An automatic target tracking task allocation method for a lightning-optical linkage system

Abstract

This invention discloses a method for automatically allocating target tracking tasks for a lightning-lightning linkage system. This method constructs and trains a target detection probability model based on a BP neural network. The trained model is used to determine the detection probability of each target by each photoelectric detection device in the lightning-lightning linkage system. A particle swarm optimization algorithm is then used to solve the optimal weighted sum of the detection probabilities of the photoelectric detection devices for the targets. The resulting photoelectric device allocation scheme is then used to track the targets. This method comprehensively considers target type, location, threat level, speed, and the task status of the photoelectric detection devices to achieve adaptive tracking task allocation, maximizing detection efficiency.

Description

Technical Field

[0001] The present application belongs to the technical field of target tracking task allocation, and in particular relates to a method for automatically allocating target tracking tasks for a lightning linkage system. Background Art

[0002] Currently, radar and photoelectric detection equipment are widely used in perimeter intrusion detection systems. Radar, as an active radio detection device, operates in all weather conditions and around the clock, but suffers from the inability to image targets and lacks visual display capabilities. Photoelectric detection equipment has the advantage of being able to image, but is significantly affected by weather. A radar-photoelectric detection equipment linkage system can effectively combine the advantages of both to remedy these issues. Currently, radar-photoelectric detection equipment linkage systems mostly employ a radar-led detection mode with photoelectric detection equipment assisting in tracking. Specifically, when the radar searches for and detects a target, it reports this to the information processing system. The information processing system then sends the target guidance information to the photoelectric detection device's detection unit, which then searches, tracks, and identifies the target and reports the target tracking information. The information processing system then fuses the information from both and displays it to the user.

[0003] Since one photoelectric detection device can only track one target at a time, when multiple targets invade and are detected by the radar, target guidance information needs to be sent to different photoelectric detection devices. Therefore, how to realize the target tracking task allocation of photoelectric detection devices is one of the important links in the linkage of radar photoelectric detection devices.

[0004] Traditional target tracking task allocation methods fall into two categories: purely manual and fixed-strategy. Manual selection involves manually selecting the optoelectronic detection devices to be linked, issuing target guidance information, and making manual adjustments based on tracking performance. Fixed-strategy automatic issuance of target guidance information is achieved based on pre-determined linkage rules.

[0005] However, purely manual coordination of radar and optoelectronic detection equipment is highly arbitrary, resulting in unscientific and irrational allocations, slow adjustments, and low efficiency, making optimal allocation impossible. Fixed strategies, which predetermine linkage rules and cannot be adjusted based on the actual target or equipment performance, also fail to achieve optimal allocation. Summary of the Invention

[0006] The purpose of this application is to provide a method for automatically allocating target tracking tasks for a lightning-light linkage system, so as to solve the problem that the target tracking tasks of the photoelectric detection equipment in the lightning-light joint tracking system cannot be optimally allocated.

[0007] In order to achieve the above objectives, the technical solutions of this application are as follows:

[0008] A method for automatically allocating target tracking tasks for a lightning-light linkage system, comprising:

[0009] Construct and train a target detection probability model based on a BP neural network, wherein the input of the target detection probability model is the target type, direction, distance and speed, and the output is the detection probability of the photoelectric detection device to the target;

[0010] Using the trained target detection probability model, the detection probability of each target by each photoelectric detection device in the lightning-light linkage system is obtained;

[0011] The particle swarm optimization algorithm is used to solve the optimal optimization problem of the weighted sum of the detection probability of the photoelectric detection equipment to the target, and the target is tracked using the solved photoelectric detection equipment allocation scheme.

[0012] Furthermore, the target detection probability model includes an input layer, a hidden layer and an output layer. The input layer has four neurons, corresponding to the target type, speed, direction and distance respectively. The direction is the angle of the target relative to the photoelectric detection device, and the distance is the distance of the target relative to the photoelectric detection device.

[0013] Furthermore, the trained target detection probability model is used to obtain the detection probability of each target by each photoelectric detection device in the lightning linkage system, including:

[0014] Obtain target number, type, position and speed information detected by the radar, and determine target type and speed;

[0015] Calculate the target's direction and distance information relative to the photoelectric detection equipment based on the position information;

[0016] The target type, speed, direction and distance are input into the trained target detection probability model, and the detection probability of the target by the photoelectric detection device is output.

[0017] Furthermore, the fitness function of the optimization problem of the weighted sum of the detection probability of the target by the photoelectric detection device is:

[0018]

[0019] Where n represents the number of targets, W i is the weight of the target, P Xi It represents the detection probability of the Xth photoelectric detection device to the i-th target, and P is the fitness value.

[0020] Furthermore, the W i The threat level of the target.

[0021] Furthermore, the threat level of the target is obtained by radar monitoring of the lightning linkage system.

