A method and system for monitoring and tracking a moving target signal by unmanned aerial vehicle

Abstract

The application discloses a kind of unmanned aerial vehicle target signal tracking monitoring method and system, belong to radio frequency spectrum sensing and passive direction finding positioning technical field, including the following contents: through the synchronous acquisition of multiple element antenna array in environment radar, communication equipment radiated radio frequency signal, pre-processing obtains multiple channel intermediate frequency digital sampling signal;Capture effective radiation signal and output signal frequency domain parameter;Multi-channel angle measurement data are weighted and fused to obtain accurate DOA data;Based on multi-channel DOA direction finding line, cross solution is carried out, and target original latitude and longitude point sequence is generated;Candidate effective point set is obtained using FCM fuzzy clustering;In candidate effective point set, based on Mahalanobis distance wave door, the pairing processing of point and existing track is completed;Smooth continuous target time sequence track is output after filtering iteration;Data storage and track visualization plotting are completed.The application uses the above method, guarantees the continuity and stability of moving target track tracking under passive monitoring scene.

Description

Technical Field

[0001] This invention relates to the technical field of radio spectrum sensing and passive direction finding and positioning, and in particular to a method and system for tracking and monitoring moving target signals carried by an unmanned aerial vehicle (UAV). Background Technology

[0002] With the widespread adoption of lightweight UAV electromagnetic situational awareness technology, real-time monitoring, precise direction finding and positioning, and continuous moving target tracking of radar and communication signals radiated by various devices in airspace, land, and sea environments, relying on the passive receiving system of UAVs, has become a core method for radiation source reconnaissance and monitoring in complex electromagnetic environments. Passive monitoring, with its advantages of no active radiation, strong concealment, and flexible deployment, is suitable for UAV-borne operational scenarios. However, existing UAV-borne electromagnetic monitoring technologies suffer from numerous core technical bottlenecks, severely limiting monitoring accuracy and tracking stability, making it difficult to meet actual operational needs.

[0003] First, existing UAV-borne signal processing technologies are poorly adapted to complex electromagnetic environments. Traditional clutter suppression and signal detection methods mostly use fixed noise reduction thresholds and uniform detection thresholds across the entire domain, which cannot distinguish between near-field and far-field clutter and are difficult to adapt to the distribution characteristics of different types of clutter on the ground and sea surface. This leads to frequent false alarms and missed alarms in complex scenarios, and the accuracy and reliability of effective radiated signal acquisition are extremely low, causing distortion of monitoring data from the source and seriously affecting the accuracy of subsequent positioning and tracking.

[0004] Secondly, existing direction finding and positioning technologies cannot simultaneously meet the requirements of airborne real-time performance and high precision. The angle measurement accuracy of a single interferometer is insufficient, and the computational cost of the full-domain MUSIC algorithm is high and prone to generating false peak interference, making it difficult to adapt to the high-speed mobile operation scenarios of UAVs. Moreover, most technologies do not perform multi-channel amplitude and phase consistency calibration, and channel hardware errors are propagated step by step. The generated original points are mixed with a large number of clutter false alarm points, and there is a lack of intelligent point filtering mechanism. False points directly participate in the trajectory calculation, resulting in large target positioning deviations and frequent false targets, making it impossible to achieve accurate positioning of real targets.

[0005] Most critically, existing moving target tracking technologies have fatal flaws: passive monitoring signals are highly susceptible to terrain obstruction, sea clutter interference, and multipath fading, resulting in intermittent frame breaks and signal flickering. Traditional technologies lack a scientific mechanism for determining track viability, easily leading to problems such as track breaks despite the target not disappearing, frequent ID jumps, and misjudgments of newly emerging targets, resulting in extremely poor target tracking continuity. Furthermore, traditional filtering algorithms have fixed parameters and cannot adapt to changes in target maneuvering or stable motion, exhibiting severe lag in tracking high-speed, variable-speed maneuvers, leading to prominent track jitter and deviation issues, making stable and continuous tracking difficult.

[0006] In summary, existing UAV-borne electromagnetic monitoring technologies generally suffer from low signal processing architecture integration, poor adaptability to complex clutter, and difficulty in balancing direction finding accuracy and airborne real-time performance. In particular, they lack adaptable track management and filtering optimization mechanisms for intermittent frame breaks in passive monitoring signals and target maneuvering scenarios, which easily leads to many defects such as track breaks, target tracking lag, and poor positioning accuracy. Overall, their monitoring accuracy, anti-interference ability, and tracking stability are insufficient, making it difficult to meet the high-precision, continuous, and stable monitoring and tracking requirements of UAV-borne moving targets in complex land and sea electromagnetic environments. Summary of the Invention

[0007] The purpose of this invention is to provide a method and system for tracking and monitoring moving target signals carried by unmanned aerial vehicles (UAVs), which improves the reliability of acquiring effective signals in complex clutter environments and ensures the continuity and stability of moving target trajectory tracking in passive monitoring scenarios.

[0008] To achieve the above objectives, the present invention provides a method for tracking and monitoring the signal of a moving target carried by an unmanned aerial vehicle (UAV), comprising the following steps: S1. The radio frequency signals radiated by radar and communication equipment in the environment are synchronously collected by a multi-element antenna array, and multi-channel intermediate frequency digital sampling signals are obtained after preprocessing. S2. Adaptive dynamic threshold lifting wavelet layered noise reduction is used to remove clutter from each intermediate frequency signal. The full frequency band spectrum is divided into near clutter region and far clutter region. The OS-CFAR detection threshold is calculated independently for each region. The decision is made by combining the multi-channel detection results with the majority channel consistent judgment to capture the effective radiated signal and output the signal frequency domain parameters. S3. The direction of arrival (DOA) is roughly calculated using an interferometer and the local search range of MUSIC is locked. Within the limited range, the angle measurement is finely corrected using the MUSIC spatial spectrum. The multi-channel angle measurement data is weighted and fused to obtain accurate DOA data. S4. Based on the multi-channel DOA direction finding line, cross calculation is performed to generate the target original latitude and longitude point sequence containing clutter isolated false alarm points; S5. Use FCM fuzzy clustering to pre-screen the original points and remove isolated clutter points with membership degrees lower than a preset threshold to obtain a set of candidate valid points. S6. In the candidate valid point track set, complete the pairing process between the point track and the existing track based on the Mahalanobis distance gate; S7. Construct a six-dimensional state equation for the target that includes position, velocity, and acceleration. Adaptively adjust the extended Kalman EKF process noise matrix online based on the target's real-time acceleration and turning rate. Output a smooth and continuous target time-series trajectory after filtering and iteration. S8. Upload the target time-series trajectory and raw monitoring data to the display and control terminal to complete data storage and trajectory visualization plotting.

