An underwater equipment single-base station passive target detection and tracking method and related device

Abstract

This application belongs to the field of target tracking technology and discloses a single-base station passive detection target tracking method and related equipment for underwater devices. It generates multiple Kalman filter calculation functions and initializes the target state parameters of each function. Each time the single-base station passive detector detects the real-time direction angle of a surface target, it updates the target state parameters of each Kalman filter calculation function using a strong tracking filter algorithm or a Kalman filter algorithm based on the real-time direction angle and the number of times the target state parameters are updated. When the number of times the target state parameters are updated reaches a preset first threshold, the target state parameters of one of the Kalman filter calculation functions are selected as the measured target state parameters of the surface target using a maximum likelihood estimation method. This improves the accuracy and practicality of surface target tracking.

Description

Technical Field

[0001] This application relates to the field of target tracking technology, and more specifically, to a passive target tracking method and related equipment for a single base station of an underwater device. Background Technology

[0002] Target tracking is one of the key technologies for underwater unmanned equipment to achieve ocean exploration. Currently, underwater equipment target detection is mainly divided into active sonar detection and passive sonar detection. Active sonar detection involves actively emitting sound waves and using the echoes to obtain target-related information, but its effective detection range is relatively small, and it easily reveals the operator's position. In contrast, passive detection determines the target's location by listening to noise sources, offering advantages such as a large detection range and high concealment. Passive detection can be further divided into single-base station passive detection and multi-base station passive detection. Multi-base station passive detection suffers from problems such as mutual interference between base stations and measurement errors, and multiple base stations increase the size, cost, and manufacturing and usage difficulty of the underwater unmanned equipment. Therefore, single-base station passive detection has higher value in practical applications.

[0003] However, current single-base station passive detection target tracking algorithms have many problems, limiting their practicality. For example, some existing target tracking methods (such as the target tracking method disclosed in the paper "Research on Pure Azimuth Target Tracking Method Based on Underwater Passive Detection") have unreasonable initial value settings before Kalman filtering. In practical applications, the maximum effective detection range of the noise source will change depending on the noise level. If the initial value differs significantly from the true value, the Kalman filtering algorithm may fail to converge or fail to converge within a reasonable time, resulting in the method only being able to calculate targets with initial true distances that are not significantly different from the initial value. Secondly, these methods require too much time to converge the error to a suitable range during filtering, making them impractical in real-world applications. They are prone to situations where the target has already left the detection range before the algorithm has converged the target state value error to a sufficiently accurate range. Furthermore, existing methods using strong tracking filtering algorithms for target state updates are prone to divergence and difficulties in converging the results to a small range.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] The purpose of this application is to provide a passive target tracking method and related equipment for underwater devices with a single base station, which aims to solve the problems of unreasonable initial value setting of Kalman filter, excessive error convergence time and easy divergence of strong tracking filter algorithm in existing passive target tracking algorithms with a single base station, thereby improving the practicality and accuracy of surface target tracking.

[0006] In a first aspect, this application provides a single-base station passive detection target tracking method for underwater equipment, used to track surface targets using a single-base station passive detector. The method includes the following steps:

[0007] A1. Generate multiple Kalman filter calculation functions;

[0008] A2. Obtain the initial orientation angle of the water surface target detected by the single base station passive detector, and initialize the target state parameters of each of the Kalman filter calculation functions based on the coordinates of multiple position points spaced apart in the direction corresponding to the initial orientation angle;

[0009] A3. After the initialization of the target state parameters is completed, each time the single base station passive detector detects the real-time direction angle of the water surface target, the target state parameters of each of the Kalman filter calculation functions are updated according to the real-time direction angle and the number of times the target state parameters are updated, using a strong tracking filter algorithm or a Kalman filter algorithm.

[0010] A4. When the number of times the target state parameter is updated reaches a preset first threshold, the target state parameter of one of the Kalman filter calculation functions is selected as the measured target state parameter of the water surface target by the maximum likelihood estimation method.

[0011] Preferably, step A1 includes:

[0012] A101. Obtain information on currently available computing resources;

[0013] A102. Determine the number of Kalman filter computation functions to be generated based on the available computing resource information;

[0014] A103. Generate a Kalman filter calculation function based on the number of generated functions.

[0015] Preferably, the target state parameters include target coordinates and target velocity, wherein the target coordinates represent the coordinates of the water surface target and the target velocity represents the velocity of the water surface target;

[0016] Step A2 includes:

[0017] A201. Obtain the initial orientation angle of the water surface target detected by the single-base station passive detector;

[0018] A202. Based on the number of Kalman filter calculation functions generated, select multiple location points at different distances from the single base station passive detector in the direction corresponding to the initial direction angle, and calculate the coordinates of each location point based on the distance between each location point and the single base station passive detector and the initial direction angle;

[0019] A203. Assign the coordinates of each of the aforementioned positions to the target coordinates in the target state parameters of each of the aforementioned Kalman filter calculation functions, and set the target velocity in the target state parameters of each of the aforementioned Kalman filter calculation functions to zero.

[0020] Preferably, in step A202, based on the number of Kalman filter calculation functions generated, multiple position points are selected at equal intervals in the direction corresponding to the initial direction angle, with a preset interval distance;

[0021] Alternatively, based on the number of Kalman filter calculation functions generated and the maximum detection distance of the single base station passive detector, the interval distance is determined, and multiple position points are selected at equal intervals in the direction corresponding to the initial direction angle according to the determined interval distance.

[0022] Preferably, step A3 includes:

[0023] A301. When the single base station passive detector detects the real-time direction angle of the water surface target each time, the total number of times the target state parameters of each Kalman filter calculation function are updated is obtained, and the number of times the target state parameters are updated is obtained.

[0024] A302. Compare the number of times the target state parameter is updated with a preset second threshold; the second threshold is less than the first threshold;

[0025] A303. If the number of times the target state parameter is updated is less than the second threshold, the target state parameter of each of the Kalman filter calculation functions is updated using the strong tracking filter algorithm;

[0026] A304. If the number of times the target state parameter is updated is not less than the second threshold, then the target state parameters of each of the Kalman filter calculation functions are updated using the standard Kalman filter algorithm or the unscented Kalman filter algorithm.