[0022] This application proposes a method for automatically allocating target tracking tasks for a lightning-light linkage system. It establishes a detection probability model based on target type, speed, direction, and distance, and designs a method for automatically allocating tracking tasks for photoelectric detection equipment based on a particle swarm algorithm. It achieves the best target detection and tracking effects as a whole, achieves the maximum detection efficiency, and solves the problem that target tracking tasks of photoelectric detection equipment in a lightning-light joint tracking system cannot be optimally allocated. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a flow chart of the method for automatically allocating target tracking tasks to the lightning linkage system in this application;

[0024] Figure 2 This is a schematic diagram of the target detection probability model structure of an embodiment of the present application. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0026] The target detection probability of the photoelectric detection device is related to the type, position, distance and speed of the target. This application first fits the curve of the target detection probability and the target type, position, distance and speed based on the BP neural network, and obtains the target detection probability model as P=F(C,V,A,L). Where C is the target type, V is the speed of the target, A is the position of the target relative to the photoelectric detection device, and L is the Euclidean distance of the target relative to the photoelectric detection device. Secondly, the calculated target detection probability is used as the basis for calculating the fitness function in the particle swarm algorithm, and the allocation of the photoelectric detection device tracking task is completed based on the particle swarm algorithm.

[0027] In one embodiment, Figure 1 As shown, a method for automatically allocating target tracking tasks for a lightning linkage system is provided, comprising:

[0028] Step S1: construct and train a target detection probability model based on a BP neural network, wherein the input of the target detection probability model is the target type, direction, distance and speed, and the output is the detection probability of the photoelectric detection device to the target.

[0029] The detection probability of the photoelectric detection equipment for different types of targets (people, large vehicles, small vehicles) at different positions, distances, and speeds was collected. 70% of the data was randomly selected as training samples, and the rest as test samples.

[0030] A three-layer BP neural network is used as the target detection probability model, such as Figure 2 As shown in the figure, it includes an input layer, a hidden layer, and an output layer. The input layer has four neurons, representing the target's type, speed, direction, and distance. The output vector is one, representing the probability of the photoelectric detection device detecting the target. The number of neurons in the hidden layer, k, is determined using the empirical formula k = 2h + 1, where h is the number of input neurons.

[0031] The target detection probability model is trained using training samples. To ensure that different input vectors have the same weight, the speed, direction, and distance data in the samples need to be normalized before training. The normalization formula is as follows:

[0032]

[0033] Where v is the target speed, v max is the maximum value of the target speed in all samples, A is the angle of the target relative to the photoelectric detection device, L is the distance of the target relative to the photoelectric detection device, and L max It is the maximum detection distance of the photoelectric detection equipment.

[0034] After training, the target detection probability model is tested using the test samples. If the target detection probability model test result is not good, the training samples are continued to be used for training until a better target detection probability model is obtained.

[0035] Step S2: using the trained target detection probability model, obtain the detection probability of each photoelectric detection device in the lightning-light linkage system for each target.

[0036] In a radar-assisted system, radar can monitor and obtain information such as the target's number, type, location (longitude, latitude, altitude), speed, and threat level. This information can then be used to determine the target's type and speed, and to calculate the target's position and distance relative to each photoelectric detection device. This information is fed into a trained target detection probability model to determine the detection probability of each photoelectric detection device for each target.

[0037] In this embodiment, the trained target detection probability model is used to obtain the detection probability of each target by each photoelectric detection device in the lightning-light linkage system, including:

[0038] Obtain target number, type, position and speed information detected by the radar, and determine target type and speed;

[0039] Calculate the target's direction and distance information relative to the photoelectric detection equipment based on the position information;

[0040] The target type, speed, direction and distance are input into the trained target detection probability model, and the detection probability of the target by the photoelectric detection device is output.

[0041] It's easy to understand that in a radar-assisted system, the positions of each optoelectronic device are known. Target type, speed, direction, and distance are derived from target information acquired by the radar. This is a relatively mature technology in the field and will not be elaborated on here. For multiple optoelectronic detection devices and multiple targets, the detection probability of each target is calculated for each device separately.

[0042] Step S3: using a particle swarm optimization algorithm to solve the optimal optimization problem of the weighted sum of the detection probabilities of the photoelectric detection equipment for the target, and using the solved photoelectric detection equipment allocation scheme to track the target.

[0043] The particle swarm optimization (PSO) algorithm is a global optimization algorithm based on swarm intelligence. It was first proposed by Kennedy and Eberhart in 1995, stemming from research on the feeding behavior of bird flocks. The basic idea behind PSO is to obtain optimal particles by using the potential solutions to evolve at a random rate through collaboration and information sharing among individuals in the swarm. PSO's advantages lie in its simplicity and ease of implementation, as well as the lack of numerous parameter adjustments. It has been widely used in function optimization, neural network training, fuzzy system control, and other applications of genetic algorithms.

