Multi-target system tracking method based on adaptive optimal control method and Voronoi distribution motion coordination algorithm

Abstract

The invention discloses a method for tracking a multi-target system based on an adaptive optimal control method and a Voronoi distribution motion coordination algorithm, which belongs to the technical field of automatic control and intelligent systems. The method includes the following specific steps: calculating the initial probability density function of the multi-target system and the spatial distribution of dynamic obstacles; obtaining the state evolution path of the multi-target system, which is represented by a time-varying Gaussian mixture model; constructing an obstacle map function including the multi-target system and the dynamic obstacles, and calculating the optimal density path of the system by an adaptive optimal control method. The invention introduces an adaptive optimal control method and a Voronoi distribution motion coordination algorithm, and can accurately describe and predict the state evolution of a large-scale multi-target system in a dynamic environment by constructing a Gaussian mixture model (GMM) and an obstacle map function.

Description

Technical Field

[0001] The present invention relates to the field of automatic control and intelligent system technology, and is specifically a multi-objective system tracking method based on an adaptive optimal control method and a Voronoi distributed motion coordination algorithm. Background Art

[0002] The tracking problem of large-scale multi-target systems is of great significance in the field of modern intelligent systems and automatic control, especially in complex environments full of dynamic obstacles. Traditional centralized control methods are unable to cope with the complexity and real-time requirements of such problems, while multi-agent systems (MAS) provide a potential way to achieve efficient tracking through collaborative control and distributed algorithms.

[0003] In a dynamic obstacle environment, the positions of targets and obstacles in a multi-target system are constantly changing. Traditional path planning methods often rely on pre-set static environmental information and are difficult to adapt to real-time changes. This requires the intelligent system to have real-time perception and decision-making capabilities, and to be able to dynamically adjust path planning and control strategies based on the current environmental state to avoid collisions and path blockages. In addition, as the target density of large-scale multi-target systems changes dynamically, the agent needs to constantly adjust its control strategy to adapt to these changes. For example, in a multi-target system, targets may be dense in some areas, while sparse in other areas. How to effectively perform fine control in dense areas while achieving wide-area coverage in sparse areas has become an important research issue.

[0004] Distributed control algorithms need to avoid conflicts in the actions of individual agents while ensuring the overall coordination of the system, and update control strategies in real time to cope with dynamic changes in the environment and goals. This involves how to efficiently allocate tasks and coordinate movements among agents to achieve global goal tracking and environmental adaptation. Traditional centralized control methods rely on a central control unit to calculate the motion trajectory of each agent through a global optimization algorithm. Although this type of method can obtain a global optimal solution, it has high computational complexity and high requirements for real-time communication and information transmission, making it difficult to adapt to dynamically changing environments.

[0005] In contrast, distributed control methods decompose the global control problem into multiple local problems, and each agent only needs to use local information to make decisions. This method reduces computational complexity and communication requirements and has better scalability and adaptability. Typical methods include distributed control algorithms based on Voronoi diagrams, cooperative control algorithms based on local rules, etc. However, these methods still have certain limitations when dealing with dynamic obstacles and large-scale systems. Specifically, these methods may not be able to quickly adjust the motion trajectory of the agent when faced with real-time changing environments, resulting in insufficient system response speed and flexibility.

[0006] Therefore, a multi-target system tracking method based on an adaptive optimal control method and a Voronoi distribution motion coordination algorithm is invented to introduce an adaptive optimization control method and update the control strategy in real time, so that the intelligent agent can dynamically adjust the motion trajectory according to the current environment and target state, thereby significantly improving the flexibility and response speed of the system. Summary of the invention

[0007] In view of the above and / or existing multi-objective system tracking method based on adaptive optimal control method and Voronoi distributed motion coordination algorithm, the present invention has been proposed.

[0008] Therefore, the object of the present invention is to provide a multi-objective system tracking method based on an adaptive optimal control method and a Voronoi distributed motion coordination algorithm, which can solve the existing problems raised above.