[0027] Preferably, step A4 includes:

[0028] A401. Perform the following steps for each Kalman filter calculation function:

[0029] B1. Based on the current target state parameters calculated by the Kalman filter calculation function, calculate the coordinates of the water surface target at each observation time point before the current time; the observation time point is the time point at which the direction angle of the water surface target is detected by the single base station passive detector.

[0030] B2. Based on the calculated coordinates, calculate the orientation and angle of the water surface target at each observation point before the current time.

[0031] B3. Based on the direction angles of each observation time point obtained by back calculation and the direction angles of each observation time point detected by the single base station passive detector, the likelihood value of each observation time point is calculated using the Gaussian probability density function.

[0032] B4. Calculate the overall likelihood value based on the likelihood values ​​at each observation time point;

[0033] A402. Compare the comprehensive likelihood values ​​corresponding to each Kalman filter calculation function, determine the largest comprehensive likelihood value, and select the current target state parameter of the Kalman filter calculation function corresponding to the largest comprehensive likelihood value as the measured target state parameter of the water surface target.

[0034] Preferably, step B1 includes:

[0035] Based on the current target state parameters calculated by the Kalman filter calculation function, and based on the assumption of uniform linear motion, the coordinates of the water surface target at each observation time point before the current time are calculated.

[0036] Preferably, in step B4, the comprehensive likelihood value is calculated according to the following formula:

[0037] ;

[0038] Where P is the overall likelihood value, and n is the total number of observation time points. Let be the likelihood value at the i-th observation time point.

[0039] Secondly, this application provides an electronic device including a processor and a memory, the memory storing a computer program executable by the processor, wherein when the processor executes the computer program, it performs the steps of the underwater device single-base station passive detection target tracking method described above.

[0040] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the steps of the underwater device single-base station passive detection target tracking method described above.

[0041] Beneficial Effects: This application provides a passive target tracking method and related equipment for underwater devices using a single base station. By generating multiple Kalman filter calculation functions and initializing them based on the coordinates of multiple position points corresponding to the initial direction angle, it effectively solves the problem in existing technologies where unreasonable initial values ​​of the Kalman filter lead to convergence failure or excessively long convergence times. Specifically, by initializing the target state parameters of each Kalman filter calculation function with multiple position points spaced apart in the direction corresponding to the initial direction angle, the initial values ​​can more comprehensively cover the possible initial positions of the target, thereby significantly improving the initial convergence speed and accuracy of the Kalman filter algorithm. Simultaneously, during the target state parameter update process, this application dynamically switches between the strong tracking filter algorithm and the Kalman filter algorithm based on the number of updates. While ensuring rapid convergence, it effectively avoids the divergence problem that may occur in the strong tracking filter algorithm, improving the stability and accuracy of the algorithm. Furthermore, when the number of target state parameter updates reaches a preset threshold, the optimal target state parameters are selected from multiple Kalman filter calculation functions using the maximum likelihood estimation method, further improving the accuracy and reliability of target tracking and overcoming the error accumulation and divergence problems that may exist with a single filter in existing technologies. In summary, the method of this application effectively solves the shortcomings of existing single-base station passive detection target tracking algorithms in terms of initial value setting, convergence speed and algorithm stability, and improves the practicality and accuracy of water surface target tracking. Attached Figure Description

[0042] Figure 1 A flowchart of a passive target tracking method for a single base station of an underwater device provided in this application.

[0043] Figure 2 A schematic diagram of the structure of the electronic device provided in this application.

[0044] Labeling explanations: 301, processor; 302, memory; 303, communication bus. Detailed Implementation

[0045] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0046] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0047] Please refer to Figure 1 This application discloses a single-base station passive detection target tracking method for underwater equipment, which is used to track surface targets using a single-base station passive detector. The method includes the following steps:

[0048] A1. Generate multiple Kalman filter calculation functions;

[0049] A2. Obtain the initial orientation angle of the water surface target detected by the single base station passive detector, and initialize the target state parameters of each Kalman filter calculation function based on the coordinates of multiple position points arranged at intervals in the direction corresponding to the initial orientation angle;

[0050] A3. After the target state parameters are initialized, each time the passive detector of a single base station detects the real-time direction angle of the water surface target, the target state parameters of each Kalman filter calculation function are updated according to the real-time direction angle and the number of times the target state parameters are updated, using a strong tracking filter algorithm or a Kalman filter algorithm.

[0051] A4. When the number of times the target state parameters are updated reaches the preset first threshold, the target state parameters of one of the Kalman filter calculation functions are selected as the measured target state parameters of the water surface target through the maximum likelihood estimation method.

[0052] This application generates multiple Kalman filter calculation functions and initializes the target state parameters of each function. This allows the single-base station passive detector to update the target state parameters of each Kalman filter calculation function based on the real-time direction angle and the number of target state parameter updates, using either a strong tracking filter algorithm or a Kalman filter algorithm. Finally, when the number of target state parameter updates reaches a preset first threshold, the target state parameters of one of the Kalman filter calculation functions are selected as the measured target state parameters of the water surface target using a maximum likelihood estimation method. This method effectively solves the problems of unreasonable initial Kalman filter values, excessively long convergence times, and the tendency of strong tracking filter algorithms to diverge in existing technologies, thus improving the accuracy and practicality of water surface target tracking.

[0053] Among them, the single-base station passive detector is an underwater device that detects targets by receiving acoustic signals generated by targets on the water surface, without actively emitting sound waves, thereby achieving covert detection.

[0054] The Kalman filter computation function is a mathematical tool for estimating the state of a system. It provides an estimate of the target state by fusing measurement data and a system model.

[0055] Target state parameters are a set of variables that describe the current state of a surface target, typically including information such as target coordinates and target velocity.

[0056] The initial orientation angle refers to the orientation of the target relative to the detector when a single-base station passive detector first detects a target on the water surface. The real-time orientation angle refers to the real-time orientation of the target relative to the detector each time the single-base station passive detector detects a target on the water surface during subsequent detection processes.

[0057] Strong tracking filtering and Kalman filtering are two different filtering strategies. The former shows better tracking performance when the target is highly maneuverable, while the latter has higher estimation accuracy when the target is moving smoothly.

[0058] Maximum likelihood estimation is a statistical method used to estimate model parameters from a set of observations in order to maximize the probability of the observed data.

[0059] The first threshold is a preset value used to determine when to select the final target state parameters.