[0044] In this application, the principle of allocating tracking tasks of photoelectric detection equipment is to comprehensively consider the number of photoelectric detection equipment and the threat level of the target, so as to maximize the detection power of intrusion targets.

[0045] For example, given m photoelectric detection devices and n intrusion targets, m ≥ n, the decision matrix for photoelectric detection device task allocation is: X = [X i ]. Where X i (i=1, 2, ... n) represents the serial number of the photoelectric detection device tracking the i-th target.

[0046] This embodiment P Xi W represents the detection probability of the Xth photoelectric detection device to the i-th target, i is the weight of target i. The optimal allocation of photoelectric detection equipment requires that the expected probability of target detection reaches the maximum value. The fitness function of the established optimization model is as follows:

[0047]

[0048] Where n represents the number of targets, W i is the weight of the target, P XiIt represents the detection probability of the Xth photoelectric detection device to the i-th target, and P is the fitness value.

[0049] Preferably, the weight W of the above objectives is i The threat level of the target can be directly adopted, and the threat level of the target is obtained by radar monitoring of the lightning linkage system.

[0050] The goal of the above optimization model is to optimize the weighted sum of the detection probabilities of the photoelectric detection equipment for the target.

[0051] This embodiment uses a particle swarm optimization algorithm to solve the above optimization problem and obtain the optimal photoelectric detection equipment allocation solution.

[0052] Specifically, solving the above process includes:

[0053] 1) Calculate the distance of each target relative to each photoelectric detection device, and select the number of photoelectric detection devices that currently have no detection and tracking tasks and whose targets are within their detection range (m).

[0054] 2) The initial population size is Z is a positive integer greater than or equal to 2, and the target optimization function is determined as formula (4), which is used as the fitness function. It is a calculation method for permutations and combinations.

[0055] 3) Randomly generate a target photoelectric detection device allocation scheme as an individual, and substitute the detection probability of each photoelectric detection device into formula (4) to calculate the individual fitness.

[0056] 4) Calculate the individual fitness extreme value and the group fitness extreme value and update them.

[0057] 5) Update particle velocity and position through particle swarm iteration formula.

[0058] 6) Determine whether the termination condition is met, such as if the error between two consecutive results is less than 0.01, or if the number of iterations has been reached. If the termination condition is met, output the allocation result corresponding to the optimal solution; otherwise, go to step 4).

[0059] It should be noted that the particle swarm optimization algorithm is already a relatively mature technology and will not be described in detail here.

[0060] After obtaining the optimal photoelectric detection device allocation scheme through the particle swarm optimization algorithm, target guidance information is generated and sent to the corresponding photoelectric detection device. The photoelectric detection device that receives the target guidance information tracks the assigned target according to the assigned target to achieve the optimal tracking result.

[0061] Compared with the fixed strategy method, this application establishes a quantifiable detection efficiency calculation method, and can comprehensively consider the target type, location, threat level, speed and task status of the photoelectric detection equipment to achieve adaptive allocation of tracking tasks, so that the detection efficiency reaches the maximum value.

[0062] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for automatically allocating target tracking tasks for a lightning linkage system, characterized in that: The method for automatically allocating target tracking tasks for a lightning linkage system includes: Construct and train a target detection probability model based on a BP neural network, wherein the input of the target detection probability model is the target type, direction, distance and speed, and the output is the detection probability of the photoelectric detection device to the target; Obtaining the target number, type, position, and speed information monitored by the radar, determining the target type and speed, calculating the target's azimuth and distance relative to the photoelectric detection device based on the position information, inputting the target type, speed, azimuth, and distance into a trained target detection probability model, and obtaining the detection probability of each target by each photoelectric detection device in the radar linkage system; The particle swarm optimization algorithm is used to solve the optimal optimization problem of the weighted sum of the detection probability of the photoelectric detection equipment to the target, and the target is tracked using the solved photoelectric detection equipment allocation scheme; The target detection probability model includes an input layer, a hidden layer, and an output layer. The input layer has four neurons, corresponding to target type, speed, direction, and distance, respectively. The direction is the angle of the target relative to the photoelectric detection device, and the distance is the distance of the target relative to the photoelectric detection device. The fitness function of the optimization problem of the weighted sum of the detection probability of the target by the photoelectric detection device is: Where n represents the number of targets, W i is the weight of the target, P Xi It represents the detection probability of the Xth photoelectric detection device to the i-th target, and P is the fitness value.

2. The method for automatically allocating target tracking tasks for a lightning-light linkage system according to claim 1, characterized in that: The W i The threat level of the target.

3. The method for automatically allocating target tracking tasks for a lightning-light linkage system according to claim 2, characterized in that: The threat level of the target is obtained by radar monitoring of the lightning linkage system.

Images