Multi-agent collaborative surrounding tracking method and system for dynamic target

Abstract

The invention provides a multi-agent collaborative surrounding tracking method and system for a dynamic target. The method comprises the steps that S1, the number of surrounding agents, the expected surrounding radius, the expected angular spacing, the expected relative surrounding speed and the communication topology between the agents are determined according to the surrounding requirement; s2, designing a distributed fixed time observer based on a communication topological relation to realize estimation of a target state; s3, acquiring an azimuth angle between the surrounding agent and the target, and constructing a radial unit vector, a tangent vector and an angular spacing coordination error pointing to the target; constructing an estimated azimuth angle change rate, an estimated relative speed, an estimated relative distance and an estimated relative speed tangential component; constructing a radial acceleration control item, a tangential acceleration control item and a dynamic target tracking item; and S4, according to the current linear velocity and course angle of the intelligent agent, calculating linear acceleration and angular velocity in a feedback linearization mode, and realizing multi-layer adjustable angular spacing circular surrounding tracking of the multi-intelligent-agent dynamic target.

Description

Technical Field

[0001] This invention belongs to the field of multi-agent cooperative control technology, specifically relating to a multi-agent cooperative orbiting tracking method and system for dynamic targets. Background Technology

[0002] Cooperative orbiting and tracking refers to a multi-agent system composed of multiple autonomous agents that, through information exchange and cooperation among themselves, orbits and tracks a target while maintaining a certain distance. Currently, the motion models for cooperative orbiting are mainly first-order integral models or nonholonomic motion models with linear velocity and angular velocity as inputs. In this model, the agents control their velocities to maintain a desired angular distance between each agent and the target, while each agent maintains the same distance from the target. To achieve orbiting and tracking of a dynamic target, each agent directly acquires the target's velocity, maintaining a relatively stationary orbiting state between the agent and the target.

[0003] In CN201810757819.0, where all agents know the global position of the moving target, a differential observer is used to estimate the target velocity. Based on this, a cooperative control method for patrol and escort missions of multiple nonholonomic robots based on vector fields is proposed, which enables multiple nonholonomic agents to orbit a moving target with a time-varying velocity within the same radius at the desired circular velocity, while maintaining the desired angular spacing between the agents.

[0004] In “MIAO Z, WANG Y, FIERRO R. Cooperative circumnavigation of a moving target with multiple nonholonomic robots using backstepping design[J].Systems & Control Letters, 2017, 103: 58-65.”, when the agents know the global position and velocity of the moving target, a feedback linearization method is used to enable multiple nonholonomic agents to circle the moving target with a time-varying velocity within the same radius at the desired circumferential velocity, while maintaining the desired angular spacing between the agents.

[0005] Currently, it is generally assumed that all orbiting agents can directly acquire the state information of the moving target. However, due to the constraints of communication resources, it cannot be guaranteed that all orbiting agents can acquire the state information of the moving target in real time. Therefore, how to acquire the target's state information under limited communication resources is an urgent problem to be solved. Meanwhile, existing orbiting control is mostly based on nonholonomic kinematic models with linear velocity and angular velocity as inputs, and research on control design with linear acceleration and angular velocity as inputs is limited. Furthermore, in complex tasks, multi-layer orbiting strategies that can achieve adjustable adjacent angular spacing are more conducive to improving execution efficiency, perception coverage, and fault tolerance, but related distributed control methods are still insufficient. Therefore, it is urgent to propose a distributed cooperative control scheme that can achieve multi-layer circular orbiting with adjustable adjacent angular spacing and stably track dynamic targets for nonholonomic agents with linear acceleration and angular velocity as control inputs, under limited communication resources. Summary of the Invention

[0006] The purpose of this invention is to propose a multi-agent cooperative orbiting tracking method and system for dynamic targets. It proposes a control method for nonholonomic motion models with linear acceleration and angular velocity as inputs. By designing a distributed fixed-time observer, it enables accurate estimation of target state information under limited communication capabilities.

[0007] To achieve the above objectives, the technical solution of the present invention is: a multi-agent cooperative orbiting tracking method for dynamic targets, comprising the following steps:

[0008] S1. Based on nonholonomic constraint multi-agent systems with linear acceleration and angular velocity as control inputs, determine the number of orbiting agents, the desired orbiting radius, the desired angular spacing, the desired relative orbiting velocity, and the communication topology between the agents according to the orbiting requirements.