[0060] The core of the underwater equipment single-base station passive target tracking method in this application lies in the fact that it effectively improves the accuracy and robustness of single-base station passive target tracking through multi-model parallel processing and adaptive filtering strategies.

[0061] Specifically, in step A1, multiple Kalman filter computation functions need to be generated. These functions can be viewed as independent trackers, each attempting to estimate the state of the surface target. There are various ways to generate these functions. For example, a fixed number can be pre-defined; these functions are structurally identical but will be assigned different initial state parameters in subsequent initialization phases. Another approach is to dynamically generate the number of Kalman filter computation functions based on available computing resources. For instance, if the system has powerful processing capabilities, more functions can be generated; if computing resources are limited, fewer functions can be generated.

[0062] In step A2, the initial orientation angle of the water surface target detected by the single-base station passive detector is obtained. Based on the coordinates of multiple location points spaced apart along the direction corresponding to the initial orientation angle, the target state parameters of each Kalman filter calculation function are initialized. The initial orientation angle can be calculated using the acoustic signal received by the single-base station passive detector, for example, by determining the target's orientation relative to the detector using a sound source localization algorithm. After obtaining the initial orientation angle, multiple location points need to be selected along that direction. These location points can be arranged at fixed intervals, for example, one location point every 100 meters, or the entire detection range can be divided into several equally spaced regions based on the maximum detection range of the single-base station passive detector, and one location point can be selected within each region. For each selected location point, its coordinates can be calculated based on the distance between the location point and the single-base station passive detector and the initial orientation angle. For example, if the initial orientation angle is θ and the distance is d, the coordinates of the location point can be represented as (d*cosθ, d*sinθ). Then, these calculated location point coordinates are assigned to the target coordinates in the target state parameters of each Kalman filter calculation function. To avoid the influence of the initial velocity on the subsequent filtering process, the target velocity in the target state parameter of each Kalman filter calculation function can be set to zero.

[0063] In step A3, after initializing the target state parameters, each time the single-base station passive detector detects the real-time orientation angle of the water surface target, the target state parameters of each Kalman filter calculation function are updated using either a strong tracking filter algorithm or a Kalman filter algorithm, based on the real-time orientation angle and the number of times the target state parameters have been updated. The single-base station passive detector periodically detects the water surface target, generating a new real-time orientation angle with each detection. After each acquisition of the real-time orientation angle, the total number of times the target state parameters of each Kalman filter calculation function have been updated needs to be recorded to distinguish the tracking stage. In the early stage, the target state may have significant uncertainty; in this case, the strong tracking filter algorithm can be used to update the target state parameters of each Kalman filter calculation function. The strong tracking filter algorithm can better handle target maneuverability and accelerate convergence. In the stable stage, the target state estimation is relatively accurate; in this case, the Kalman filter algorithm can be used to update the target state parameters of each Kalman filter calculation function. The Kalman filter algorithm has higher estimation accuracy when the target motion is stable.

[0064] In step A4, when the number of target state parameter updates reaches a preset first threshold, the target state parameter of one of the Kalman filter calculation functions is selected as the measured target state parameter of the water surface target using the maximum likelihood estimation method. When the number of target state parameter updates reaches the first threshold, it indicates that each Kalman filter calculation function has performed enough updates, and its target state parameter has become relatively stable. At this point, it is necessary to evaluate each Kalman filter calculation function to select the one that best represents the true state of the water surface target.

[0065] The underwater equipment single-base station passive detection target tracking method proposed in this application effectively solves many problems existing in the single-base station passive detection target tracking technology by introducing parallel processing of multi-Kalman filter calculation functions and adaptive filtering strategies.

[0066] Specifically, traditional target tracking methods often fail to set appropriate initial values ​​before Kalman filtering, leading to potential convergence issues or failure to converge within a reasonable timeframe. This application addresses this by generating multiple Kalman filter computation functions and initializing the target state parameters of each function at multiple points spaced apart along the direction corresponding to the initial directional angle. This multi-model initialization strategy ensures that at least one Kalman filter computation function's initial value closely approximates the target's true state, significantly improving both the convergence accuracy and speed of the Kalman filter algorithm.

[0067] Secondly, existing methods require excessive time to converge the error to a suitable range during filtering, easily leading to situations where the target has already left the detection range before the algorithm has converged the target state value error to a sufficiently accurate range. This application addresses this by using a strong tracking filter algorithm to update the target state parameters in the early stages of tracking, and a Kalman filter algorithm to update them in the later stages. The strong tracking filter algorithm can converge the error faster and cope with target maneuverability, while the Kalman filter algorithm provides higher accuracy in the stable tracking phase. This adaptive filtering strategy allows the algorithm to select the most suitable filtering algorithm based on the characteristics of the tracking phase, thereby significantly shortening the error convergence time and improving the real-time performance and effectiveness of tracking.

[0068] Furthermore, existing methods using strong tracking filtering algorithms for target state updates are prone to divergence and difficulties in converging the results to a small range. This application addresses this issue by employing maximum likelihood estimation when the number of target state parameter updates reaches a preset first threshold, selecting the optimal target state parameters from multiple Kalman filter calculation functions. This method avoids the divergence problems that may arise from a single strong tracking filtering algorithm, and through parallel estimation of multiple models and subsequent optimal selection, ensures the accuracy and stability of the final output target state parameters.

[0069] In summary, the underwater device single-base station passive target tracking method of this application significantly improves the robustness, convergence speed, and accuracy of single-base station passive target tracking through parallel initialization of the multi-Kalman filter calculation function, adaptive filter algorithm selection, and optimal state parameter selection based on maximum likelihood estimation. It overcomes the shortcomings of existing technologies such as unreasonable initial value settings, excessively long convergence time, and easy divergence of filter algorithms, and provides strong technical support for underwater unmanned equipment to achieve efficient and covert ocean exploration.

[0070] While a fixed number of Kalman filter calculation functions can be used, this may lead to underutilization of available resources or system overload under resource constraints in practical applications, affecting the real-time performance and accuracy of target tracking. If these issues are not addressed, the performance and stability of the surface target tracking system will be limited under different deployment environments or when facing dynamically changing computational loads. To address this, this application proposes a scheme to optimize the Kalman filter calculation function generation process by dynamically assessing available computational resources to determine the number of functions to generate, thereby improving the system's resource utilization efficiency and environmental adaptability.