[0009] Preferably, in step S1, the preprocessing includes bandpass filtering, low-noise amplification, frequency mixing and downconversion, synchronous ADC analog-to-digital conversion, and multi-channel amplitude-phase consistency calibration, and finally outputs multi-channel intermediate frequency digital sampling signals in the form of synchronous I / Q orthogonal complex numbers.

[0010] Preferably, step S2 specifically includes the following steps: S21. Perform layered wavelet decomposition on each intermediate frequency digital sampling signal, and adopt an adaptive dynamic threshold lifting wavelet layered noise reduction algorithm. Configure appropriate adaptive dynamic noise reduction thresholds for ground clutter that follows a log-normal distribution and sea clutter that follows a Rayleigh distribution respectively to complete clutter removal and reconstruct clean signals. S22. Perform full-band spectrum analysis on the clean signal after noise reduction. Based on the location of the receiving device, divide the full-band spectrum into a near clutter region of less than 1km and a far clutter region of greater than or equal to 1km. Statistically count the clutter level in the two clutter regions independently and calculate the OS-CFAR signal detection threshold separately. S23. Each channel completes frequency signal detection based on the OS-CFAR detection threshold of the corresponding partition to obtain single-channel signal detection results. Combining the multi-channel detection results, a majority-channel consistent judgment fusion decision method is adopted to identify and capture real and effective radiated signals and eliminate random clutter interference signals. S24. Perform parameter calculation on the captured effective radiation signal, accurately extract and output the frequency domain parameters of the signal center frequency, occupied bandwidth, and average signal power, and provide effective data support for the subsequent accurate angle measurement process.

[0011] Preferably, step S3 specifically includes the following steps: S31. Based on the effective radiation signal frequency domain parameters and multiple intermediate frequency digital sampling signals output by S2, the target wave direction angle is quickly and roughly calculated using the interferometer algorithm by utilizing the channel phase difference characteristics of the multi-element antenna array to obtain low-precision initial angle measurement data. S32. Using the direction of arrival angle coarsely measured by the interferometer as the center, define a local search interval of ±5° for the coarsely measured azimuth angle to lock the fine search range of the MUSIC algorithm. S33. Within the limited local search interval, construct the covariance matrix for the multi-channel array signals, complete the decomposition of the signal subspace and noise subspace, and complete the fine correction of angle measurement through MUSIC spatial spectrum peak search to obtain single-channel high-precision angle measurement data. S34. The high-precision angle measurement data output independently from each channel are weighted and fused to suppress random interference and measurement errors in single channels, and the fused output is accurate DOA data with strong anti-interference and stable accuracy, providing reliable angle parameters for subsequent target positioning calculation.

[0012] Preferably, step S4 specifically includes the following steps: S41. Using the center of the multi-element antenna array as the coordinate reference point, based on the accurate DOA data output in step S3, construct multiple spatial direction-finding rays corresponding to different receiving channels to obtain multi-channel DOA direction-finding lines. S42. Real-time acquisition of the geodetic coordinates and elevation information of the local array center, establishment of a spherical geographic coordinate system, mapping of multiple spatial direction finding rays to the geodetic coordinate system, and completion of geometric cross-solution calculation using the angle differences of multiple channel direction finding lines to obtain the geodetic latitude and longitude coordinates of the radiation source target. S43. According to the fixed sampling period of the system, the position of the dynamic radiation source target is continuously calculated, and the real-time latitude and longitude points of the target are output frame by frame to generate a continuous sequence of changes in the original position of the target. S44. Retain all original point data, including isolated false alarm points caused by land-sea clutter and multipath interference, to form a target original latitude and longitude point sequence containing isolated false alarm points with clutter, for subsequent clutter screening and processing.

[0013] Preferably, step S5 specifically includes the following steps: S51. Import the target original latitude and longitude point sequence of isolated false alarm points with clutter generated in step S4, and construct all the original points into a point dataset to be clustered. S52. The FCM fuzzy clustering algorithm is used to perform iterative clustering operations on the dataset of points to be clustered, dividing the effective target clusters and clutter discrete point clusters, and calculating the cluster membership value corresponding to each original point. S53. Based on the preset membership threshold, compare the cluster membership of each point, identify and remove isolated clutter false alarm points with membership below the preset threshold, and filter out invalid discrete points. S54. Retain valid target points with high membership, complete the preliminary screening of original points, and organize them to obtain a clean set of candidate valid points, providing effective data support for subsequent point track pairing.

[0014] Preferably, step S6 specifically includes the following steps: S61. Import the set of candidate valid points and tracks that have been filtered in step S5, and at the same time retrieve the historical target track data cached locally in the system. Determine whether there are valid historical tracks at present by the track survival time difference rule. S62. If a valid historical track exists, a Mahalanobis distance gate is constructed based on the target position and velocity status information, and the Mahalanobis distance between each candidate valid track and the existing track is calculated one by one; if no valid historical track exists, the current candidate track is directly determined to be a new target track. S63. For cases with historical tracks, the calculated Mahalanobis distance is compared with the preset gate threshold to determine whether the candidate track falls within the matching gate of the corresponding track, and to distinguish between matching tracks and isolated new tracks. S64. The candidate points falling within the Mahalanobis distance gate are associated and paired with the corresponding existing tracks to achieve the matching and connection of new and old track data; the isolated points that are not matched and the newly created points without corresponding historical tracks are initialized to generate new target tracks, providing a complete data foundation for subsequent track filtering iteration processing.