[0009] S2. Based on the communication topology relationship constructed in step S1, design a distributed fixed-time observer. Under the condition that only some of the surrounding agents obtain the target state, the global position, velocity and acceleration of the target can be accurately estimated within a predetermined time through local information interaction between the surrounding agents.

[0010] S3. Obtain the azimuth angle between the surrounding agent and the target using a monocular camera mounted on the surrounding agent, and construct the radial unit vector, tangential vector, and angular spacing coordination error pointing towards the target; combine the results of steps S1 and S2 to construct the estimated azimuth angle change rate, estimated relative velocity, estimated relative distance, and estimated relative velocity tangential component, and on this basis, construct the radial acceleration control term, tangential acceleration control term, and dynamic target tracking term;

[0011] S4. Based on the current linear velocity and heading angle of the agent, calculate the required linear acceleration and angular velocity through feedback linearization to achieve multi-agent dynamic target multi-layer adjustable angular spacing circular orbit tracking.

[0012] Preferably, the orbiting agent adopts a nonholonomic motion model with linear acceleration and angular velocity as control inputs, as follows:

[0013]

[0014] in For surrounding intelligent agents Position in the global coordinate system The heading angle around the agent. The linear velocity orbiting the agent, and the input orbiting the agent. , For linear acceleration around the intelligent agent, ω represents the angular velocity orbiting the intelligent agent.

[0015] Preferably, the target agent adopts a second-order integral dynamics model, as follows:

[0016]

[0017] in , , These represent the actual position, velocity, and acceleration of the target agent in the global coordinate system, respectively. , , , These represent the position, velocity, and acceleration components of the target agent in the x and y directions in the global coordinate system, respectively.

[0018] Preferably, the communication topology between agents is based on an undirected graph. Describe:

[0019] Node set Let n be a finite non-empty set, and n be the number of surrounding agents. Let be the set of edges. Let be a weighted adjacency matrix; where the edges are... Represents a node With nodes They can communicate with each other; adjacency matrix elements satisfy: ,when , ,otherwise ,for set up ;node The neighbor set is represented as ;

[0020] The communication topology between the agents is represented by the Laplace matrix. express:

[0021] Define the Laplace matrix ,and ,in An undirected graph The degree matrix, the diagonal elements of the degree matrix are ;

[0022] The communication topology between surrounding agents and between surrounding agents and the target is determined by the augmented Laplace matrix. express:

[0023] Define the augmented Laplace matrix , where the target coupling matrix ,element Defined as: if the surrounding intelligent agent If the target's state information can be obtained, then Conversely .

[0024] Preferably, the estimation of the target's global position, velocity, and acceleration in step S2 is as follows:

[0025]

[0026] Where: function x is the function input; , , They represent the surrounding intelligent agents respectively. Estimation of the position, velocity, and acceleration of a moving target. , , Its weight Represents the surrounding intelligent agent Estimation of the position, velocity, and acceleration of a moving target in the x and y directions of the global coordinate system; and All are positive odd numbers, and satisfy the following conditions: By setting parameters and Adjust the upper limit of convergence time. , , This is the observation gain coefficient.

[0027] Preferably, the azimuth angle between the surrounding agent and the target is obtained through a monocular camera mounted on the surrounding agent, and the radial unit vector, tangent vector, and angular spacing error are constructed pointing towards the target, as follows:

[0028] Surrounding intelligent agent Equipped with a monocular camera for observing targets, it surrounds the intelligent agent. Azimuth angle between the target and the target Represented as:

[0029]

[0030] Surrounding intelligent agent unit vector between the target and the target Represented as:

[0031]

[0032] relative position relative distance ;

[0033] Unit vector Orthogonal vector rotated 90° clockwise Represented as:

[0034]

[0035] The angle between the line of sight and the horizontal axis of the coordinate system centered on the target. The following relationship exists between the azimuth and the azimuth:

[0036]

[0037] The angular spacing between adjacent surrounding smart agents Defined in the following form:

[0038]

[0039] Based on the communication topology between agents, an angular spacing coordination error is constructed. :

[0040]

[0041] in, This represents the desired angular spacing between adjacent surrounding agents.

[0042] Preferably, by combining the results of steps S1 and S2, the estimated azimuth rate of change, estimated relative velocity, estimated relative distance, and estimated relative velocity tangential component are constructed, as follows:

[0043] Surrounding intelligent agent Estimated rate of change of azimuth angle between the target and the target for:

[0044]

[0045] For surrounding intelligent agents Estimated distance to the target:

[0046]

[0047] Estimating relative velocity :

[0048]

[0049] Among them, the velocity around the intelligent agent ;

[0050] Estimate the tangential component of relative velocity :

[0051] .