[0009] To solve the above technical problems, according to one aspect of the present invention, the present invention provides the following technical solutions:

[0010] A multi-objective system tracking method based on an adaptive optimal control method and a Voronoi distributed motion coordination algorithm includes specific steps as follows:

[0011] S1: Calculate the initial probability density function and spatial distribution of dynamic obstacles in a multi-objective system;

[0012] S2: Obtain the state evolution path of the multi-objective system, which is represented by the time-varying Gaussian hybrid model;

[0013] S3: Construct an obstacle map function containing a multi-objective system and dynamic obstacles, and calculate the optimal density path of the system through an adaptive optimal control method;

[0014] S4: A Voronoi distribution motion coordination algorithm is used to assign micro-control inputs to each target in the system and distribute micro-control inputs among the targets in the system to ensure collision avoidance while tracking dense paths;

[0015] S5: Simulate the macroscopic state evolution of the system to evaluate the effectiveness of the proposed algorithm.

[0016] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, the specific steps of S1 are as follows:

[0017] S11: The initial probability density function is represented by a Gaussian mixture model. The specific formula is as follows:

[0018]

[0019] Among them, π i is the weight of the i-th Gaussian component, is the probability density function of the i-th Gaussian component, μ i and Σ i are the mean and covariance matrices respectively, and N represents the total number of Gaussian distributions;

[0020] S12: The spatial distribution of dynamic obstacles is obtained through the environment perception system or prior information, and then an obstacle distribution model is constructed, which will be used for subsequent path planning and obstacle avoidance.

[0021] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, wherein: the state evolution path obtained in S2 is specifically:

[0022] It is represented by a time-varying Gaussian mixture model. The specific formula is as follows:

[0023]

[0024] Among them, π i (t), μ i (t) and Σ i (t) represent the weight, mean and covariance matrix at time t respectively.

[0025] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, the obstacle map function constructed in S3 is specifically:

[0026]

[0027] Among them, C(x, y, t) is the obstacle map function, H(x, y) is the obstacle function, x j (t) and y j (t) is the coordinate of the jth obstacle at time t, and h represents the total number of obstacles.

[0028] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, wherein: the optimal density path calculated by the adaptive optimal control method in S3 is specifically:

[0029]

[0030] Among them, x(t) is the system state, x d (t) is the desired state, u(t) is the control input, R(u(t)) is the control cost function, and T is the end time of the optimization process. By solving the optimization problem, an optimal control input function u(t) that varies with time can be obtained in the entire time interval, so that the system state x(t) is close to the desired state x d (t), while minimizing the control cost;

[0031] Among them, the cost function R(u(t)) includes the following forms:

[0032] R(u(t))=u(t) T Ru(t)

[0033] where R is a positive definite matrix that weighs the cost of the control input.

[0034] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, the Voronoi distribution motion coordination algorithm used in S4 is specifically:

[0035] Based on the density in each agent’s Voronoi cell, the agent’s microscopic control input is determined by the control law that moves to the center of mass, specifically:

[0036] u i (t) = k(c i (t)-x i (t))

[0037] The control law is used to determine the control input u of each agent. i (t) rules, u i (t) is the control input of the ith agent, c i (t) is the centroid of the Voronoi cell of the ith agent, x i (t) is the position of the ith agent, and k is the control gain constant, which is used to adjust the amplitude of the control input.

[0038] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, the Voronoi distribution motion coordination algorithm further includes the following steps:

[0039] Calculate the centroid c of each agent's Voronoi cell i (t) and density ρ i (t), specifically:

[0040]

[0041]

[0042] Among them, V i is the Voronoi cell of the ith agent, ρ(x, t) is the density function at position x and time t, which can be approximated using a Gaussian mixture model; ρ i (t) represents the density of the ith agent at time t, N i (t) is the number of the ith agent at time t, A i (t) is the Voronoi cell area of ​​the ith agent.

[0043] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, wherein: the macro state evolution of the simulated system in S5 is specifically:

[0044] The simulation results show the effectiveness of the proposed method in avoiding obstacles and team collisions. The simulation results include the following:

[0045] P = {CAE, PE, AT}

[0046] Among them, P represents system performance, CAE represents collision avoidance effect, PE represents path efficiency, and AT represents the time to reach the target.

