptability. This provides a new physical foundation for constructing adaptive array localization systems for complex fields, but also places higher demands on array architecture and spatial modeling methods. Therefore, it is urgent to propose a novel reconfigurable array system that integrates a fluid antenna architecture and establish a unified spatial modeling and source localization method to achieve high-precision modeling, flexible structural control, and engineering feasibility in multi-scale mixed fields such as near-field, Fresnel region, and far-field, overcoming the performance bottlenecks and adaptability limitations of traditional arrays in source localization in complex scenarios. Summary of the Invention

[0005] The purpose of this invention is to provide a fluid antenna source localization method and apparatus based on aperture dynamic reconfiguration, so as to solve the problems existing in the prior art.

[0006] The fluid antenna source localization method based on aperture dynamic reconfiguration described in this invention includes the following steps:

[0007] 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. Perform precise spatial geometric modeling and response compensation on the spatial coherence characteristics under different array configurations.

[0008] 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;

[0009] 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;

[0010] 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;

[0011] 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;

[0012] 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:

[0013] ;

[0014] 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;

[0015] 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:

[0016] ;

[0017] 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;

[0018] Constructing the directional spectrum function under compressed configuration:

[0019] ;

[0020] in, Far-field direction vector center OK;

[0021] 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. ;

[0022] S4. In the expanded aperture configuration Under the condition of data collected in step S2 Calculate the sample covariance matrix:

[0023] ;

[0024] 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;

[0025] An optimization estimation is performed using a two-stage local search strategy guided by directional priors:

[0026] 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:

[0027] ;

[0028] 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. ;

[0029] 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:

[0030] ;

[0031] Finally, the peak position was selected. This is the final, detailed estimation result.

[0032] The fluid antenna device described in this invention includes the fluid antenna system and utilizes the source localization method to locate the signal source.

[0033] The fluid antenna source localization method and apparatus based on dynamic aperture reconfiguration described in this invention have the advantage of fully releasing the reconfigurable potential of the fluid antenna elements in physical layout by taking adaptive reconstruction of the array element spatial configuration as the core, and constructing a dynamic aperture array architecture with dual-state switching capabilities of compression and expansion. In the initial stage, the system controls the liquid antenna array elements to converge in a local area to form a high-density compressed aperture mode. Using the standard MUSIC algorithm, the target direction is quickly and coarsely estimated under low sample and low complexity conditions, completing the direction pre-positioning in the compressed aperture stage. Subsequently, the system drives the array element space to expand and enters the expanded aperture stage in conjunction with ESG. A higher angular resolution is obtained through the spectral peak refinement mechanism, and then the target distance and angle are calculated in conjunction to achieve high-precision reconstruction of the spatial position. This mechanism cleverly integrates near-field and far-field propagation characteristics, has signal independence and frequency band adaptability, and can adapt to different modulation waveforms and electromagnetic environments, showing good positioning consistency in various propagation regions such as near-field, Fresnel region, and far-field. Experimental results show that the proposed scheme exhibits superior robustness and resolution under typical low-altitude complex environments, multi-source interference backgrounds, and low-sample conditions. It is particularly suitable for high-dynamic communication and sensing fusion scenarios such as urban low-altitude security, electromagnetic spectrum intelligent sensing, illegal signal source monitoring, and dynamic frequency scheduling. It has good system compatibility and broad prospects for engineering applications and industrialization. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the overall process of the fluid antenna source localization method described in this invention.

[0035] Figure 2A schematic diagram of the initial orientation estimation process under the compression aperture configuration.

[0036] Figure 3 A schematic diagram of the joint angle-distance fine estimation process under extended aperture configuration.

[0037] Figure 4 This is the initial angle spectrum curve of the compressed aperture MUSIC algorithm.

[0038] Figure 5 This is the angle spectrum curve of the traditional far-field MUSIC algorithm.

[0039] Figure 6 To estimate the angle in the initial stage The following is a one-dimensional distance spectrum curve.

[0040] Figure 7 To estimate the angle in the initial stage The following is a one-dimensional distance spectrum curve.

[0041] Figure 8 To estimate the angle in the initial stage The following is a one-dimensional distance spectrum curve.

[0042] Figure 9 To estimate the angle in the initial stage The following is a one-dimensional distance spectrum curve.

[0043] Figure 10 For source 1 Two-dimensional angle-distance joint spectral distribution map.

[0044] Figure 11 Source 2 Two-dimensional angle-distance joint spectral distribution map.

[0045] Figure 12 Source 3 Two-dimensional angle-distance joint spectral distribution map.

[0046] Figure 13 Source 4 Two-dimensional angle-distance joint spectral distribution map.