[0052] Preferably, radial acceleration control term, tangential acceleration control term, and dynamic target tracking term are constructed as follows:

[0053]

[0054] in Represents the surrounding intelligent agent In the inertial coordinate system, the desired acceleration vector is... , They are respectively surrounding intelligent agents The components of the desired acceleration vector in the x and y directions in the inertial coordinate system; radial acceleration control quantity. , Indicates the desired orbital radius; tangential acceleration control value. , Indicates the desired relative orbital velocity; , , , It is a positive gain constant.

[0055] Preferably, the calculation of the required linear acceleration and angular velocity is as follows:

[0056] .

[0057] A multi-agent cooperative orbiting tracking system for dynamic targets includes a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs any of the steps in the above-described multi-agent cooperative orbiting tracking method.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] First, compared to traditional cooperative orbiting studies, the controlled agents are usually first-order integral systems or nonholonomic motion models with linear velocity and angular velocity as inputs. This invention targets nonholonomic motion models with linear acceleration and angular velocity as inputs, and adopts a progressive target decomposition and joint control method to achieve multi-layer circular adjustable angular spacing cooperative orbiting tracking of dynamic targets, effectively improving the flexibility, fault tolerance and execution efficiency of the overall system.

[0060] Second, compared to methods that involve communication between the target agent and all agents, this invention, by designing a distributed fixed-time observer, achieves the estimation of the target agent's position, velocity, and acceleration through local information interaction between agents when only some agents obtain the target state information, effectively reducing communication pressure. Attached Figure Description

[0061] Figure 1 This is a control framework diagram of the present invention;

[0062] Figure 2 A schematic diagram of a multi-agent, multi-layered, adjustable-angle-spacing loop orbiting a dynamic target;

[0063] Figure 3 This is a schematic diagram of a multi-agent undirected communication topology.

[0064] Figure 4 A global trajectory map of a dynamic target coordinating multiple agents;

[0065] Figure 5 A multi-agent collaborative orbiting dynamic target relative to the target trajectory diagram;

[0066] Figure 6 The graphs show the norm of the position estimation error for the observer (a), the norm of the velocity estimation error (b), and the norm of the acceleration estimation error (c).

[0067] Figure 7A graph showing the distance error curve for multi-agent cooperative orbiting;

[0068] Figure 8 A graph showing the relative velocity norm of multiple agents cooperating in orbit;

[0069] Figure 9 The diagram shows the angular spacing coordination error curve for multi-agent cooperative orbiting. Detailed Implementation

[0070] The following is in conjunction with the appendix Figure 1-9 The technical solution of the present invention will be described in detail below.

[0071] Existing cooperative orbiting tracking methods suffer from limitations in model input and formation structure: most methods rely on incomplete models with linear and angular velocities as inputs for orbiting control, lacking control design for models with linear acceleration and angular velocity as inputs. This makes it difficult to meet the demands of high dynamic performance or rapid trajectory adjustment scenarios. Typical solutions employ single-layer orbiting with adjustable angular spacing, which is ill-suited to adapting to varying requirements arising from load differences and heterogeneous sensing capabilities in complex tasks. Furthermore, they do not support multi-layer cooperative orbiting with adjustable adjacent angular spacing, resulting in limited system execution efficiency and fault tolerance. In addition, existing cooperative orbiting tracking methods heavily rely on global target information and have insufficient applicability: they generally assume that each agent can acquire the target's global position information in real time. In complex scenarios with limited communication resources and unstable channel quality, these assumptions are difficult to meet, easily leading to a significant decrease in control performance and even the inability to maintain stable orbiting.

[0072] To address the aforementioned shortcomings of existing cooperative orbiting tracking methods, this invention designs a distributed orbiting control law for nonholonomic models with linear acceleration and angular velocity as control inputs. This law enables multiple agents to operate at desired relative speeds within a given orbiting radius while maintaining adjustable desired angular spacing with their neighbors. Furthermore, under conditions of limited communication and when only some orbiting agents acquire the target state, this invention achieves accurate estimation of the target's global position, velocity, and acceleration through local information interaction between orbiting agents. This reduces reliance on global communication and provides usable target state inputs for orbiting control.