[0071] Specifically, step A1 includes:

[0072] A101. Obtain information on currently available computing resources;

[0073] A102. Determine the number of Kalman filter computation functions to be generated based on available computing resource information;

[0074] A103. Generate the Kalman filter calculation function based on the number of generated samples.

[0075] Specifically, in step A101, obtaining the current available computing resource information refers to the system monitoring and collecting the status of currently available computing resources in real time during runtime, such as the utilization rate of the central processing unit (CPU), the available space of memory (RAM), the load of the graphics processing unit (GPU), and the idle status of other parallel computing units. The purpose is to provide a data foundation for subsequently dynamically adjusting the number of Kalman filter calculation functions generated.

[0076] In step A102, determining the number of Kalman filter computation functions to be generated based on available computing resource information can be understood as the system intelligently calculating the most suitable number of Kalman filter computation functions to be generated under the current environment, using preset strategies, algorithms, or models, based on the acquired computing resource information. For example, when sufficient computing resources are detected, more Kalman filter computation functions can be generated to improve tracking accuracy; when computing resources are scarce, the number generated can be appropriately reduced to avoid system overload and ensure basic tracking performance. The aim is to achieve a balance between computing resources and tracking performance.

[0077] In practical applications, step A103 involves generating Kalman filter calculation functions based on the number of generated targets. Specifically, this means the system instantiates a corresponding number of Kalman filter calculation functions based on the number determined in step A102. These functions will be used for subsequent initialization and updating of the water surface target state parameters. The purpose is to provide multi-model parallel processing capabilities for water surface target tracking.

[0078] The proposed solution obtains information on available computing resources before generating the Kalman filter calculation function, enabling the system to perceive its own computing capabilities in real time. Subsequently, based on this real-time resource information, the number of Kalman filter calculation functions generated is dynamically determined, rather than using a fixed preset value. This allows the system to flexibly adjust the computing resources used according to the actual hardware environment and load conditions, avoiding resource waste or system overload caused by generating too many functions, and also avoiding underutilization of resources due to generating too few functions. It is precisely this dynamic adjustment mechanism that allows the surface target tracking method to better adapt to different deployment environments and dynamically changing computing needs.

[0079] Through the above technical solution, this application can intelligently adjust the number of generated Kalman filter calculation functions according to the actual computing resource conditions, thereby significantly improving the resource utilization efficiency of the system. Compared with using a fixed number of Kalman filter calculation functions, the solution of this application can effectively avoid the problem of system performance degradation due to excessive computation when resources are limited, or the problem of not fully utilizing computing power when resources are sufficient. In addition, this dynamic adaptability enables the surface target tracking method to maintain good performance and stability under different hardware configurations and operating loads, enhancing the robustness and practicality of the method.

[0080] Preferably, the target state parameters include target coordinates and target velocity, where the target coordinates represent the coordinates of the water surface target and the target velocity represents the velocity of the water surface target;

[0081] Step A2 includes:

[0082] A201. Obtain the initial orientation angle of the water surface target detected by the single-base station passive detector;

[0083] A202. Based on the number of generated functions of the Kalman filter, select multiple location points at different distances from the single base station passive detector in the direction corresponding to the initial direction angle, and calculate the coordinates of each location point based on the distance between each location point and the single base station passive detector and the initial direction angle.

[0084] A203. Assign the coordinates of each location point to the target coordinates in the target state parameters of each Kalman filter calculation function, and set the target velocity in the target state parameters of each Kalman filter calculation function to zero.

[0085] The target state parameters can be understood as vectors describing the motion state of the water surface target, for example, they can be represented in the form [x, y, vx, vy]. In this representation, x and y represent the coordinate components of the water surface target in the two-dimensional plane, i.e., the target coordinates; vx and vy represent the velocity components of the water surface target in the x-axis and y-axis directions, i.e., the target velocity. This parameter definition method can comprehensively describe the instantaneous position and motion trend of the water surface target, providing the necessary state information for subsequent Kalman filter calculations.

[0086] Specifically, step A201 refers to the process where a single-base station passive detector acquires an initial orientation angle when it first detects a water surface target. This initial orientation angle is the angle information of the line connecting the single-base station passive detector and the water surface target (e.g., the angle between the line and the x-axis), and it serves as the initial observation data for passive detection and tracking.

[0087] In step A202, based on the number of generated Kalman filter calculation functions, multiple location points with different distances from the single-base station passive detector are selected along the direction corresponding to the initial direction angle. Since the single-base station passive detector cannot directly obtain the target's distance information, it is necessary to assume that the target may be located at different distances along this initial direction angle. These location points are the assumed target positions used to initialize different Kalman filter calculation functions. Subsequently, based on the distance between each location point and the single-base station passive detector, as well as the initial direction angle, the specific coordinates of these location points in a preset coordinate system can be calculated.

[0088] Further, in step A203, the coordinates of each location point calculated in step A202 are assigned to the target coordinate components in the target state parameters of each Kalman filter calculation function. Simultaneously, since the velocity information of the water surface target is usually unknown at the initial moment, the target velocity component in the target state parameters of each Kalman filter calculation function is set to zero. This initialization method provides each Kalman filter calculation function with an initial target state based on different distance assumptions, thereby constructing an initial state set covering the possible locations of the target.

[0089] The above technical solution clarifies the specific composition of the target state parameters and provides detailed initialization steps, enabling effective initialization of multiple Kalman filter calculation functions in a single-base station passive detection environment. This initialization method overcomes the limitation of single-base station passive detectors being unable to directly obtain target distance information. By setting multiple assumed distance points at the initial direction angle, reasonable initial target coordinates are provided for each Kalman filter calculation function. Simultaneously, setting the initial target velocity to zero simplifies the initialization process and avoids errors introduced by inaccurate initial velocity estimation. Therefore, the solution proposed in this application can improve the initial convergence speed and accuracy of target tracking, providing a solid foundation for subsequent real-time direction angle updates and accurate estimation of target state parameters, thus enhancing the robustness and applicability of the entire tracking method.

[0090] Preferably, in step A202, multiple location points can be selected at equal intervals in the direction corresponding to the initial direction angle, based on the number of Kalman filter calculation functions generated and at preset intervals (hereinafter referred to as Method 1).

[0091] Alternatively, based on the number of Kalman filter calculation functions generated and the maximum detection distance of a single base station passive detector, the interval distance is determined, and multiple location points are selected at equal intervals in the direction corresponding to the initial direction angle according to the determined interval distance (hereinafter referred to as Method Two).