[0015] Preferably, step S7 specifically includes the following steps: S71. Based on the paired associated point traces and newly generated track data, construct a six-dimensional motion state equation containing the target's two-dimensional position, two-dimensional velocity, and two-dimensional acceleration, and establish an extended Kalman EKF filter iterative model. S72. Real-time calculation of target motion state parameters, continuous monitoring of target instantaneous acceleration and turning rate, and real-time determination of target maneuver intensity; S73. Adaptively adjust the EKF process noise matrix online based on the target's maneuvering state, when the target's instantaneous acceleration is greater than 0.5 m / s². 2 When the turning speed is greater than 2° / s, the process noise matrix is ​​increased to adapt to the target's maneuvering state; when the target moves smoothly, the process noise matrix is ​​decreased to ensure the smoothness of the filtering. S74. Iterative updates are completed through adaptively adjusted extended Kalman filters to correct point observation errors and motion deviations, and output smooth, stable and continuous target time-series tracks.

[0016] Preferably, step S8 specifically includes the following steps: S81. Collect the smooth and continuous target time-series trajectory data output in step S7, and synchronously integrate the corresponding signal frequency domain parameters, accurate DOA data, and a complete set of monitoring data of target positioning points throughout the entire process to construct a complete target monitoring data package. S82. Target monitoring data packets are uploaded to the ground display and control terminal in real time via the airborne transmission link to ensure the real-time performance and integrity of data transmission; S83, the ground display and control terminal parses and processes the received target monitoring data packets, and locally stores the track data and raw monitoring data to achieve data archiving and retention, which facilitates subsequent data backtracking and analysis. S84. Based on the parsed target time-series trajectory data, the real-time dynamic visualization plotting of the target trajectory is completed on the display and control terminal interface, intuitively displaying the target's real-time position and continuous motion trajectory, and completing the entire process of UAV-borne target signal tracking and monitoring.

[0017] A UAV-borne target signal tracking and monitoring system includes a signal acquisition and preprocessing module, a clutter suppression and effective signal detection module, a high-precision DOA angle measurement module, a target point track calculation module, a point track filtering module, a point track track matching module, an adaptive track filtering module, and a data visualization, storage, and transmission module. The signal acquisition and preprocessing module is used to acquire ambient radio frequency signals through a multi-element antenna array, and sequentially complete bandpass filtering, low-noise amplification, frequency mixing and downconversion, synchronous ADC analog-to-digital conversion and multi-channel amplitude and phase calibration, and output multiple synchronous I / Q orthogonal complex intermediate frequency digital sampling signals; The clutter suppression and effective signal detection module is used to perform adaptive dynamic threshold lifting wavelet layered noise reduction on each intermediate frequency digital sampling signal. Different noise reduction thresholds are configured for ground clutter and sea clutter. The OS-CFAR detection threshold is calculated independently for near and far clutter zones. The effective radiated signal is captured by the consensus decision of most channels and the center frequency, bandwidth and power frequency domain parameters of the output signal are extracted. The high-precision DOA angle measurement module is used to roughly calculate the target direction of arrival angle through an interferometer algorithm, lock the local search range of ±5°, and complete the high-precision angle measurement correction within the limited range by combining the MUSIC algorithm. It also performs weighted fusion of multi-channel angle measurement data to output accurate DOA data. The target point trace calculation module is used to construct spatial direction finding rays based on multi-channel DOA direction finding lines, and combine the geodetic coordinate system to complete geometric cross-combined calculations to continuously generate the target original latitude and longitude trace sequence containing clutter isolated false alarm points. The point filtering module is used to iteratively cluster the original point dataset using the FCM fuzzy clustering algorithm, calculate the membership degree of each point, remove isolated clutter false alarm points with low membership degree, and filter to obtain a pure set of candidate valid points. The point track pairing module is used to effectively determine the survival status of historical tracks. It completes the association and pairing of valid points with historical tracks through the Mahalanobis distance gate, and completes track initialization for new targets without valid historical tracks. The adaptive track filtering module is used to construct the target's six-dimensional motion state equation, adaptively adjust the EKF process noise matrix according to the target's instantaneous acceleration and turning rate, correct errors through iterative filtering, and output a smooth and continuous target time-series track. The data visualization, storage, and transmission module is used to integrate monitoring data from the entire process to construct monitoring data packets, which are then uploaded to the ground display and control terminal in real time. This completes data parsing, local storage, and real-time dynamic visualization of target tracks.

[0018] Therefore, the present invention employs the above-described method and system for tracking and monitoring unmanned aerial vehicle (UAV) targets, which has the following advantages: (1) In this invention, an adaptive dynamic threshold lifting wavelet hierarchical denoising algorithm is adopted in combination with the near and far clutter partitioning OS-CFAR detection mechanism. Differentiated denoising thresholds are adapted for different clutter types, which can effectively adapt to complex land and sea clutter environments, accurately remove clutter interference and capture effective radiation signals, greatly improve the detection probability and acquisition reliability of radar and communication effective signals under complex clutter backgrounds, and ensure the accuracy of monitoring data from the source.

[0019] (2) In this invention, a fusion direction finding method combining coarse angle measurement with local interval MUSIC fine angle measurement is adopted, and a fuzzy clustering point screening mechanism is used. This avoids the defects of low accuracy and high computing power overhead of single angle measurement algorithm, adapts to the real-time computing needs of UAV, and can effectively eliminate clutter false alarm points, significantly improving the passive positioning accuracy and target discrimination capability of target.