[0073] This invention proposes a multi-agent cooperative orbital tracking method for dynamic targets, comprising the following steps:

[0074] a) For nonholonomic constrained multi-agent systems with linear acceleration and angular velocity as control inputs, determine the number of orbiting agents, the desired orbiting radius, the desired angular spacing, the desired relative orbiting velocity, and the communication topology between the agents based on the orbiting requirements. The specific steps are as follows:

[0075] a1) The orbiting agent adopts a nonholonomic motion model with linear acceleration and angular velocity as control inputs, as detailed below:

[0076]

[0077] in For surrounding intelligent agents Position in the global coordinate system The heading angle around the agent. The linear velocity orbiting the agent, and the input orbiting the agent. , For linear acceleration around the intelligent agent, ω represents the angular velocity orbiting the intelligent agent.

[0078] The target agent adopts a second-order integral dynamics model, as follows:

[0079]

[0080] in , , These represent the actual position, velocity, and acceleration of the target agent in the global coordinate system.

[0081] a2) Determine the number of orbiting agents n and the desired orbiting radius based on the orbiting requirements. Expected angular spacing Desired relative orbital velocity .

[0082] The communication topology between agents is an undirected graph. To describe, node set For a finite nonempty set, Let be the set of edges. This is a weighted adjacency matrix. Edges Represents a node With nodes They can communicate with each other; adjacency matrix The elements satisfy: ,when , ,otherwise Assuming , .node The neighbor set can be represented as .

[0083] Define the Laplace matrix ,and ,in An undirected graph The degree matrix, the diagonal elements of the degree matrix are .

[0084] Define the augmented Laplace matrix Target coupling matrix ,element Defined as: if the surrounding intelligent agent If the target's state information can be obtained, then Conversely .

[0085] Figure 3 The communication topology shown is represented by the Laplace matrix. Coupled matrix with target The details are as follows:

[0086]

[0087] ;

[0088] b) Based on the communication topology established in step a), a distributed fixed-time observer is designed. Under the condition that only a portion of the orbiting agents acquire the target state, accurate estimation of the target's global position, velocity, and acceleration is achieved within a predetermined time period through local information exchange between the orbiting agents. The specific calculations are as follows:

[0089]

[0090] Where: function x is the function input; , and They represent the surrounding intelligent agents respectively. Estimation of the position, velocity, and acceleration of a moving target. and All are positive odd numbers, and satisfy the following conditions: By setting parameters and Adjust the upper limit of convergence time. , , This is the observation gain coefficient.

[0091] c) Obtain the azimuth angle between the orbiting agent and the target using a monocular camera mounted on the orbiting agent, and construct the radial unit vector, tangential vector, and angular spacing coordination error pointing towards the target. Combining the results of steps a) and b), construct the estimated azimuth angle change rate, estimated relative velocity, estimated relative distance, and estimated relative velocity tangential component. Based on this, construct the radial acceleration control term, tangential acceleration control term, and dynamic target tracking term. The specific steps are as follows:

[0092] c1) Surrounding the intelligent agent Equipped with a monocular camera for observing targets, such as Figure 2 As shown, surrounding the intelligent agent Azimuth angle between the target and the target It can be represented as:

[0093]

[0094] Surrounding intelligent agent unit vector between the target and the target Represented as:

[0095]

[0096] relative position relative distance ;

[0097] Unit vector Orthogonal vector rotated 90° clockwise Represented as:

[0098]

[0099] The angle between the line of sight and the horizontal axis of the coordinate system centered on the target. The following relationship exists between the azimuth and the azimuth:

[0100]

[0101] The angular spacing between adjacent surrounding smart agents Defined in the following form:

[0102]

[0103] Based on the communication topology between agents, an angular spacing coordination error is constructed. :

[0104]

[0105] c2) Combining the results of steps a) and b), the intelligent agent is... Estimated rate of change of azimuth angle between the target and the target for:

[0106]

[0107] Surrounding intelligent agent Estimated distance to the target for:

[0108]

[0109] Estimating relative velocity :

[0110]

[0111] Among them, the velocity around the intelligent agent .

[0112] Estimate the tangential component of relative velocity :

[0113]

[0114] c3) Multi-agent, multi-layer circular adjustable angular spacing, multi-layer circular dynamic target tracking control law:

[0115]

[0116] in Represents intelligent agents Desired acceleration vector in inertial coordinate system, radial acceleration control quantity Tangential acceleration control quantity , , , , It is a positive gain constant.