[0047] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, the simulation results also show the performance of the method under different environments and obstacle distribution conditions, specifically including:

[0048] S={SOE,DOE,ME}

[0049] Among them, S represents different environmental scenarios, SOE represents static obstacle environment, DOE represents dynamic obstacle environment, and ME represents mixed environment.

[0050] As a preferred solution of the multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm described in the present invention, the multi-target system tracking method can also be extended to different types of multi-target systems, specifically including the following:

[0051] A={RC,AV,SN}

[0052] Among them, A stands for application field, RC stands for robot swarm, AV stands for autonomous vehicle, and SN stands for sensor network.

[0053] Compared with existing technologies:

[0054] The present invention introduces an adaptive optimal control method and a Voronoi distribution motion coordination algorithm. Through the construction of a Gaussian mixture model (GMM) and an obstacle map function, the state evolution of a large-scale multi-objective system in a dynamic environment can be accurately described and predicted. The optimal density path is calculated by an adaptive optimal control method, and combined with the Voronoi distribution motion coordination algorithm, collision avoidance between intelligent agents and obstacles is ensured while achieving efficient tracking. The simulation results verify the effectiveness and adaptability of the proposed method in different environments and conditions, and it has broad application prospects, including but not limited to robot swarms, autonomous vehicles, and sensor networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The flowchart of the method is shown in FIG.

[0056] Figure 2 Schematic diagram of the initial probability density function (IPDF) and obstacle distribution.

[0057] Figure 3 Schematic diagram of the state evolution path of the system, where (a) t = 0, (b) t = 300t, (c) t = 400t, and (d) t = 720t.

[0058] Figure 4 Schematic diagram of the obstacle map function and the optimal density path. The gray area represents the set occupied by dynamic obstacles (a) t = 0, (b) t = 300t, (c) t = 400t, and (d) t = 720t.

[0059] Figure 5 Schematic diagram of simulation results for simulating the macroscopic state evolution of the target system. DETAILED DESCRIPTION

[0060] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0061] The present invention provides a multi-target system tracking method based on an adaptive optimal control method and a Voronoi distribution motion coordination algorithm. Figure 1-Figure 5 , including the following specific steps:

[0062] S1: Calculate the initial probability density function (IPDF) of the multi-target system and the spatial distribution of dynamic obstacles;

[0063] The initial probability density function (IPDF) is represented by a Gaussian mixture model (GMM) as follows:

[0064]

[0065] Among them, π i is the weight of the i-th Gaussian component, is the probability density function of the i-th Gaussian component, μ i and Σ i are the mean and covariance matrices respectively, and N represents the total number of Gaussian distributions.

[0066] The spatial distribution of dynamic obstacles is obtained through the environmental perception system or prior information, and an obstacle distribution model is constructed for subsequent path planning and obstacle avoidance.

[0067] The initial probability density function (IPDF) is used to describe the state distribution of the system at the initial moment. As time goes by, the state evolution path of the system is represented by the Time-Varying Gaussian Mixture Model (TVGMM), which is called the State Evolution Path (SEP). This representation method allows us to dynamically describe the state distribution of the system for better trajectory tracking and obstacle avoidance.

[0068] S2: Obtain the state evolution path of the multi-objective system, where the path is represented by a time-varying Gaussian mixture model (TVGMM);

[0069] The state evolution path is represented by the time-varying Gaussian mixture model (SVGMM), and the formula is as follows:

[0070]

[0071] Among them, π i (t), μ i (t) and ∑ i (t) represent the weight, mean and covariance matrix at time t respectively.

[0072] S3: Construct an obstacle map function including a multi-objective system and dynamic obstacles, and calculate the optimal density path of the system through an adaptive optimal control method;

[0073] The formula for the obstacle map function is as follows:

[0074]

[0075] Among them, C(x, y, t) is the obstacle map function, H(x, y) is the obstacle function, x j (t) and y j (t) is the coordinate of the jth obstacle at time t, and h is the total number of obstacles. The optimal control input is determined by solving the following optimization problem:

[0076]

[0077] Among them, x(t) is the system state, x d (t) is the desired state, u(t) is the control input, R(u(t)) is the control cost function, and T is the end time of the optimization process. By solving this optimization problem, an optimal control input function u(t) that varies with time can be obtained in the entire time interval, so that the system state x(t) is as close as possible to the desired state x d (t), while minimizing the control cost.