[0020] (3) In this invention, the track survival time difference judgment rule is set to realize the effective management of the track. At the same time, the extended Kalman filter algorithm that can adapt to the target maneuvering state is adopted, which can effectively adapt to the characteristics of the intermittent frame breakage of passive signal, avoid track breakage and target ID jump problem, and solve the problems of target maneuvering tracking lag and track jitter, greatly improving the continuity, stability and stability of moving target track tracking.

[0021] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating a method for tracking and monitoring moving target signals carried by an unmanned aerial vehicle (UAV) according to the present invention. Figure 2 This is a schematic diagram of the structure of a UAV-borne target signal tracking and monitoring system according to the present invention; Figure 3 This is a comparison chart of monitoring indicators between embodiments of the present invention and traditional UAV monitoring technology. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Specific model specifications need to be selected and determined according to the actual specifications of the device, etc. The specific selection calculation method adopts existing technology in the art, and therefore will not be described in detail.

[0024] Example like Figure 1As shown, the present invention provides a method for tracking and monitoring the signal of a moving target carried by an unmanned aerial vehicle (UAV), comprising the following steps: S1. The radio frequency signals radiated by radar and communication equipment in the environment are synchronously collected by a multi-element antenna array, and multi-channel intermediate frequency digital sampling signals are obtained after preprocessing. The preprocessing process includes bandpass filtering, low-noise amplification, frequency mixing and downconversion, synchronous ADC analog-to-digital conversion, and multi-channel amplitude-phase consistency calibration, and finally outputs multiple intermediate frequency digital sampling signals in the form of synchronous I / Q orthogonal complex numbers. The channel amplitude calibration formula is: ; in, For the first i The original received signal amplitude of the road channel, For the first i The preset amplitude calibration coefficient for each channel is used to compensate for differences in hardware channel gain. To ensure consistent signal amplitude and channel phase across all channels after calibration, the calibration formula is as follows: ; in, For the first i The original received signal phase of the road channel, For the first i The inherent phase deviation compensation value of the road channel hardware. The signal phase is aligned for each channel after calibration.

[0025] S2. Adaptive dynamic threshold lifting wavelet layered noise reduction is used to remove clutter from each intermediate frequency signal. The full frequency band spectrum is divided into near clutter region and far clutter region. The OS-CFAR (Ordered Statistical Constant False Alarm) detection threshold is calculated independently for each region. The majority channel consistent judgment is used to fuse the decision based on the multi-channel detection results, capture the effective radiated signal and output the signal frequency domain parameters. S21. Perform layered wavelet decomposition on each intermediate frequency digital sampling signal, and adopt an adaptive dynamic threshold lifting wavelet layered noise reduction algorithm. Configure appropriate adaptive dynamic noise reduction thresholds for ground clutter that follows a log-normal distribution and sea clutter that follows a Rayleigh distribution respectively to complete clutter removal and reconstruct clean signals. The adaptive dynamic wavelet denoising formula is as follows: ; in, The standard deviation of the clutter noise estimate for the current frame characterizes the real-time clutter intensity. This represents the total number of sampling points for a single frame of signal; a ground clutter correction threshold is derived based on this basic threshold. Sea clutter correction threshold , These are correction coefficients specific to land and sea clutter, respectively.

[0026] S22. Perform full-band spectrum analysis on the clean signal after noise reduction. Based on the location of the receiving device, divide the full-band spectrum into a near clutter region of less than 1km and a far clutter region of greater than or equal to 1km. Statistically count the clutter level in the two clutter regions independently and calculate the OS-CFAR signal detection threshold separately. The OS-CFAR signal detection threshold formula is: ; in, After sorting the spectrum of the partitioned reference unit, the first k Level noise, This is the constant false alarm rate ratio, used to maintain a constant false alarm probability in the system. This is the final signal detection threshold for the partition.

[0027] S23. Each channel completes frequency signal detection based on the OS-CFAR detection threshold of the corresponding partition to obtain single-channel signal detection results. Combining the multi-channel detection results, a majority-channel consistent judgment fusion decision method is adopted to identify and capture real and effective radiated signals and eliminate random clutter interference signals. The consensus decision-making method for most channels is as follows: when the proportion of effective detection channels exceeds the preset half threshold, it is determined that there is a real and effective radiated signal in the current frequency band; otherwise, it is determined to be random clutter interference and is eliminated. S24. Perform parameter calculation on the captured effective radiation signal, accurately extract and output the frequency domain parameters of the signal center frequency, occupied bandwidth, and average signal power, and provide effective data support for the subsequent accurate angle measurement process.

[0028] S3. The direction of arrival (DOA) is roughly calculated using an interferometer and the local search range of MUSIC (Multi-Signal Classification Algorithm) is locked. Within the limited range, the angle measurement is finely corrected using the MUSIC spatial spectrum. The multi-channel angle measurement data is weighted and fused to obtain accurate DOA data. S31. Based on the effective radiation signal frequency domain parameters and multiple intermediate frequency digital sampling signals output by S2, the target wave direction angle is quickly and roughly calculated using the interferometer algorithm by utilizing the channel phase difference characteristics of the multi-element antenna array to obtain low-precision initial angle measurement data. The formula for coarse angle measurement using an interferometer is: ; in, For the phase difference received by the array element channel, To fix the element spacing of the array, The effective signal electromagnetic wave wavelength, The initial azimuth of the incoming wave to the target; S32. Using the direction of arrival angle coarsely measured by the interferometer as the center, define a local search interval of ±5° for the coarsely measured azimuth angle to lock the fine search range of the MUSIC algorithm. S33. Within the limited local search interval, construct the covariance matrix for the multi-channel array signals, complete the decomposition of the signal subspace and noise subspace, and complete the fine correction of angle measurement through MUSIC spatial spectrum peak search to obtain single-channel high-precision angle measurement data. The array covariance matrix is: ; in, Let be the signal covariance matrix. For multiple intermediate frequency signals, complex vectors The conjugate transpose of the signal vector. For mathematical expectation operations; The formula for the MUSIC spatial spectrum is: ; in, Angle-guided vector, The noise subspace matrix, The value represents the spatial spectral amplitude, with the peak value corresponding to the precise angle measurement result.