[0092] Specifically, the "preset interval distance" refers to a predetermined, fixed distance value used to evenly distribute the various location points along the initial direction angle. This preset value can be set based on prior knowledge of typical target motion patterns, required tracking accuracy, or specific environmental conditions. "Equally spaced selection" means that the selected multiple location points are evenly distributed along the direction corresponding to the initial direction angle, the distance between adjacent location points remains consistent, and the distance between the first location point and the single-base station passive detector is also the preset interval distance. This method ensures systematic coverage of the potential initial target location.

[0093] "Maximum detection range" refers to the limit at which a single-base passive detector can reliably detect targets on the water surface, and is a key parameter defining the operating range of the tracking system. "Determined interval distance" represents a method for dynamically calculating the interval distance between location points. Unlike using fixed preset values, this interval distance is calculated based on the number of Kalman filter calculation functions generated and the maximum detection range of the single-base passive detector. For example, if it is necessary to initialize N Kalman filters (i.e., Kalman filter calculation functions), the maximum detection range can be divided by N-1 (or N) to determine a suitable interval, thereby ensuring that the initial location points can cover the entire detectable range.

[0094] This application's solution significantly enhances the robustness and efficiency of the Kalman filter initialization process by providing two specific strategies for selecting initial position points. When using a preset interval distance, the system benefits from a consistent and predictable distribution of initial assumptions, which is particularly advantageous when the initial target distance uncertainty is relatively stable or known. This method simplifies the initialization process by relying on a fixed, empirically determined interval. On the other hand, by dynamically determining the interval distance based on the number of generated Kalman filter calculation functions and the maximum detection range of a single-base station passive detector, the system can adaptively distribute the initial position points across the entire observable range. This ensures that the initial assumptions effectively cover all possible target distances within the detector's capability range, thereby increasing the likelihood of accurately capturing the target's true initial state, especially when the initial distance uncertainty is large or unknown. Both methods contribute to more efficient initialization, laying a solid foundation for subsequent accurate target tracking.

[0095] The initialization process of the Kalman filter calculation function is significantly improved through the above technical solution. By using a preset interval distance, or dynamically determining the interval based on the detector's maximum detection range and the number of function generators, the initial position points can be distributed in a more systematic and optimized manner. This systematic distribution ensures that the initial assumptions about the target state more accurately represent the actual possible target positions, thereby reducing the initial uncertainty of the target distance. Therefore, the tracking system can converge to the true target state faster and exhibit higher accuracy in the early stages of tracking, especially when the initial distance information is limited or inaccurate. This optimization directly solves the problem of low initialization efficiency or inaccuracy that may exist in ambiguous position point selection methods, thus achieving more reliable and robust passive target tracking performance.

[0096] In practical applications, the motion state of the water surface target and the measurement characteristics of the detector may differ at different tracking stages. If only a single filtering algorithm is used, or if a dynamic selection mechanism is lacking, it may lead to slow convergence speed and insufficient tracking accuracy in the early stages of target tracking, or problems such as tracking divergence when the target maneuvers, thus affecting the overall tracking performance and robustness.

[0097] Therefore, in some preferred embodiments, step A3 includes:

[0098] A301. When a single base station passive detector detects the real-time direction angle of a water surface target each time, obtain the total number of times the target state parameters of each Kalman filter calculation function have been updated, and obtain the number of times the target state parameters have been updated.

[0099] A302. Compare the number of times the target state parameters are updated with a preset second threshold; if the second threshold is less than the first threshold;

[0100] A303. If the number of times the target state parameter is updated is less than the second threshold, then the target state parameter of each Kalman filter calculation function is updated using the strong tracking filter algorithm;

[0101] A304. If the number of times the target state parameters are updated is not less than the second threshold, then the target state parameters of each Kalman filter calculation function are updated using the standard Kalman filter algorithm or the unscented Kalman filter algorithm.

[0102] Specifically, the target state parameter update count refers to the total number of times the target state parameters have been successfully updated since the Kalman filter calculation function was initialized. This count reflects the length of time the target has been tracked or the degree of accumulation of observation data. The preset second threshold is a key parameter used to distinguish the tracking stages, and its value is set to be less than the preset first threshold. The preset first threshold is used to determine when to select the measured target state parameters using the maximum likelihood estimation method. The strong tracking filter algorithm is a filtering algorithm that can adaptively adjust the gain matrix. Its characteristic is that it can effectively suppress model errors and measurement noise, especially suitable for target maneuvering or the initial tracking stage, to improve the robustness and convergence speed of tracking. The standard Kalman filter algorithm is a linear filtering algorithm that can provide optimal estimation under the assumption that the system model and noise statistics are known and satisfy the Gaussian distribution. It can provide high tracking accuracy when the target is in stable motion and a large amount of observation data has accumulated. The unscented Kalman filter (UKF) is an important extension of the Kalman filter, primarily used to solve state estimation problems in nonlinear systems. Unlike the standard Kalman filter, which requires both the system and observation models to be linear, the UKF can more effectively handle nonlinear relationships in the system or observations. If updating the target state parameters involves nonlinear dynamics or nonlinear observations, the UKF is preferred, providing more accurate and reliable estimation results. Strong tracking filtering, the standard Kalman filter, and the UKF are all existing technologies; their computational processes will not be detailed here.

[0103] This application's solution effectively solves the aforementioned problems by introducing the target state parameter update frequency and a second threshold, thus achieving adaptive switching of the filtering algorithm. In the initial stage of target tracking, due to limited observation data, the estimation uncertainty of the target state parameters is high, and because the velocity in the initial target state parameters of the Kalman filter calculation function is usually set to 0, it differs significantly from the actual target velocity. At this time, the target state parameter update frequency is usually less than the preset second threshold, and the system is configured to use a strong tracking filter algorithm to update the target state parameters. The strong tracking filter algorithm, through its adaptive characteristics, can quickly converge and effectively suppress large errors in the initial stage, enhancing the robustness of tracking and avoiding tracking divergence. As the tracking process progresses, observation data gradually accumulates, and the target state parameter update frequency gradually increases. When it is not less than the preset second threshold, it indicates that target tracking has entered a relatively stable stage. At this point, the system is configured to switch to the standard Kalman filter algorithm or the unscented Kalman filter algorithm to update the target state parameters. The standard Kalman filter algorithm or the unscented Kalman filter algorithm can provide higher estimation accuracy in the stable tracking stage, thereby further optimizing the tracking effect. This dynamic switching mechanism ensures that the most suitable filtering strategy is used in different tracking stages, thereby improving the overall tracking performance.