[0047] 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.

[0048] 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.

[0049] 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.

[0050] Figure 17 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

[0051] 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:

[0052] 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.

[0053] 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 follows:

[0054] ;

[0055] This structure provides a controllable physical basis for subsequent spatial resolution adjustment and modeling.

[0056] 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.

[0057] 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.

[0058] 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.

[0059] 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. :

[0060] ;

[0061] in, Rayleigh distance representing the baseline aperture, The operating wavelength is indicated. The Rayleigh distance criterion serves as the core basis for model selection, used to determine in real-time the spatial propagation mechanism between the signal source and the array—whether it's a spherical wave or a plane wave—thus acting as the physical boundary for switching between near-field and far-field modeling methods. This fully demonstrates the real-time control effect of array aperture adjustment on near-field and far-field boundaries: when... When the aperture is reduced, i.e., the aperture configuration is compressed, the effective aperture of the array shrinks, resulting in As the area decreases, the near-field region shrinks accordingly; when When the aperture is increased, i.e., the aperture configuration is expanded, the increase in aperture leads to The range increases significantly, the far-field boundary recedes, and some mid-range sources may cross the boundary into the near field.

[0062] During actual operation, after each array structure switch, the fluid antenna system will automatically invoke the Rayleigh distance criterion to estimate the distance to the signal source. Compared with current criteria Perform real-time comparison: If If the information source is determined to be in the near field, an ESG modeling and joint angle-distance estimation strategy should be employed; if If the condition is far-field, then a MUSIC-type angle estimation method can be used to reduce computational complexity.

[0063] The Rayleigh distance criterion index has a closed analytical expression and a relation to... Due to its monotonic dependence, it possesses real-time computing capabilities. Under various operating scenarios, it can assist in automatically switching modeling methods, adjusting estimation algorithms, and evaluating resolution performance, becoming a crucial supporting module for the entire positioning system to achieve closed-loop control of field identification and strategy scheduling.

[0064] 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 mechanism, which can dynamically adjust the spacing between adjacent array elements, enabling 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, ensuring cross-configuration coherence of data collected within the same task cycle.

[0065] 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 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.

[0066] 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.

[0067] 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.

[0068] 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:

[0069] ;

[0070] 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.

[0071] 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 feature decomposition to extract the signal subspace and noise subspace:

[0072] ;

[0073] 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.

[0074] Based on this, the directional spectrum function under the compressed configuration is constructed using the classic MUSIC method:

[0075] ;

[0076] in, Far-field direction vector center OK.

[0077] Finally, the function with the largest amplitude is selected from the directional spectrum functions. The angles corresponding to each peak constitute the initial DOA estimation set under compressed aperture. The direction estimation result will be used as the initial value for the direction search in step S4, and further optimized in conjunction with the distance dimension estimation to achieve high-precision two-dimensional source localization.

[0078] S4. In the expanded aperture configuration Under these conditions, as the physical size of the array increases and the spacing between array elements widens, the mutual coupling effect between elements is significantly weakened. Therefore, this step no longer introduces a mutual coupling modeling correction term. At this point, the Rayleigh distance of the array... As the amplitude increases, it exhibits obvious spherical wave characteristics. Therefore, the fluid antenna system employs a unified near-field and far-field modeling approach based on the ESG model at this stage to jointly and finely estimate the incident angle and range of the target source. This is based on the data collected in step S2. Calculate the sample covariance matrix:

[0079] ;

[0080] 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. is the diagonal matrix of eigenvalues ​​for the corresponding noise subspace.

[0081] To avoid in two-dimensional polar coordinate parameter space To address the high computational complexity caused by directly performing a global search, this invention proposes a two-stage local search strategy guided by directional priors for optimization estimation.

[0082] 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:

[0083] ;

[0084] 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, the second stage begins with 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:

[0085] ;

[0086] Finally, the peak position was selected. This is the final, detailed estimate of the joint direction and distance of the k-th target source.

[0087] The fluid antenna device described in this invention includes the fluid antenna system and utilizes the source localization method to locate the signal source.

[0088] The effectiveness of the method proposed in this invention will be demonstrated through numerical results below.

[0089] To comprehensively verify the performance of the technical solution of this invention under different observation conditions, the experiment adopted a uniform array structure and channel parameter configuration. Specifically, the number of array antennas was set to a fixed value. The carrier wavelength remains constant. This is to eliminate the interference of frequency differences on estimation performance. The number of snapshots is fixed at [value missing]. This is used to suppress random noise disturbances and ensure the stability and repeatability of the results. In the snapshot-number dependency analysis, the signal-to-noise ratio is kept constant at a fixed value. Only change the number of snapshots The value of is used to quantify the impact of data volume on estimation accuracy and convergence speed. This dual-axis evaluation strategy effectively avoids cross-coupling between multiple parameters and can comprehensively reflect the robustness and stability of the method under various noise environments and snapshot settings.