[0029] S34. The high-precision angle measurement data output independently by each channel is weighted and fused to suppress random interference and measurement errors in single channels, and the fused output is accurate DOA data with strong anti-interference and stable accuracy, providing reliable angle parameters for subsequent target positioning calculation. The multi-channel weighted fusion formula is as follows: ; in, To ultimately integrate the DOA perspective, The total number of channels. For the first i Dynamic weighting of road channels This is a single-channel high-precision angle measurement result.

[0030] S4. Based on the multi-channel DOA direction finding line, cross calculation is performed to generate the target original latitude and longitude point sequence containing clutter isolated false alarm points; S41. Using the center of the multi-element antenna array as the coordinate reference point, based on the accurate DOA data output in step S3, construct multiple spatial direction-finding rays corresponding to different receiving channels to obtain multi-channel DOA direction-finding lines. S42. Real-time acquisition of the geodetic coordinates and elevation information of the local array center, establishment of a spherical geographic coordinate system, mapping of multiple spatial direction finding rays to the geodetic coordinate system, and completion of geometric cross-solution calculation using the angle differences of multiple channel direction finding lines to obtain the geodetic latitude and longitude coordinates of the radiation source target. S43. According to the fixed sampling period of the system, the position of the dynamic radiation source target is continuously calculated, and the real-time latitude and longitude points of the target are output frame by frame to generate a continuous sequence of changes in the original position of the target. S44. Retain all original point data, including isolated false alarm points caused by land-sea clutter and multipath interference, to form a target original latitude and longitude point sequence containing isolated false alarm points with clutter, for subsequent clutter screening and processing.

[0031] S5. Use FCM fuzzy clustering to pre-screen the original points and remove isolated clutter points with membership degrees lower than a preset threshold to obtain a set of candidate valid points. S51. Import the target original latitude and longitude point sequence of isolated false alarm points with clutter generated in step S4, and construct all the original points into a point dataset to be clustered. S52. The FCM fuzzy clustering algorithm is used to perform iterative clustering operations on the dataset of points to be clustered, dividing the effective target clusters and clutter discrete point clusters, and calculating the cluster membership value corresponding to each original point. The objective function of FCM is: ; in, This is the clustering loss function; the smaller the value, the better the clustering effect. This represents the total number of original points to be clustered in a single frame. For the number of clusters, For the first i The point belongs to the first j The membership degree of each cluster. The coordinates of the cluster centers For fuzzy weighted index, Euclidean distance calculation; The membership degree iteration formula is: ; The iterative formula for cluster centers is: ; S53. Based on the preset membership threshold, compare the cluster membership of each point, identify and remove isolated clutter false alarm points with membership below the preset threshold, and filter out invalid discrete points. S54. Retain valid target points with high membership, complete the preliminary screening of original points, and organize them to obtain a clean set of candidate valid points, providing effective data support for subsequent point track pairing.

[0032] S6. In the candidate valid point track set, complete the pairing process between the point track and the existing track based on the Mahalanobis distance gate; S61. Import the set of candidate valid points and tracks that have been filtered in step S5, and at the same time retrieve the historical target track data cached locally in the system. Determine whether there are valid historical tracks at present by the track survival time difference rule. S62. If a valid historical track exists, a Mahalanobis distance gate is constructed based on the target position and velocity status information, and the Mahalanobis distance between each candidate valid track and the existing track is calculated one by one; if no valid historical track exists, the current candidate track is directly determined to be a new target track. S63. For cases with historical tracks, the calculated Mahalanobis distance is compared with the preset gate threshold to determine whether the candidate track falls within the matching gate of the corresponding track, and to distinguish between matching tracks and isolated new tracks. S64. The candidate points falling within the Mahalanobis distance gate are associated and paired with the corresponding existing tracks to achieve the matching and connection of new and old track data; the isolated points that are not matched and the newly created points without corresponding historical tracks are initialized to generate new target tracks, providing a complete data foundation for subsequent track filtering iteration processing.

[0033] S7. Construct a six-dimensional state equation for the target that includes position, velocity, and acceleration. Adaptively adjust the extended Kalman EKF process noise matrix online based on the target's real-time acceleration and turning rate. Output a smooth and continuous target time-series trajectory after filtering and iteration. S71. Based on the paired associated point traces and newly generated track data, construct a six-dimensional motion state equation containing the target's two-dimensional position, two-dimensional velocity, and two-dimensional acceleration, and establish an extended Kalman EKF filter iterative model. The six-dimensional state vector is: ; in, For the target two-dimensional position, For two-dimensional motion velocity, It is the acceleration of two-dimensional motion; The state transition equation is: ; The observation equation is: ; in, , These are the target's six-dimensional motion state vectors at the current and previous time steps, respectively. Here is the state transition matrix. For the observation matrix, For process noise, To observe noise; S72. Real-time calculation of target motion state parameters, continuous monitoring of target instantaneous acceleration and turning rate, and real-time determination of target maneuver intensity; S73. Adaptively adjust the EKF process noise matrix online based on the target's maneuvering state, when the target's instantaneous acceleration is greater than 0.5 m / s². 2 When the turning speed is greater than 2° / s, the process noise matrix is ​​increased to adapt to the target's maneuvering state; when the target moves smoothly, the process noise matrix is ​​decreased to ensure the smoothness of the filtering. S74. Iterative updates are completed through adaptively adjusted extended Kalman filters to correct point observation errors and motion deviations, and output smooth, stable and continuous target time-series tracks.

