A distributed formation control method for multi-agent swarm

Abstract

The present invention belongs to the field of intelligent flight control technology, and proposes a multi-agent cluster distributed formation control method, including: obtaining a target formation and expressing it as a group of grid points; each intelligent agent receives the target formation, and each intelligent agent negotiates the target formation parameters with its neighboring intelligent agents to obtain unified target formation parameters; based on the hop count algorithm of distributed network propagation, the light intensity coefficient of the grid points within the perception range of each intelligent agent is calculated, and the light intensity of the grid points within the perception range of each intelligent agent is determined; each intelligent agent selects a target grid point set according to the light intensity; the target speed is determined according to the current position, current speed, current position, current speed of neighboring intelligent agents, and target grid points of each intelligent agent; and the movement of each intelligent agent is controlled according to the target speed of each intelligent agent. The multi-agent cluster distributed formation control method of the present invention has strong flexibility, robustness and scalability, and realizes formation control in a fully distributed manner.

Description

Technical Field

[0001] The present invention relates to the field of intelligent body flight control technology, and in particular to a multi-agent cluster distributed formation control method. Background Art

[0002] Distributed formation control of multi-agent swarms refers to the formation, maintenance or reconstruction of a specific formation in a fully distributed manner through a group of agents. Each agent only relies on local observations and negotiates with its neighbors to move. Distributed formation control of multi-agent swarms has broad application prospects in many fields such as unmanned systems, logistics, and sensor networks.

[0003] In the field of unmanned systems, such as drone swarms and unmanned vehicle fleets, multi-agent cluster distributed formation control is a key technology for achieving collaborative combat, reconnaissance, transportation and other tasks. Through precise formation control and coordinated cooperation, the overall effectiveness and mission success rate of unmanned systems can be improved.

[0004] In the field of logistics, multi-agent cluster distributed formation control is widely used in industrial automation, intelligent manufacturing, service robots, etc. For example, in an automated warehouse, by controlling the formation of a robot team, tasks such as cargo handling, sorting, and stacking can be completed efficiently.

[0005] In sensor networks, network nodes usually need to cover a specific area to perform tasks such as environmental monitoring and data collection. Multi-agent cluster distributed formation control can optimize the layout and movement strategy of sensors, improve monitoring efficiency and data collection accuracy. Especially in complex environments, through flexible formation control and coordinated cooperation, comprehensive coverage and monitoring of the target area can be achieved.

[0006] In the field of aerospace, multi-agent cluster distributed formation control is used to achieve tasks such as satellite formation flight and drone group reconnaissance. Through collaborative control and optimized formation, the operating efficiency and mission execution capability of spacecraft can be improved. In addition, this technology can also be used for tasks such as space exploration, rescue and maintenance, providing strong support for the development of aerospace industry.

[0007] In the field of transportation, multi-agent swarm distributed formation control can realize the coordinated dispatch and management of traffic elements such as vehicles and pedestrians. By forming a specific formation, it can improve traffic flow and safety and reduce traffic congestion and accidents.

[0008] At present, the distributed formation control methods of multi-agent clusters used in the above fields mainly have the following problems:

[0009] In the prior art, agents are often heterogeneous, that is, there are global labels between them, and each agent has a fixed position and role in the formation. Although this strategy can achieve formation control to a certain extent, it lacks flexibility and scalability. For example, when an agent fails or cannot perform a task, the entire formation may be affected or even unable to continue to maintain.

[0010] In the prior art, formation description usually relies on the global coordinate system, which means that the agent needs to know the global information of the entire cluster to determine its position in the formation. This not only increases the communication overhead, but also reduces the robustness of the system. Once there is an error or delay in the global information, it may lead to the failure of formation control.

[0011] In the existing technology, when realizing formation transformation, it is usually necessary to pre-assign the target position to the intelligent agent, and realize the adjustment and reconstruction of the formation through centralized control. This method not only limits the autonomy and self-organization of the intelligent agent, but also may lead to insufficient adaptability in complex environments. Summary of the invention

[0012] In view of the deficiencies in the prior art, the present invention proposes a multi-agent cluster distributed formation control method, which has strong flexibility, robustness and scalability and can realize formation control in a fully distributed manner.