[0117] d) Based on the agent's current linear velocity and heading angle, the required linear acceleration and angular velocity are calculated through feedback linearization to achieve multi-agent dynamic target multi-layer adjustable angular spacing circular orbiting tracking. Specifically:

[0118]

[0119] The present invention also proposes a multi-agent cooperative orbiting tracking system for dynamic targets, including a processor, a memory, and a computer program stored in the memory. When the processor executes the computer program, it specifically performs any of the steps in the above-described multi-agent cooperative orbiting tracking method.

[0120] The following simulation presents a multi-agent cooperative orbiting tracking method for dynamic targets, with 6 orbiting agents. To avoid the initial linear velocity being zero or very small, an initial linear velocity is set... initial angular velocity Duration .

[0121] 1) Determine the parameters for multi-circle wrapping as shown in Table 1.

[0122] Table 1. Parameters for Multi-Circle Wrapping

[0123]

[0124] 2) The initial position and heading angle of the agent are shown in Table 2.

[0125] Table 2 Initial position and heading angle of the agent

[0126]

[0127] 3) Initial position of the target agent initial velocity acceleration .

[0128] 4) The initial estimates of the fixed-time observer are shown in Table 3.

[0129] Table 3 Initial estimates for fixed-time observers

[0130]

[0131] Fixed-time observer observation gain , , , , .

[0132] 5) Control gain set to , , , .

[0133] 6) The global motion trajectories of the six agents and the target are as follows: Figure 4 As shown, the trajectories of the six agents relative to the target are as follows: Figure 5 As shown, the movement trajectory of the agent demonstrates that the agent can quickly form the required circular formation. Figure 6 (a)–6(c) represent the target location estimation error norms, respectively. Velocity estimation error norm With acceleration estimation error norm The curves change over time, and each curve rapidly decays to near zero, indicating that the distributed fixed-time observer can accurately estimate the target state. Figure 7 Distance error curve ,when Converging to 0, the formation radius reaches the expected value. . Figure 8 Relative velocity norm curve ,exist At that time, agents 1, 3, and 5 converged to 2.0 m / s, and agents 2, 4, and 6 converged to 1.75 m / s, achieving the expected velocity distribution. Figure 9 The curve of adjacent angular spacing error ,exist Converging to 0, and Figure 8 The fact that the relative velocities are stable at the same time indicates that the angular spacing between adjacent vertices is synchronously achieved after the velocity distribution is completed.

[0134] In summary, unlike current nonholonomic motion models that use linear velocity and angular velocity as inputs, this invention addresses nonholonomic motion models that use linear acceleration and angular velocity as inputs. The proposed control method enables multi-loop cooperative orbiting with adjustable adjacent angular spacing among multiple agents, effectively improving the overall system's execution efficiency and fault tolerance. Unlike other inventions where all agents can directly obtain the target's state information, this invention designs distributed fixed-time observers. Even when only some agents can obtain the target's state information, accurate estimation of the target's state information is achieved through information interaction between orbiting agents.

[0135] The above are preferred embodiments of the present invention. Any changes made to the technical solution of the present invention that do not exceed the scope of the technical solution of the present invention shall fall within the protection scope of the present invention.

Claims

1. A dynamic target-oriented multi-agent cooperative circling tracking method, characterized in that, The method comprises the following steps: S1, based on the non-holonomic constraint multi-agent with linear acceleration and angular velocity as control input, determining the number of surrounding agents, the expected surrounding radius, the expected angular interval, the expected relative surrounding speed and the communication topology between agents according to the surrounding requirement; S2, designing a distributed fixed-time observer based on the communication topology relationship constructed in step S1, under the condition that only part of the surrounding agents obtain the target state, realizing accurate estimation of the global position, velocity and acceleration of the target within a predetermined time through local information interaction between the surrounding agents; S3, obtaining the azimuth angle between the surrounding agent and the target through the monocular camera carried by the surrounding agent, constructing the radial unit vector pointing to the target, the tangent vector, the angular interval coordination error, combining the results of steps S1 and S2 to construct the estimated azimuth angle rate, the estimated relative speed, the estimated relative distance and the estimated relative speed tangent component, and on this basis, constructing the radial acceleration control item, the tangent acceleration control item and the dynamic target tracking item; S4, according to the current linear velocity and heading angle of the agent, calculating the required linear acceleration and angular velocity through feedback linearization, realizing multi-layer adjustable angular interval circular surrounding tracking of the multi-agent dynamic target.