[0034] S8. Upload the target time-series trajectory and raw monitoring data to the display and control terminal to complete data storage and trajectory visualization plotting. S81. Collect the smooth and continuous target time-series trajectory data output in step S7, and synchronously integrate the corresponding signal frequency domain parameters, accurate DOA data, and a complete set of monitoring data of target positioning points throughout the entire process to construct a complete target monitoring data package. S82. Target monitoring data packets are uploaded to the ground display and control terminal in real time via the airborne transmission link to ensure the real-time performance and integrity of data transmission; S83, the ground display and control terminal parses and processes the received target monitoring data packets, and locally stores the track data and raw monitoring data to achieve data archiving and retention, which facilitates subsequent data backtracking and analysis. S84. Based on the parsed target time-series trajectory data, the real-time dynamic visualization plotting of the target trajectory is completed on the display and control terminal interface, intuitively displaying the target's real-time position and continuous motion trajectory, and completing the entire process of UAV-borne target signal tracking and monitoring.

[0035] like Figure 2 As shown, a UAV-borne target signal tracking and monitoring system includes a signal acquisition and preprocessing module, a clutter suppression and effective signal detection module, a high-precision DOA angle measurement module, a target point track calculation module, a point track screening module, a point track track matching module, an adaptive track filtering module, and a data visualization storage and transmission module. The modules are serially coupled, iteratively optimized at each stage, and have unidirectional closed-loop flow. The signal acquisition and preprocessing module is used to acquire ambient radio frequency signals through a multi-element antenna array, and sequentially complete bandpass filtering, low-noise amplification, frequency mixing and downconversion, synchronous ADC analog-to-digital conversion and multi-channel amplitude and phase calibration, and output multiple synchronous I / Q orthogonal complex intermediate frequency digital sampling signals; The clutter suppression and effective signal detection module is used to perform adaptive dynamic threshold lifting wavelet layered noise reduction on each intermediate frequency digital sampling signal. Different noise reduction thresholds are configured for ground clutter and sea clutter. The OS-CFAR detection threshold is calculated independently for near and far clutter zones. The effective radiated signal is captured by the consensus decision of most channels and the center frequency, bandwidth and power frequency domain parameters of the output signal are extracted. The high-precision DOA angle measurement module is used to roughly calculate the target direction of arrival angle through an interferometer algorithm, lock the local search range of ±5°, and complete the high-precision angle measurement correction within the limited range by combining the MUSIC algorithm. It also performs weighted fusion of multi-channel angle measurement data to output accurate DOA data. The target point trace calculation module is used to construct spatial direction finding rays based on multi-channel DOA direction finding lines, and combine the geodetic coordinate system to complete geometric cross-combined calculations to continuously generate the target original latitude and longitude trace sequence containing clutter isolated false alarm points. The point filtering module is used to iteratively cluster the original point dataset using the FCM fuzzy clustering algorithm, calculate the membership degree of each point, remove isolated clutter false alarm points with low membership degree, and filter to obtain a pure set of candidate valid points. The point track pairing module is used to effectively determine the survival status of historical tracks. It completes the association and pairing of valid points with historical tracks through the Mahalanobis distance gate, and completes track initialization for new targets without valid historical tracks. The adaptive track filtering module is used to construct the target's six-dimensional motion state equation, adaptively adjust the EKF process noise matrix according to the target's instantaneous acceleration and turning rate, correct errors through iterative filtering, and output a smooth and continuous target time-series track. The data visualization, storage, and transmission module is used to integrate monitoring data from the entire process to construct monitoring data packets, which are then uploaded to the ground display and control terminal in real time. This completes data parsing, local storage, and real-time dynamic visualization of target tracks.

[0036] like Figure 3As shown, this invention achieves higher signal detection accuracy and track continuity and integrity rates than traditional UAV-borne monitoring technologies, while lower false alarm rates in spectrum detection and average target positioning errors. Employing a universal integrated processing architecture for radar and communication signals, it can adapt to complex situations where the same physical target simultaneously radiates radar pulse signals and continuous communication signals. It can simultaneously complete RF acquisition and preprocessing of both types of signals, achieving frequency domain separation, independent detection, and parameter extraction based on the differences in frequency domain characteristics. High-precision direction finding and passive positioning are then performed to generate two co-source tracks. Subsequently, dual-signal co-source correlation fusion is achieved through clustering spatial correlation discrimination and Mahalanobis distance track matching. The alternating and complementary characteristics of the two signals solve the track breakage problem caused by short-term frame breaks in passive signals. Finally, the integrated summary display of single-target dual-mode signals is achieved on the display and control terminal, effectively avoiding the defects of traditional equipment such as dual-signal aliasing misjudgment, incomplete target monitoring, and unstable track tracking. This significantly improves the robustness and integrity of UAV-borne multi-mode radiation source monitoring in complex electromagnetic environments.