[0104] Through the above technical solution, this application can dynamically adjust the filtering algorithm used according to the number of times the target state parameters are updated, thus effectively solving the problem of poor performance of a single filtering algorithm in different tracking stages in traditional schemes. Specifically, in the early stage of tracking, the use of a strong tracking filtering algorithm can significantly improve the convergence speed and robustness of tracking, effectively dealing with the uncertainty and potential maneuvers in the initial stage. After the tracking enters a stable stage, switching to the standard Kalman filter algorithm or the unscented Kalman filter algorithm can maximize the tracking accuracy and ensure more accurate estimation of the target state parameters. Thus, the solution of this application ensures both tracking robustness and tracking accuracy, enabling the surface target tracking method to exhibit excellent performance in various complex environments, reducing the risk of tracking divergence, and improving the reliability of target tracking.

[0105] Preferably, step A4 includes:

[0106] A401. Perform the following steps for each Kalman filter calculation function:

[0107] B1. Based on the current target state parameters calculated by the Kalman filter calculation function, calculate the coordinates of the water surface target at each observation time point before the current time; the observation time point is the time point when the direction angle of the water surface target is detected by the single base station passive detector.

[0108] B2. Based on the calculated coordinates, calculate the orientation and angle of the water surface target at each observation point before the current time.

[0109] B3. Based on the direction angles of each observation time point obtained by back calculation and the direction angles of each observation time point obtained by the passive detector of a single base station, the likelihood value of each observation time point is calculated using the Gaussian probability density function.

[0110] B4. Calculate the overall likelihood value based on the likelihood values ​​at each observation time point;

[0111] A402. Compare the comprehensive likelihood values ​​corresponding to each Kalman filter calculation function, determine the largest comprehensive likelihood value, and select the target state parameter of the Kalman filter calculation function corresponding to the largest comprehensive likelihood value as the measured target state parameter of the water surface target.

[0112] Specifically, in step A401, for each Kalman filter calculation function, step B1 is first executed, which involves calculating the coordinates of the surface target at each observation time point before the current moment based on the current target state parameters of the Kalman filter calculation function. Here, "observation time" refers to all time points at which the single-base station passive detector detects the directional angle of the surface target, including the time point corresponding to the initial directional angle and the time points corresponding to each subsequent real-time directional angle. The purpose of this step is to retrospectively reconstruct the possible positions of the target at each past observation time based on the estimation results of the current filter.

[0113] Further, in step B2, based on the coordinates calculated in step B1, the orientation angles of the water surface target at each observation time point before the current time are calculated. These calculated orientation angles are the observation angles predicted based on the trajectory estimated by the Kalman filter calculation function.

[0114] Subsequently, in step B3, based on the direction angles at each observation time point obtained from the back-calculation in step B2, and the direction angles at each observation time point actually detected by the single-base station passive detector, the likelihood value for each observation time point is calculated using the Gaussian probability density function. The Gaussian probability density function is used here to quantify the degree of consistency or matching between the direction angle predicted by the Kalman filter calculation function and the actual observed direction angle at each observation time. The closer the predicted angle is to the actual observed angle, the higher the likelihood value.

[0115] In step B4, a composite likelihood value is calculated based on the likelihood values ​​at each observation time point. This composite likelihood value is an overall assessment of how well a single Kalman filter calculation function matches the actual observation data across all historical observation time points.

[0116] Finally, in step A402, the comprehensive likelihood values ​​corresponding to all Kalman filter calculation functions are compared, and the largest comprehensive likelihood value is determined. The target state parameter of the Kalman filter calculation function corresponding to the largest comprehensive likelihood value is selected as the measured target state parameter of the water surface target. This step is based on the maximum likelihood criterion, which considers that the Kalman filter calculation function with the largest comprehensive likelihood value best represents the true motion state of the water surface target.

[0117] The proposed solution effectively addresses the issues of insufficient accuracy and robustness in the selection of the multi-Kalman filter calculation function in the basic scheme by introducing a detailed maximum likelihood estimation step.

[0118] Specifically, in step A401, for each Kalman filter calculation function, the coordinates of the surface target at each observation time point before the current time are first calculated using step B1. This allows each Kalman filter calculation function to construct a complete historical trajectory based on its current estimated state. Subsequently, step B2 converts these calculated coordinates into corresponding direction angles, thus providing a unified benchmark for comparison with actual observation data. Further, step B3 uses the Gaussian probability density function to calculate the likelihood value for each observation time point. This is an effective method for quantifying the degree of agreement between the observations predicted by each Kalman filter calculation function and the actual observations. By considering the statistical characteristics of observation noise, the Gaussian probability density function can objectively evaluate the matching degree of each observation point. Finally, step B4 combines the likelihood values ​​of these individual observation points to form a comprehensive likelihood value, which comprehensively reflects the performance of each Kalman filter calculation function throughout the entire observation history.

[0119] Therefore, in step A402, by comparing the combined likelihood values ​​of all Kalman filter calculation functions and selecting the largest one, this application can select the trajectory that best matches all historical observation data and has the highest probability from multiple possible target trajectories. This evaluation mechanism based on full historical observation data significantly enhances the accuracy and anti-interference capability of target state parameter selection, especially in information-limited scenarios such as passive detection by a single base station, enabling more reliable identification of the true motion pattern of water surface targets.

[0120] Through the above technical solution, this application provides a more accurate and robust mechanism for selecting surface target state parameters. This solution fully utilizes all historical observation data from the initial moment to the current moment, and by performing a comprehensive likelihood assessment on each Kalman filter calculation function, it avoids errors that may arise from relying solely on the latest observation data. Therefore, even in a single-base station passive detection environment, facing challenges such as observation noise and target maneuvering, it can more accurately identify the estimation result that best reflects the true motion state of the surface target from multiple parallel Kalman filter calculation functions, thereby significantly improving the accuracy and reliability of surface target tracking.