[0078] Among them, the cost function R(u(t)) includes the following forms:

[0079] R(u(t))=u(t) T Ru(t)

[0080] where R is a positive definite matrix that weighs the cost of the control input.

[0081] S4: A Voronoi distribution motion coordination algorithm is used to distribute microscopic control inputs among the various goals of the system to ensure collision avoidance while tracking a dense path;

[0082] Based on the density in each agent’s Voronoi cell, the agent’s microscopic control input is determined by the control law that moves to the center of mass, as follows:

[0083] u i (t) = k(c i (t)-x i (t))

[0084] The control law is used to determine the control input u of each agent. i (t) rules, u i (t) is the control input of the ith agent, ci (t) is the centroid of the Voronoi cell of the ith agent, x i (t) is the position of the ith agent, and k is the control gain constant, which is used to adjust the amplitude of the control input.

[0085] The Voronoi distribution motion coordination algorithm also involves calculating the centroid and density of the Voronoi cells for each agent:

[0086]

[0087]

[0088] Among them, V i is the Voronoi cell of the ith agent, ρ(x, t) is the density function at position x and time t, which can be approximated using the Gaussian mixture model (GMM); ρ i (t) represents the density of the ith agent at time t, N i (t) is the number of the ith agent at time t, A i (t) is the Voronoi cell area of ​​the ith agent.

[0089] S5: Simulate the macroscopic state evolution of the system to evaluate the effectiveness of the proposed algorithm.

[0090] The simulation results demonstrate the effectiveness of the proposed method in avoiding obstacles and team collisions. The simulation results include the following:

[0091] P = {CAE, PE, AT}

[0092] Among them, P (Performance) represents system performance, CAE (Collision Avoidance Effect) represents collision avoidance effect, PE (Path Efficiency) represents path efficiency, and AT (Arrival Time) represents the time to arrive at the target.

[0093] The simulation results also show the performance of the method under different environmental and obstacle distribution conditions, including:

[0094] S={SOE,DOE,ME}

[0095] Among them, S (Scenarios) represents different environmental scenarios, SOE (Static Obstacle Environments) represents static obstacle environment, DOE (Dynamic Obstacles Environments) represents dynamic obstacle environment, and ME (Mixed Environments) represents mixed environment.

[0096] The multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm can be extended to different types of multi-target systems, including:

[0097] A={RC,AV,SN}

[0098] Among them, A represents the application field, RC (Robotics Community) represents the robot community, AV (Autonomous Vehicles) represents autonomous vehicles, and SN (Sensor Networks) represents sensor networks.

[0099] In order to verify the prediction accuracy of the adaptive optimal control method and Voronoi distribution motion coordination algorithm (hereinafter referred to as "this method") proposed in this embodiment, it is compared with the existing method. Figure 2 As shown in the schematic diagram of the initial probability density function (IPDF) and the obstacle distribution, under the framework of leave-one-out cross-validation, the area under the curve (AUC) of this method is significantly better than other methods, which fully demonstrates that this method performs well in avoiding obstacles and improving path efficiency, and has higher accuracy and efficiency.

[0100] In a series of simulation experiments, the method of the present invention performs well under different environments and obstacle distribution conditions. The specific experimental settings are as follows:

[0101] 1. Static obstacle environment:

[0102] In a static obstacle environment, the system demonstrates strong obstacle avoidance capabilities and efficient path planning capabilities. Figure 2 Schematic diagram of the initial probability density function (IPDF) and obstacle distribution. The system can effectively avoid static obstacles and has a high path efficiency. The AUC value reaches 0.8757, which is significantly better than other comparison methods (for example, traditional centralized control methods and distributed control methods based on local rules). This shows that the proposed method has high robustness and accuracy in static environments.