[0037] Therefore, this invention employs a method and system for tracking and monitoring moving targets carried by unmanned aerial vehicles (UAVs). Through an integrated and collaborative design that combines multi-level clutter suppression, zoned constant false alarm rate (CFAR) detection, coarse and fine combined angle measurement, intelligent clutter screening, track survival validity determination, and adaptive maneuvering filtering, it effectively compensates for the technical shortcomings of traditional monitoring schemes. This significantly improves the reliability of effective signal acquisition and target positioning accuracy in complex sea and land clutter environments, ensuring the continuity, stability, and smoothness of moving target track tracking in passive monitoring scenarios. It possesses advantages such as strong anti-interference capability, good real-time performance, high tracking accuracy, and wide environmental adaptability, effectively meeting the high-precision, continuous, and stable monitoring and tracking requirements of UAV-carried moving targets in complex electromagnetic environments.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for tracking and monitoring signals of a UAV-borne moving target, characterized in that: Includes the following steps: S1. The radio frequency signals radiated by radar and communication equipment in the environment are synchronously collected by a multi-element antenna array, and multi-channel intermediate frequency digital sampling signals are obtained after preprocessing. S2. Adaptive dynamic threshold lifting wavelet layered noise reduction is used to remove clutter from each intermediate frequency signal. The full frequency band spectrum is divided into near clutter region and far clutter region. The OS-CFAR detection threshold is calculated independently for each region. The decision is made by combining the multi-channel detection results with the majority channel consistent judgment to capture the effective radiated signal and output the signal frequency domain parameters. S3. The direction of arrival (DOA) is roughly calculated using an interferometer and the local search range of MUSIC is locked. Within the limited range, the angle measurement is finely corrected using the MUSIC spatial spectrum. The multi-channel angle measurement data is weighted and fused to obtain accurate DOA data. S4. Based on the multi-channel DOA direction finding line, cross calculation is performed to generate the target original latitude and longitude point sequence containing clutter isolated false alarm points; S5. Use FCM fuzzy clustering to pre-screen the original points and remove isolated clutter points with membership degrees lower than a preset threshold to obtain a set of candidate valid points. S6. In the candidate valid point track set, complete the pairing process between the point track and the existing track based on the Mahalanobis distance gate; S7. Construct a six-dimensional state equation for the target that includes position, velocity, and acceleration. Adaptively adjust the extended Kalman EKF process noise matrix online based on the target's real-time acceleration and turning rate. Output a smooth and continuous target time-series trajectory after filtering and iteration. S8. Upload the target time-series trajectory and raw monitoring data to the display and control terminal to complete data storage and trajectory visualization plotting.

2. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 1, characterized in that: In step S1, the preprocessing includes bandpass filtering, low-noise amplification, frequency mixing and downconversion, synchronous ADC analog-to-digital conversion, and multi-channel amplitude-phase consistency calibration, and finally outputs multi-channel intermediate frequency digital sampling signals in the form of synchronous I / Q orthogonal complex numbers.

3. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 2, characterized in that: Step S2 specifically includes the following steps: S21. Perform layered wavelet decomposition on each intermediate frequency digital sampling signal, and adopt an adaptive dynamic threshold lifting wavelet layered noise reduction algorithm. Configure appropriate adaptive dynamic noise reduction thresholds for ground clutter that follows a log-normal distribution and sea clutter that follows a Rayleigh distribution respectively to complete clutter removal and reconstruct clean signals. S22. Perform full-band spectrum analysis on the clean signal after noise reduction. Based on the location of the receiving device, divide the full-band spectrum into a near clutter region of less than 1km and a far clutter region of greater than or equal to 1km. Statistically count the clutter level in the two clutter regions independently and calculate the OS-CFAR signal detection threshold separately. S23. Each channel completes frequency signal detection based on the OS-CFAR detection threshold of the corresponding partition to obtain single-channel signal detection results. Combining the multi-channel detection results, a majority-channel consistent judgment fusion decision method is adopted to identify and capture real and effective radiated signals and eliminate random clutter interference signals. S24. Perform parameter calculation on the captured effective radiation signal, accurately extract and output the frequency domain parameters of the signal center frequency, occupied bandwidth, and average signal power, and provide effective data support for the subsequent accurate angle measurement process.

4. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 3, characterized in that: Step S3 specifically includes the following steps: S31. Based on the effective radiation signal frequency domain parameters and multiple intermediate frequency digital sampling signals output by S2, the target wave direction angle is quickly and roughly calculated using the interferometer algorithm by utilizing the channel phase difference characteristics of the multi-element antenna array to obtain low-precision initial angle measurement data. S32. Using the direction of arrival angle coarsely measured by the interferometer as the center, define a local search interval of ±5° for the coarsely measured azimuth angle to lock the fine search range of the MUSIC algorithm. S33. Within the limited local search interval, construct the covariance matrix for the multi-channel array signals, complete the decomposition of the signal subspace and noise subspace, and complete the fine correction of angle measurement through MUSIC spatial spectrum peak search to obtain single-channel high-precision angle measurement data. S34. The high-precision angle measurement data output independently from each channel are weighted and fused to suppress random interference and measurement errors in single channels, and the fused output is accurate DOA data with strong anti-interference and stable accuracy, providing reliable angle parameters for subsequent target positioning calculation.

5. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 4, characterized in that: Step S4 Specifically, the following steps are included: S41. Using the center of the multi-element antenna array as the coordinate reference point, based on the accurate DOA data output in step S3, construct multiple spatial direction-finding rays corresponding to different receiving channels to obtain multi-channel DOA direction-finding lines. S42. Real-time acquisition of the geodetic coordinates and elevation information of the local array center, establishment of a spherical geographic coordinate system, mapping of multiple spatial direction finding rays to the geodetic coordinate system, and completion of geometric cross-solution calculation using the angle differences of multiple channel direction finding lines to obtain the geodetic latitude and longitude coordinates of the radiation source target. S43. According to the fixed sampling period of the system, the position of the dynamic radiation source target is continuously calculated, and the real-time latitude and longitude points of the target are output frame by frame to generate a continuous sequence of changes in the original position of the target. S44. Retain all original point data, including isolated false alarm points caused by land-sea clutter and multipath interference, to form a target original latitude and longitude point sequence containing isolated false alarm points with clutter, for subsequent clutter screening and processing.

6. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 5, characterized in that: Step S5 specifically includes the following steps: S51. Import the target original latitude and longitude point sequence of isolated false alarm points with clutter generated in step S4, and construct all the original points into a point dataset to be clustered. S52. The FCM fuzzy clustering algorithm is used to perform iterative clustering operations on the dataset of points to be clustered, dividing the effective target clusters and clutter discrete point clusters, and calculating the cluster membership value corresponding to each original point. S53. Based on the preset membership threshold, compare the cluster membership of each point, identify and remove isolated clutter false alarm points with membership below the preset threshold, and filter out invalid discrete points. S54. Retain valid target points with high membership, complete the preliminary screening of original points, and organize them to obtain a clean set of candidate valid points, providing effective data support for subsequent point track pairing.

7. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 6, characterized in that: Step S6 specifically includes the following steps: S61. Import the set of candidate valid points and tracks that have been filtered in step S5, and at the same time retrieve the historical target track data cached locally in the system. Determine whether there are valid historical tracks at present by the track survival time difference rule. S62. If a valid historical track exists, a Mahalanobis distance gate is constructed based on the target position and velocity status information, and the Mahalanobis distance between each candidate valid track and the existing track is calculated one by one; if no valid historical track exists, the current candidate track is directly determined to be a new target track. S63. For cases with historical tracks, the calculated Mahalanobis distance is compared with the preset gate threshold to determine whether the candidate track falls within the matching gate of the corresponding track, and to distinguish between matching tracks and isolated new tracks. S64. The candidate points falling within the Mahalanobis distance gate are associated and paired with the corresponding existing tracks to achieve the matching and connection of new and old track data; the isolated points that are not matched and the newly created points without corresponding historical tracks are initialized to generate new target tracks, providing a complete data foundation for subsequent track filtering iteration processing.

8. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 7, characterized in that: Step S7 specifically includes the following steps: S71. Based on the paired associated point traces and newly generated track data, construct a six-dimensional motion state equation containing the target's two-dimensional position, two-dimensional velocity, and two-dimensional acceleration, and establish an extended Kalman EKF filter iterative model. S72. Real-time calculation of target motion state parameters, continuous monitoring of target instantaneous acceleration and turning rate, and real-time determination of target maneuver intensity; S73. Adaptively adjust the EKF process noise matrix online based on the target's maneuvering state, when the target's instantaneous acceleration is greater than 0.5 m / s². 2 When the turning speed is greater than 2° / s, the process noise matrix is ​​increased to adapt to the target's maneuvering state; when the target moves smoothly, the process noise matrix is ​​decreased to ensure the smoothness of the filtering. S74. Iterative updates are completed through adaptively adjusted extended Kalman filters to correct point observation errors and motion deviations, and output smooth, stable and continuous target time-series tracks.

9. The method for tracking and monitoring unmanned aerial vehicle (UAV) target signals according to claim 8, characterized in that: Step S8 specifically includes the following steps: S81. Collect the smooth and continuous target time-series trajectory data output in step S7, and synchronously integrate the corresponding signal frequency domain parameters, accurate DOA data, and a complete set of monitoring data of target positioning points throughout the entire process to construct a complete target monitoring data package. S82. Target monitoring data packets are uploaded to the ground display and control terminal in real time via the airborne transmission link to ensure the real-time performance and integrity of data transmission; S83, the ground display and control terminal parses and processes the received target monitoring data packets, and locally stores the track data and raw monitoring data to achieve data archiving and retention, which facilitates subsequent data backtracking and analysis. S84. Based on the parsed target time-series trajectory data, the real-time dynamic visualization plotting of the target trajectory is completed on the display and control terminal interface, intuitively displaying the target's real-time position and continuous motion trajectory, and completing the entire process of UAV-borne target signal tracking and monitoring.

10. A UAV-borne target signal tracking and monitoring system, used to implement the UAV-borne target signal tracking and monitoring method according to any one of claims 1-9, characterized in that: It includes a signal acquisition and preprocessing module, a clutter suppression and effective signal detection module, a high-precision DOA angle measurement module, a target point track calculation module, a point track filtering module, a point track track matching module, an adaptive track filtering module, and a data visualization, storage, and transmission module; The signal acquisition and preprocessing module is used to acquire ambient radio frequency signals through a multi-element antenna array, and sequentially complete bandpass filtering, low-noise amplification, frequency mixing and downconversion, synchronous ADC analog-to-digital conversion and multi-channel amplitude and phase calibration, and output multiple synchronous I / Q orthogonal complex intermediate frequency digital sampling signals; The clutter suppression and effective signal detection module is used to perform adaptive dynamic threshold lifting wavelet layered noise reduction on each intermediate frequency digital sampling signal. Different noise reduction thresholds are configured for ground clutter and sea clutter. The OS-CFAR detection threshold is calculated independently for near and far clutter zones. The effective radiated signal is captured by the consensus decision of most channels and the center frequency, bandwidth and power frequency domain parameters of the output signal are extracted. The high-precision DOA angle measurement module is used to roughly calculate the target direction of arrival angle through an interferometer algorithm, lock the local search range of ±5°, and complete the high-precision angle measurement correction within the limited range by combining the MUSIC algorithm. It also performs weighted fusion of multi-channel angle measurement data to output accurate DOA data. The target point trace calculation module is used to construct spatial direction finding rays based on multi-channel DOA direction finding lines, and combine the geodetic coordinate system to complete geometric cross-combined calculations to continuously generate the target original latitude and longitude trace sequence containing clutter isolated false alarm points. The point filtering module is used to iteratively cluster the original point dataset using the FCM fuzzy clustering algorithm, calculate the membership degree of each point, remove isolated clutter false alarm points with low membership degree, and filter to obtain a pure set of candidate valid points. The point track pairing module is used to effectively determine the survival status of historical tracks. It completes the association and pairing of valid points with historical tracks through the Mahalanobis distance gate, and completes track initialization for new targets without valid historical tracks. The adaptive track filtering module is used to construct the target's six-dimensional motion state equation, adaptively adjust the EKF process noise matrix according to the target's instantaneous acceleration and turning rate, correct errors through iterative filtering, and output a smooth and continuous target time-series track. The data visualization, storage, and transmission module is used to integrate monitoring data from the entire process to construct monitoring data packets, which are then uploaded to the ground display and control terminal in real time. This completes data parsing, local storage, and real-time dynamic visualization of target tracks.

Images