[0090] To ensure the objectivity of the comparison results, the experiment set up several representative comparison algorithms as performance baselines, including classic methods such as MILE and SDM. All algorithms were tested under the same array structure, noise environment, and initialization conditions, with the element spacing uniformly set to [value missing]. To establish a fair comparison platform, the following key variables are defined to unify the terminology used in the estimation stage: ACC represents the initial angle estimation result obtained in the compressed aperture stage; AAR represents the joint angle-distance fine estimation result performed in the expanded aperture stage. The two stages synergistically characterize the multi-scale estimation mechanism under the "coarse-fine" process of this method.

[0091] Furthermore, to evaluate the algorithm's ability to approximate the theoretical performance upper limit, the experiment introduced the Cramé-Rao bound (CRB) as a theoretical lower bound standard. Among these, In the standard array structure The lower bound of the estimated variance Corresponding extended array structure The theoretical optimal estimation performance is used to measure the theoretical upper limit of the accuracy of long-distance sensing. These two types of lower bounds provide a quantitative reference for the approximation capability of the algorithm under different array configurations, which helps to comprehensively verify the effectiveness of the algorithm from both theoretical and experimental perspectives.

[0092] To further verify the versatility and adaptability of this invention in near-field-Fresnel region-far-field mixed scenarios, a representative non-uniform spatial test scenario was constructed. Four signal sources were set up with incident angles of... The corresponding radial distances are respectively The arrays cover typical near-field, Fresnel transition, and far-field regions, forming a composite field distribution. To ensure fairness in the horizontal comparison, the array structure and channel parameters are kept consistent during the testing process, ensuring the repeatability of the experimental results and the reliability of the methodological conclusions.

[0093] First stage compression orifice configuration In step S3, the subspace spectral function required for direction estimation is constructed. For example... Figure 4 As shown, the method proposed in this invention can effectively obtain coarse incident angle estimation results from four information sources. Compared to Figure 5 The traditional far-field MUSIC algorithm, as shown, exhibits significant peak drift, spurious peaks, and spectral leakage in the near field and Fresnel region due to its failure to consider mutual coupling and mixed-field modeling errors, resulting in severe degradation of identification performance. This stage, through the mechanism of "mutually coupled absorption incident source covariance + manifold far-field approximation modeling," significantly improves directional discriminability and peak stability in the mixed-field region, providing reliable initial values ​​for subsequent estimations.

[0094] Second-stage expanded aperture configuration In the algorithm, the initial value of the incident angle obtained in the first stage is used... Centered on a point, a one-dimensional search is performed along the radial distance dimension for each signal source to construct a distance spectrum function to extract the initial distance estimate. For example... Figure 6 and Figure 7 As shown, for two sources in the near field and Fresnel region, their range spectra are respectively in... and The presence of a sharp main peak nearby indicates that this method has good near-mid distance resolution capability. Figure 8 and Figure 9 Demonstrating far-field sources With far-field sources The distance spectrum results show that the spectral peaks remain focused and there is no main lobe broadening or side lobe drift, fully verifying that the ESG modeling framework constructed in this invention still has good main lobe control capability and estimation accuracy even under far-field conditions. Finally, the obtained four spectral peak positions constitute the initial distance estimates of the four sources, which serve as the search centers for the subsequent two-dimensional joint refinement estimation stage, effectively reducing the search space and improving estimation efficiency.

[0095] Figure 10 , Figure 11 , Figure 12 as well as Figure 13 This paper comprehensively demonstrates the two-dimensional joint estimation performance of the unified ESG modeling and estimation algorithm proposed in this invention under different electromagnetic propagation regions, and verifies its broad adaptability and accurate positioning capability under near-field, Fresnel region, and far-field conditions. Figure 10 In the near-field scene shown, the source location is... The two-dimensional spectrum exhibits sharp point peaks in both the angle and distance dimensions, verifying its excellent two-dimensional resolution. Figure 11 The estimation results of Fresnel zone sources are presented, when the source location is... At that time, the spectral peaks broadened slightly in the distance dimension, but remained well focused in the angular dimension, indicating that the fluid antenna system has a natural adaptability to field sources in the intermediate region. Figure 12 and Figure 13 The corresponding far-field source locations are: The location of the far-field source is In the scene, the spectrum exhibits a horizontal stretching trend in the distance direction, while the focusing characteristics in the angular direction remain clear, fully conforming to the physical characteristics of far-field propagation where distance is indistinguishable and angle dominates. The evolution of the spectral peak shape from point to line clearly depicts the essential differences in the spatial characteristics of near-field and far-field sources, and also demonstrates the regional adaptive estimation capability of the method of this invention within a unified modeling framework.