[0103] 2. Dynamic obstacle environment:

[0104] In a dynamic obstacle environment, the system needs to adjust the path in real time to avoid constantly moving obstacles. Figure 3The schematic diagram of the state evolution path of the system shows that the method of the present invention performs well, can respond quickly to environmental changes, and adjust the path in time to avoid collisions. Its AUC value is 0.8353±0.0045, which is far higher than other comparison methods (e.g., classic distributed control methods and path planning based on potential field method), which strongly proves the adaptability and real-time performance of the present method in dynamic environments.

[0105] 3. Hybrid Environment:

[0106] In mixed environments (including static and dynamic obstacles), the method of the present invention demonstrates its adaptability and effectiveness in complex environments. The system can effectively avoid collisions and maintain efficient path planning in complex mixed environments. The experimental results show that the method performs well in both static and dynamic obstacle environments, and the AUC values ​​are significantly higher than those of other comparison methods.

[0107] In order to verify the effectiveness of the method of the present invention in practical applications, we conducted tests in different fields such as robot groups, autonomous vehicles, and sensor networks. The specific applications are as follows:

[0108] 1. Application in robot groups:

[0109] Figure 3 It is a schematic diagram of the state evolution path of the system. In the robot group application, each robot works together through the Voronoi distribution motion coordination algorithm, effectively avoiding collisions and completing the designated tasks.

[0110] This method can theoretically achieve efficient teamwork and task allocation.

[0111] 2. Autonomous Vehicle Applications:

[0112] Figure 4 Schematic diagram of obstacle map function and optimal density path. In the application of autonomous vehicles, the method of the present invention can effectively avoid traffic obstacles and optimize the driving path. The distributed control strategy of the present method has the potential to significantly improve driving safety and efficiency, and the autonomous vehicle can drive smoothly in a complex traffic environment.

[0113] 3. Sensor network applications:

[0114] In the sensor network, the present method theoretically adjusts the position of each sensor node through the adaptive optimal control method and the Voronoi distribution motion coordination algorithm, optimizes network coverage, and improves monitoring accuracy and data transmission efficiency. The method of the present invention performs well in improving the coverage and data transmission performance of the sensor network, and has the potential to significantly improve the overall performance of the system.

[0115] The multi-target system tracking method based on adaptive optimal control and Voronoi distribution motion coordination algorithm proposed in the present invention has demonstrated excellent performance in different environments and application fields, and has broad application prospects. In addition, an adaptive optimization control method is introduced in the present invention. This method enables the intelligent agent to dynamically adjust the motion trajectory according to the current environment and target state by updating the control strategy in real time, thereby significantly improving the flexibility and response speed of the system. This control method can not only effectively cope with the dynamic changes of the environment, but also show superior performance when dealing with uncertainty. Compared with the traditional distributed control method, it has obvious innovation.

[0116] In addition, the application of multi-agent systems has shown its unique innovative advantages. Through a reasonably designed collaborative control algorithm, this system can flexibly adjust when the number of agents and the scale of tasks change, and maintain the normal operation of the system when individual agents fail. Its practical application in the fields of drone formations, autonomous driving fleets, and intelligent warehouse robots has further verified its innovation and efficiency, especially in large-scale, multi-target tracking tasks.