[0013] The technical solution of the present invention is as follows:

[0014] A multi-agent cluster distributed formation control method comprises the following steps:

[0015] Step S1: Obtain the target formation and represent it as a set of grid points;

[0016] Step S2: Each agent receives the target formation, and each agent negotiates the target formation parameters with its neighboring agents to obtain unified target formation parameters;

[0017] Step S3: Based on the hop count algorithm of distributed network propagation, the light intensity coefficient of the grid points within the perception range of each intelligent agent is calculated, and the light intensity of the grid points within the perception range of each intelligent agent is determined;

[0018] Step S4: each agent selects a target grid point set according to the light intensity;

[0019] Step S5: Determine the target speed of each agent according to the current position, current speed, current position and current speed of neighboring agents and the target grid point of each agent;

[0020] Step S6: Control the movement of each agent according to the target speed of each agent to achieve distributed formation control of multi-agent cluster.

[0021] Preferably, the step S1 specifically includes:

[0022] Step S1-1: Generate a target formation image manually or automatically through a program;

[0023] Step S1-2: converting the target formation image into a binary image, wherein black pixels represent the target formation;

[0024] Step S1-3: Calculate the number of black pixels in the binary image m and the number of agents n The ratio between r ;

[0025] Step S1-3: Use n Agent coverage m black pixels, that is, each agent covers r black pixels, thus we get n Grid points.

[0026] Preferably, the step S2 specifically includes:

[0027] Step S2-1: Each agent receives a target formation and determines the target formation parameters of each agent, wherein the target formation parameters include the target formation position, the target formation size and the target formation direction;

[0028] Step S2-2: Each agent negotiates the target formation parameters with its neighboring agents, and the negotiation protocol includes the target formation position change rate, the target formation size change rate, and the target formation direction change rate;

[0029] Step S2-3: Obtain unified target formation parameters.

[0030] Preferably, the step S3 specifically includes:

[0031] Step S3-1: Determine whether there are unoccupied grid points within the perception range of each intelligent agent, and calculate the hop count of each intelligent agent using a hop count algorithm of distributed network propagation according to the determination result;

[0032] Step S3-2: determining the illumination intensity coefficient of the grid points within the sensing range of each agent based on the hop count of each agent;

[0033] Step S3-3: Each agent regards all its neighboring agents as point light sources, and the sum of the light intensities emitted by all point light sources at the grid point is the light intensity of the grid point within the agent's perception range.

[0034] Preferably, the step S4 specifically includes:

[0035] Determine the target grid point set of each agent based on whether there are grid points within the perception range of each agent :

[0036] (1) When the agent is outside the target formation and there is no grid point within the range, it selects the unoccupied grid point closest to itself as the target grid point;

[0037] (2) When the agent is outside the target formation and there are grid points within the range, the target grid point is selected based on the light intensity of the grid points within the agent's perception range.

[0038] Preferably, the step S5 specifically includes:

[0039] Step S5-1: the agent determines the target formation speed item according to the distance between the current position and the target grid point;

[0040] Step S5-2: the agent determines the target formation interaction speed item according to the current position, current speed, current position of the neighboring agent, and current speed of the neighboring agent;

[0041] Step S5-3: The vector sum of the target formation speed term and the target formation interaction speed term is the target speed.

[0042] Preferably, in step S2-1, the agent i Receive target formation and determine target formation parameters , and ,in, Representing an Agent i The target formation position, Representing an Agent i The target formation size, Representing an Agent i The target formation direction;

[0043] The target formation position change rate in step S2-2 The expression is as follows:

[0044] ;

[0045] Target formation size change rate The expression is as follows:

[0046] ;

[0047] Target formation direction change rate The expression is as follows:

[0048] ;

[0049] in, , , , , , are all constant coefficients greater than 0. Representing an Agent i The set of neighbor agents, Representing an Agent i Neighbor Agent j The target formation position, represents the time-varying formation position, represents the time-varying formation speed, Representing an Agent i Neighbor Agent j The target formation size, The size of the formation changes with time. is the rate of change of the time-varying formation size, Representing an Agent i Neighbor Agent j The target formation direction, represents the time-varying formation direction, Represents the rate of change of the direction of the time-varying formation.