2. The multi-agent cooperative circling tracking method for dynamic target according to claim 1, characterized in that, The surrounding agent adopts a non-holonomic motion model with linear acceleration and angular velocity as control input, which is as follows: wherein is a surrounding agent is a position in a global coordinate system, is a heading angle of the surrounding agent, is a linear velocity of the surrounding agent, the input of the surrounding agent , is a linear acceleration of the surrounding agent, is an angular velocity of the surrounding agent.

3. The multi-agent cooperative circling tracking method for dynamic target of claim 2, wherein, The target agent adopts a second-order integral dynamic model, which is as follows: wherein , , respectively represent the actual position, velocity and acceleration of the target agent in the global coordinate system, , , wherein respectively represent the components of the position, velocity and acceleration of the target agent in the global coordinate system in x and y directions.

4. The multi-agent cooperative circling tracking method for dynamic target according to claim 3, characterized in that, The communication topology between the agents is based on an undirected graph Description is made: Node set is a finite non-empty set, n is the number of agents surrounding the agent, is an edge set, is a weighted adjacency matrix; wherein, edge represents node can communicate with node , the element of the adjacency matrix satisfies: , when , , otherwise , for , set ; the neighbor set of node is denoted as ; The communication topology relationship among the agents is represented by a Laplacian matrix : Definition of Laplacian matrix , and where is the degree matrix of the undirected graph , the diagonal elements of the degree matrix being ; The communication topology relationship among the agents, between the agents and the target is represented by an augmented Laplacian matrix denotes: Defining an augmented Laplacian matrix where the target coupling matrix , element is defined as: if the agent can obtain state information of the target, otherwise .

5. The multi-agent cooperative circling tracking method for dynamic target according to claim 4, characterized in that, The estimation of the global position, velocity and acceleration of the target in step S2 is as follows: where the function , x is the function input; , , respectively represent the estimates of the position, velocity and acceleration of the moving target around the agent , , , , the components represent the estimates of the position, velocity and acceleration of the moving target in the global coordinate system x, y direction around the agent ; and are positive odd numbers, and satisfy , by setting the parameters and to adjust the upper limit of the convergence time, , , are the observation gain coefficients.

6. The multi-agent cooperative circling tracking method for dynamic target according to claim 5, characterized in that, The azimuth angle between the surrounding agent and the target is obtained through the monocular camera carried by the surrounding agent, and the radial unit vector pointing to the target, the tangent vector and the angular interval coordination error are constructed, which are as follows: Surrounding agent Equipped with a monocular camera for observing a target, the surrounding agent Azimuth angle between the surrounding agent Is expressed as: Surrounding agent Unit vector between target Is represented as: relative position , relative distance ; unit vector orthogonal vector rotated 90° clockwise is represented as: Surrounding agent the angle between the agent's line of sight and the horizontal axis of the target-centric coordinate system has the following relationship with the azimuth angle: The angular spacing between adjacent surrounding agents is defined as is defined as: Based on the communication topology between the agents, construct the inter-angle spacing coordination error : wherein, denotes the desired angular separation between adjacent surrounding agents.

7. The multi-agent cooperative circling tracking method for dynamic target according to claim 6, characterized in that, The results of steps S1 and S2 are combined to construct the estimated azimuth angle rate, the estimated relative speed, the estimated relative distance and the estimated relative speed tangent component, which are as follows: Surrounding agent Rate of change of estimated azimuth between the target is: for the agent estimated distance between the agent and the target: Estimating relative speed : where the surrounding agent velocity ; Estimating a relative speed tangential component : 。 8. The multi-agent cooperative circling tracking method for dynamic target according to claim 7, characterized in that, The radial acceleration control item, the tangent acceleration control item and the dynamic target tracking item are constructed, which are as follows: wherein denotes the circumnavigation of the agent the desired acceleration vector in the inertial coordinate system, , denotes the circumnavigation of the agent the components of the desired acceleration vector in the inertial coordinate system in the x, y direction; the radial acceleration control variable , denotes the desired circumnavigation radius; the tangential acceleration control variable , denotes the desired relative circumnavigation speed; , , , is a positive gain constant.

9. The multi-agent cooperative circling tracking method for dynamic target according to claim 8, characterized in that, The calculation of the required linear acceleration and angular velocity is as follows: 。 10. A dynamic target oriented multi-agent cooperative circling tracking system, characterized in that, The method comprises a processor, a memory and a computer program stored in the memory, wherein when the processor executes the computer program, the steps in the multi-agent cooperative surrounding tracking method of any one of claims 1-9 are executed.

Images