[0121] As a specific implementation, assume that a single-base station passive detector detects the direction angle of a water surface target n times over a period of time, corresponding to n observation time points. When the number of times the target state parameters are updated reaches a preset first threshold, M Kalman filter calculation functions run in parallel within the system. At this time, for each Kalman filter calculation function, for example, the k-th Kalman filter calculation function, it first calculates the coordinates of the water surface target at the previous n-1 observation time points based on its current target state parameters (e.g., [x,y,vx,vy]) and the assumption of uniform linear motion (or a more complex motion model). The target coordinates in the target state parameters at the n-th observation time point (i.e., the current time) are then used as the coordinates calculated at the n-th observation time point.

[0122] Subsequently, based on these inversely calculated coordinates, the directional angles of the water surface target relative to the single-base station passive detector at N observation time points are calculated. For example, for the i-th observation time point, the inversely calculated directional angle is... Meanwhile, the actual direction angle detected by the single-base station passive detector at the i-th observation time point is... .

[0123] Next, the likelihood value at each observation time point is calculated using the Gaussian probability density function. .For example, , It is a Gaussian probability density function.

[0124] After obtaining the likelihood values ​​at n observation time points , ..., Then, calculate the comprehensive likelihood value P of the k-th Kalman filter computation function. For example, it can be... to Multiplying these together yields the overall likelihood value P.

[0125] Repeat the above process to calculate the comprehensive likelihood values ​​corresponding to all M Kalman filter calculation functions. Finally, compare these comprehensive likelihood values, select the largest comprehensive likelihood value, and use the corresponding Kalman filter calculation function's current target state parameters as the measured target state parameters of the water surface target.

[0126] Preferably, step B1 may include:

[0127] Based on the Kalman filter calculation function to calculate the current target state parameters, and based on the assumption of uniform linear motion, the coordinates of the water surface target at each observation time point before the current time are calculated.

[0128] Specifically, the uniform linear motion assumption means that when calculating the coordinates of a water surface target at each observation time point before the current moment, it is assumed that the water surface target moves in a straight line at a constant speed and direction between the observation time points. This means that once the coordinates and target speed of the water surface target at a certain moment (such as the current moment) are determined, simple kinematic formulas can be used to推算 the target state parameters forward or backward to obtain its coordinates at other observation time points. This assumption simplifies the description of the motion trajectory and makes the coordinate calculation process more direct and efficient.

[0129] The solution of this application provides a clear and computationally efficient model for calculating the coordinates of a water surface target at each observation time point before the current moment by introducing the uniform linear motion assumption. In practical applications, the motion trajectory of a water surface target may be relatively complex, but for the observation data within a short time interval, the uniform linear motion assumption usually provides a sufficiently accurate approximation. By adopting this simplified model, complex non-linear optimization or iterative calculations can be avoided during the calculation process, thus significantly reducing the consumption of computing resources and computing time. In addition, this assumption makes the calculation process have good interpretability and stability, which helps to ensure the reliability of subsequent likelihood value calculations.

[0130] Through the above technical solution, when calculating the historical coordinates of a water surface target, the calculation model can be effectively simplified and the calculation efficiency can be improved. Compared with using a more complex motion model for calculation, the uniform linear motion assumption significantly reduces the complexity of the algorithm while ensuring a certain accuracy, making the entire target tracking method more feasible in resource-constrained underwater devices. This not only speeds up the screening speed of target state parameters but also improves the system's response ability to real-time tracking of water surface targets.

[0131] In some preferred embodiments, it is assumed that at a certain moment t0, the target state parameters of the Kalman filter calculation function are [x0, y0, vx0, vy0], where x0 and y0 represent the target coordinates, and vx0 and vy0 represent the velocity components of the target speed in the x-axis and y-axis directions. If it is necessary to calculate the coordinates of the water surface target at a previous observation moment t1 (t1 < t0), the calculation can be based on the uniform linear motion assumption. Specifically, the time difference between the observation moment t1 and the current moment t0 is Δt = t0 - t1. Then the coordinates (x1, y1) of the water surface target at the observation moment t1 can be calculated according to the following formula:

[0132] x1 = x0 - vx0 * Δt;

[0133] y1 = y0 - vy0 * Δt;

[0134] In this way, the coordinates of water surface targets at various historical observation points can be calculated quickly and directly, providing basic data for subsequent direction and angle calculations and likelihood calculations.

[0135] Preferably, in step B4, the comprehensive likelihood value can be calculated according to the following formula:

[0136] ;

[0137] Where P is the overall likelihood value, and n is the total number of observation time points. Let be the likelihood value at the i-th observation time point.

[0138] The solution presented in this application effectively addresses the numerical precision issues that may arise from direct multiplication by converting the multiplication operation of likelihood values ​​into the accumulation operation of log-likelihood values. When multiple likelihood values... When probabilities (usually between 0 and 1) are multiplied, the product decreases rapidly, especially when there are many observation points n, easily causing the result to approach zero, thus triggering floating-point underflow. By considering each likelihood value... Take the logarithm and convert it to a negative value (because) (The range is between 0 and 1), and then these log-likelihood values ​​are accumulated to avoid numerical underflow. Logarithmic operations convert multiplication into addition, making the numerical range of the calculation result more stable, thus ensuring the accuracy and reliability of the comprehensive likelihood value calculation.

[0139] By employing the aforementioned technical solution, the numerical underflow problem caused by multiplying multiple low-probability values ​​can be effectively avoided when calculating the overall likelihood value of a surface target, significantly improving the numerical stability and computational accuracy of maximum likelihood estimation. This allows for a more accurate assessment of the likelihood of target state parameters calculated using different Kalman filter functions, enabling more reliable decisions when selecting measured target state parameters, thereby enhancing the overall performance and robustness of surface target tracking.

[0140] Please refer to Figure 2This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device includes a processor 301 and a memory 302. The processor 301 and the memory 302 are interconnected and communicate with each other via a communication bus 303 and / or other forms of connection mechanisms (not shown). The memory 302 stores a computer program executable by the processor 301. When the electronic device is running, the processor 301 executes the computer program to perform the underwater device single-base station passive detection target tracking method in any optional implementation of the above embodiments, to achieve the following functions: generating multiple Kalman filter calculation functions; obtaining the initial value of the surface target detected by the single-base station passive detector. The target state parameters of each Kalman filter calculation function are initialized based on the direction angle and the coordinates of multiple position points spaced apart along the direction corresponding to the initial direction angle. After the target state parameters are initialized, each time the single-base station passive detector detects the real-time direction angle of the water surface target, the target state parameters of each Kalman filter calculation function are updated using a strong tracking filter algorithm or a Kalman filter algorithm based on the real-time direction angle and the number of times the target state parameters are updated. When the number of times the target state parameters are updated reaches a preset first threshold, the target state parameters of one of the Kalman filter calculation functions are selected as the measured target state parameters of the water surface target using the maximum likelihood estimation method.