[0050] Preferably, the step S3-1 specifically includes:

[0051] (1) Agent i By Euclidean distance Determine whether there are unoccupied grid points within the sensing range, where: Represents a grid point k location, Representing an Agent i All grid points within the sensing range, Representing an Agent i Neighbor Agent j Current location, Indicates the size radius of the agent;

[0052] (2) If there are unoccupied grid points, the agent i is the seed agent, whose hop count is 0; otherwise, ,in Representing an Agent i Neighbor Agent j The number of hops.

[0053] Preferably, in step S3-2, the agent i Grid points within sensing range k Light intensity coefficient The method of determining is:

[0054] If the agent i is the seed agent, ,otherwise,

[0055] in, Representing an Agent i Current location, Representing an Agent The gradient direction is calculated as follows:

[0056]

[0057] in, For intelligent agents i Gradient Agent location, The calculation method is as follows:

[0058] .

[0059] Preferably, in step S3-3, the agent i Grid points within sensing range k Light intensity The expression is as follows:

[0060]

[0061] in, is a constant and , Is a normal number.

[0062] Preferably, the target speed in step S5 is The expression is as follows:

[0063]

[0064] in, represents the target formation speed term, represents the target formation interaction speed term;

[0065]

[0066] in, is a constant coefficient greater than zero, is the weight coefficient,

[0067]

[0068] in, Represents the target grid point set The elements in

[0069]

[0070] in, is a constant coefficient greater than zero, is the weight coefficient, Representing an Agent j The current speed, Representing an Agent i The current speed.

[0071]

[0072] in, Represents the agent's obstacle avoidance range.

[0073] Compared with the prior art, the present invention has the following advantages:

[0074] 1. The multi-agent cluster distributed formation control method proposed in the present invention adopts a formation description method that is not based on agent labels, allowing the agents to determine formation parameters through distributed negotiation and dynamically select target points during movement, thereby greatly improving the flexibility and self-organization of the system.

[0075] 2. The multi-agent cluster distributed formation control method proposed in this invention introduces the idea of ​​exploration, that is, the agent will actively abandon the current position and move to the adjacent target position where there are relatively few agents, which can reduce the conflict of target selection between different agents due to local information. This exploration behavior can be achieved by simulating the light avoidance of organisms in nature, constructing an artificial light field to guide the movement of the agent, and effectively avoiding the conflict of target selection.

[0076] 3. The multi-agent cluster distributed formation control method proposed in the present invention has no requirements on the initial position of the agents and can control the unlabeled multi-agent cluster to form, maintain or reconstruct a certain target formation from any initial state while meeting the task requirements and environmental constraints.

[0077] 4. The multi-agent cluster distributed formation control method proposed in the present invention enables the agents to dynamically adjust their target positions according to local information during the movement process through distributed negotiation and dynamic target allocation, thereby enhancing the robustness and scalability of the system. Even if some agents fail or are lost, the system can still maintain the stability of the overall formation and the ability to execute tasks.

[0078] 5. The multi-agent cluster distributed formation control method proposed by the invention introduces exploration ideas and artificial light field construction, so that the agents can actively explore and repair defects or blank areas in the formation during movement, thus having certain self-adjustment and self-repair capabilities. This ability is of great significance for improving the adaptability and reliability of the system in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. The features and advantages of the present invention can be more clearly understood by referring to the drawings. The drawings are schematic and should not be understood as limiting the present invention in any way. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0080] Figure 1 It is a flow chart of the multi-agent cluster distributed formation control method of the present invention.