[0096] Figure 14 and Figure 15 Furthermore, the estimation performance of different algorithms under various SNR conditions was compared in typical mixed-source scenarios. Comparative algorithms such as MILE and SDM generally exhibited performance degradation under low SNR conditions, with estimation errors significantly deviating from the theoretical limit, and they also struggled to effectively handle mixed-field sources in the medium-to-high SNR range. In contrast, the proposed algorithm, without requiring pre-classification of near-field and far-field source types, achieved robust estimation across the entire domain based on a unified ESG modeling strategy, maintaining stable and reliable estimation accuracy across the entire SNR range. Especially under medium-to-high SNR conditions, its estimation error rapidly approached the lower bound of the CRB theory, demonstrating outstanding noise resistance and robustness of the estimation algorithm.

[0097] Figure 16 and Figure 17 The results show the error convergence trend of each algorithm under different source conditions as the number of snapshots increases under a fixed signal-to-noise ratio. The results indicate that traditional methods generally fail to converge when the number of samples is insufficient or the source type is unknown. In contrast, the proposed AAR method achieves positioning accuracy close to the theoretical lower bound with only a few snapshots, and exhibits consistent and robust performance in the near field, Fresnel zone, and far field, significantly outperforming existing comparative algorithms. These results demonstrate that the proposed method possesses good sample utilization efficiency and generalization ability, is suitable for practical deployment needs in complex dynamic scenarios, and has high engineering practical value and promotion potential.

[0098] For those skilled in the art, various other corresponding changes and modifications can be made based on the technical solutions and concepts described above, and all such changes and modifications should fall within the protection scope of the claims of this invention.

Claims

1. A fluid antenna source localization method based on aperture dynamic 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. Perform precise spatial geometric modeling and response compensation on the spatial coherence characteristics under different array configurations. Precise spatial geometric modeling specifically refers to: 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. It is the scaling factor of the spatial response vector; 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 corresponding sample covariance matrices based on two groups of data matrices respectively With To characterize the spatial statistical characteristics under the configurations of compressed aperture and extended 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 The mutual coupling behavior between elements in an array with shift-invariance property is described, and the receiving model of the array at the time snapshot is constructed as ； wherein, is the far-field array manifold under compressed configuration, is the set of incidence angles of the signal sources to be estimated, denotes the far-field direction vector, is the observation noise vector, is the effective source signal vector, denotes the array scaling factor under compressed aperture configuration, M is the total number of antenna elements; Constructing selection matrix Extracting central sub-matrix observation , compute sample covariance matrix and perform eigen-decomposition to extract signal subspace and noise subspace: ； wherein denotes the eigenvector matrix of the signal subspace under the compressed configuration, denotes the eigenvector matrix of the noise subspace under the compressed configuration, and is the diagonal matrix of the corresponding eigenvalues. Constructing the directional spectrum function under compressed configuration: ； wherein is a far field direction vector center of row; 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 extended aperture configuration Based on the data acquired in step S2 , the sample covariance matrix is computed: ； wherein, denotes a matrix of eigenvectors of the signal subspace under extended aperture configuration, denotes a matrix of eigenvectors of the noise subspace under extended aperture configuration, denotes a diagonal matrix of eigenvalues of the corresponding signal subspace, denotes a diagonal matrix of eigenvalues of 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: ； performing one-dimensional range profile peak search on each angle by the range profile function to obtain initial range estimation value under corresponding direction ; On this basis, the second stage is entered, with the initial parameters as a center, a two-dimensional polar local window , is constructed within the joint direction-distance two-dimensional spatial spectrum function: ； Finally, the peak position is selected as the final fine estimation result.

2. The method of claim 1, wherein, 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 method of claim 2, wherein, In said step S1, the Rayleigh distance criterion indicator is defined : ； wherein, the Rayleigh distance representing the baseline aperture, represents the working wavelength; the Rayleigh distance criterion index serves as the core basis of modeling selection, which is used to judge the spatial propagation mechanism between the signal source and the array in real time; thus serving as the physical boundary of the near-field / far-field modeling mode switching.

4. The method of claim 3, wherein, During operation, the fluid antenna system automatically invokes the Rayleigh 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.

5. A fluid antenna device, characterized by The fluid antenna system is provided, and the signal source is located using the fluid antenna source localization method based on aperture dynamic reconfiguration as described in any one of claims 1-4.

Images