[0141] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it executes the underwater device single-base station passive detection target tracking method in any optional implementation of the above embodiments to achieve the following functions: generating multiple Kalman filter calculation functions; obtaining the initial direction angle of the water surface target detected by the single-base station passive detector, and initializing the target state parameters of each Kalman filter calculation function based on the coordinates of multiple position points spaced apart in the direction corresponding to the initial direction angle; after completing the initialization of the target state parameters, each time the single-base station passive detector detects the real-time direction angle of the water surface target, updating the target state parameters of each Kalman filter calculation function according to the real-time direction angle and the number of target state parameter updates, using a strong tracking filter algorithm or a Kalman filter algorithm; when the number of target state parameter updates reaches a preset first threshold, selecting the target state parameter of one of the Kalman filter calculation functions as the measured target state parameter of the water surface target through a maximum likelihood estimation method. The computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0142] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. An underwater equipment single base station passive detection target tracking method for tracking a surface target by a single base station passive detector, characterized in that, The method includes the following steps: A1. Generate multiple Kalman filter calculation functions; A2. Obtain the initial orientation angle of the water surface target detected by the single base station passive detector, and initialize the target state parameters of each of the Kalman filter calculation functions based on the coordinates of multiple position points spaced apart in the direction corresponding to the initial orientation angle; A3. After the initialization of the target state parameters is completed, each time the single base station passive detector detects the real-time direction angle of the water surface target, the target state parameters of each of the Kalman filter calculation functions are updated according to the real-time direction angle and the number of times the target state parameters are updated, using a strong tracking filter algorithm or a Kalman filter algorithm. A4. When the number of times the target state parameter is updated reaches a preset first threshold, the target state parameter of one of the Kalman filter calculation functions is selected as the measured target state parameter of the water surface target by the maximum likelihood estimation method. Step A4 includes: A401. Perform the following steps for each Kalman filter calculation function: B1. Based on the current target state parameters calculated by the Kalman filter calculation function, calculate the coordinates of the water surface target at each observation time point before the current time; the observation time point is the time point at which the direction angle of the water surface target is detected by the single base station passive detector. B2. Based on the calculated coordinates, calculate the orientation and angle of the water surface target at each observation point before the current time. B3. Based on the direction angles of each observation time point obtained by back calculation and the direction angles of each observation time point detected by the single base station passive detector, the likelihood value of each observation time point is calculated using the Gaussian probability density function. B4. Calculate the overall likelihood value based on the likelihood values ​​at each observation time point; A402. Compare the comprehensive likelihood values ​​corresponding to each Kalman filter calculation function, determine the largest comprehensive likelihood value, and select the current target state parameter of the Kalman filter calculation function corresponding to the largest comprehensive likelihood value as the measured target state parameter of the water surface target.

2. The underwater equipment single-base station passive target detection and tracking method according to claim 1, characterized in that, Step A1 includes: A101. Obtain information on currently available computing resources; A102. Determine the number of Kalman filter computation functions to be generated based on the available computing resource information; A103. Generate a Kalman filter calculation function based on the number of generated functions.

3. The underwater equipment single-base station passive target detection and tracking method according to claim 1, characterized in that, The target state parameters include target coordinates and target velocity, where the target coordinates represent the coordinates of the water surface target and the target velocity represents the velocity of the water surface target. Step A2 includes: A201. Obtain the initial orientation angle of the water surface target detected by the single-base station passive detector; A202. Based on the number of Kalman filter calculation functions generated, select multiple location points at different distances from the single base station passive detector in the direction corresponding to the initial direction angle, and calculate the coordinates of each location point based on the distance between each location point and the single base station passive detector and the initial direction angle; A203. Assign the coordinates of each of the aforementioned positions to the target coordinates in the target state parameters of each of the aforementioned Kalman filter calculation functions, and set the target velocity in the target state parameters of each of the aforementioned Kalman filter calculation functions to zero.

4. The underwater equipment single-base station passive target detection and tracking method according to claim 3, characterized in that, In step A202, based on the number of Kalman filter calculation functions generated, multiple position points are selected at equal intervals in the direction corresponding to the initial direction angle, with a preset interval distance. Alternatively, based on the number of Kalman filter calculation functions generated and the maximum detection distance of the single base station passive detector, the interval distance is determined, and multiple position points are selected at equal intervals in the direction corresponding to the initial direction angle according to the determined interval distance.

5. The underwater equipment single-base station passive target detection and tracking method according to claim 1, characterized in that, Step A3 includes: A301. When the single base station passive detector detects the real-time direction angle of the water surface target each time, the total number of times the target state parameters of each Kalman filter calculation function are updated is obtained, and the number of times the target state parameters are updated is obtained. A302. Compare the number of times the target state parameter is updated with a preset second threshold; the second threshold is less than the first threshold; A303. If the number of times the target state parameter is updated is less than the second threshold, the target state parameter of each of the Kalman filter calculation functions is updated using the strong tracking filter algorithm; A304. If the number of times the target state parameter is updated is not less than the second threshold, then the target state parameters of each of the Kalman filter calculation functions are updated using the standard Kalman filter algorithm or the unscented Kalman filter algorithm.

6. The underwater equipment single-base station passive target detection and tracking method according to claim 1, characterized in that, Step B1 includes: Based on the current target state parameters calculated by the Kalman filter calculation function, and based on the assumption of uniform linear motion, the coordinates of the water surface target at each observation time point before the current time are calculated.

7. The underwater equipment single-base station passive target detection and tracking method according to claim 1, characterized in that, In step B4, the comprehensive likelihood value is calculated according to the following formula: ； Where P is the overall likelihood value, and n is the total number of observation time points. Let be the likelihood value at the i-th observation time point.

8. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a computer program executable by the processor, and when the processor executes the computer program, it performs the steps of the underwater device single-base station passive detection target tracking method as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it performs the steps of the underwater equipment single-base station passive detection target tracking method as described in any one of claims 1-7.

Images