[0117] Although the present invention has been described above with reference to the embodiments, various modifications may be made thereto and parts thereof may be replaced by equivalents without departing from the scope of the present invention. In particular, as long as there is no structural conflict, the various features in the embodiments disclosed in the present invention may be used in combination with each other in any manner, and the fact that these combinations are not exhaustively described in this specification is only for the sake of omitting space and saving resources. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A multi-target system tracking method based on an adaptive optimal control method and a Voronoi distribution motion coordination algorithm, characterized in that: The specific steps are as follows: S1: Calculate the initial probability density function of the multi-target system and the spatial distribution of dynamic obstacles; S2: Obtain a state evolution path of the multi-objective system, where the path is represented by a time-varying Gaussian mixture model; S3: Construct an obstacle map function including a multi-objective system and dynamic obstacles, and calculate the optimal density path of the system through an adaptive optimal control method; S4: A Voronoi distribution motion coordination algorithm is used to assign micro-control inputs to each target in the system and distribute micro-control inputs among the targets in the system to ensure collision avoidance while tracking dense paths; S5: simulate the macro state evolution of the system to evaluate the effectiveness of the proposed algorithm; The specific steps of S1 are as follows: S11: The initial probability density function is represented by a Gaussian mixture model. The specific formula is as follows: Among them, π i is the weight of the i-th Gaussian component, is the probability density function of the i-th Gaussian component, μ i and∑ i are the mean and covariance matrices respectively, and N represents the total number of Gaussian distributions; S12: The spatial distribution of dynamic obstacles is obtained through the environment perception system or prior information, and then an obstacle distribution model is constructed, which will be used for subsequent path planning and obstacle avoidance; The state evolution path obtained in S2 is specifically: It is represented by a time-varying Gaussian mixture model. The specific formula is as follows: Among them, π i (t), μ i (t) and ∑ i (t) represents the weight, mean and covariance matrix at time t respectively; The obstacle map function constructed in S3 is specifically: Among them, C(x, y, t) is the obstacle map function, H(x, y) is the obstacle function, x j (t) and y j (t) is the coordinate of the jth obstacle at time t, and h represents the total number of obstacles; The optimal density path calculated by the adaptive optimal control method in S3 is specifically: Among them, x(t) is the system state, x d (t) is the desired state, u(t) is the control input, R(u(t)) is the control cost function, and T is the end time of the optimization process. By solving the optimization problem, an optimal control input function u(t) that varies with time can be obtained in the entire time interval, so that the system state x(t) is close to the desired state x d (t), while minimizing the control cost; Among them, the cost function R(u(t)) includes the following forms: R(u(t))=u(t) T Ru(t) Where R is a positive definite matrix used to weigh the cost of control input; The Voronoi distribution motion coordination algorithm used in S4 is specifically: Based on the density in the Voronoi cell of each agent, the microscopic control input of the agent is determined by the control law moving to the center of mass, specifically: u i (t)=k(c i (t)-x i (t)) The control law is used to determine the control input u of each agent. i (t) rules, u i (t) is the control input of the ith agent, c i (t) is the centroid of the Voronoi cell of the ith agent, x i (t) is the position of the ith agent, k is the control gain constant, which is used to adjust the amplitude of the control input; The Voronoi distribution motion coordination algorithm further comprises the following steps: Calculate the centroid c of each agent's Voronoi cell i (t) and density ρ i (t), specifically: Among them, V i is the Voronoi cell of the ith agent, ρ(x, t) is the density function at position x and time t, which can be approximated using a Gaussian mixture model; ρ i (t) represents the density of the ith agent at time t, N i (t) is the number of the ith agent at time t, A i (t) is the Voronoi cell area of ​​the ith agent.

2. The multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm according to claim 1 is characterized in that: The macro state evolution of the simulation system in S5 is specifically as follows: The simulation results show the effectiveness of the proposed method in avoiding obstacles and team collisions. The simulation results include the following: P={CAE,PE,AT} Among them, P represents system performance, CAE represents collision avoidance effect, PE represents path efficiency, and AT represents the time to reach the target.

3. The multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm according to claim 2 is characterized in that: The simulation results also demonstrate the performance of the method under different environments and obstacle distribution conditions, including: S={SOE,DOE,ME} Among them, S represents different environmental scenarios, SOE represents static obstacle environment, DOE represents dynamic obstacle environment, and ME represents mixed environment.

4. The multi-target system tracking method based on the adaptive optimal control method and the Voronoi distribution motion coordination algorithm according to claim 3 is characterized in that: The multi-target system tracking method according to any one of claims 1 to 3 can also be extended to be applied to different types of multi-target systems, specifically including the following: A={RC,AV,SN} Among them, A stands for application field, RC stands for robot swarm, AV stands for autonomous vehicle, and SN stands for sensor network.

Images