[0081] Figure 2 This is an embodiment of the present invention for generating a specified number of grid points based on a binary image. DETAILED DESCRIPTION

[0082] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0083] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0084] Example 1

[0085] The main process of the multi-agent cluster distributed formation control method of the present invention is as follows: Figure 1 As shown, in this embodiment, the distributed formation control of a multi-agent cluster in a two-dimensional plane is taken as an example to further illustrate the control method of the present invention.

[0086] Step S1, obtain the target formation and represent it as a set of grid points.

[0087] In this embodiment, the target formation image can be manually specified or automatically generated by a program according to the required formation. The process of representing the target formation as a set of grid points is as follows:

[0088] like Figure 2As shown in Figure 1, first, the target formation image is converted into a binary image, where black pixels represent the target formation. Then, the number of black pixels in the binary image is calculated. m and the number of agents n The ratio between r Finally, use n Agent coverage m black pixels, that is, each agent covers r black pixels, thus we get n The absolute coordinates of each grid point are determined by the target formation parameters, which are obtained by multi-agents through distributed consensus.

[0089] Step S2: Each intelligent agent receives the target formation, and each intelligent agent negotiates the target formation parameters with its neighboring intelligent agents to obtain unified target formation parameters.

[0090] Each intelligent agent receives the target formation and determines the target formation parameters of each intelligent agent, wherein the target formation parameters include the target formation position, the target formation size and the target formation direction.

[0091] Agent i Receive target formation and determine target formation parameters , and ,in, Representing an Agent i The target formation position, Representing an Agent i The target formation size, Representing an Agent i Target formation direction.

[0092] To improve the flexibility and autonomy of the multi-agent swarm, the agents reach consensus on the target formation parameters through local negotiation. i get , and After that, the agent in the distributed negotiation algorithm The negotiation protocol should include the target formation position change rate , Target formation size change rate and the target formation direction change rate .

[0093] ;

[0094] ;

[0095] ;

[0096] in, , , , , , are all constant coefficients greater than 0. Representing an Agent i The set of neighbor agents, Representing an Agent i Neighbor Agent j The target formation position, represents the time-varying formation position, represents the time-varying formation speed, Representing an Agent i Neighbor Agent j The target formation size, The size of the formation changes with time. is the rate of change of the time-varying formation size, Representing an Agent i Neighbor Agent j The target formation direction, represents the time-varying formation direction, Represents the rate of change of the direction of the time-varying formation.

[0097] At this point, each agent can obtain unified target formation parameters through a distributed negotiation algorithm, including target formation position, target formation size, and target formation direction.

[0098] In actual control, according to the formation parameters, each agent can obtain the absolute coordinates of the target formation grid coordinates in the real world through the method of coordinate system conversion. Step S3, based on the hop count algorithm of distributed network propagation, calculate the light intensity coefficient of the grid points within the perception range of each agent, and determine the light intensity of the grid points within the perception range of each agent.

[0099] Agent i By Euclidean distance Determine whether there are unoccupied grid points within the sensing range, where: Represents a grid point k location, Representing an Agent i All grid points within the sensing range, Representing an Agent i Neighbor Agent j Current location, Indicates the size radius of the agent;

[0100] If there are unoccupied grid points, the agent i is the seed agent, whose hop count is 0; otherwise, ,in Representing an Agent i Neighbor Agent j The number of hops;

[0101] According to the intelligent i The number of hops, calculating the agent i Grid points within sensing range k Light intensity coefficient , used to reflect the absence of the agent i The completion status of the grid points within the sensing range.

[0102] If the agent i is the seed agent, ,otherwise,

[0103] in, Representing an Agent i Current location, Representing an Agent i The gradient direction is calculated as follows:

[0104]

[0105] in, For intelligent agents i Gradient Agent The location of the agent i Gradient Agent It can be calculated based on the number of hops as follows:

[0106] .

[0107] When the intelligent i With grid points k The angle between the direction of and the gradient direction is less than hour, Greater than 1, the grid points will be increased k The light field intensity of the intelligent agent i With grid points k The angle between the direction of and the gradient direction is greater than hour, Less than 1, the grid points will be reduced k From the perspective of the construction of the entire artificial light field, if two grid points have the same illumination intensity coefficient, the denser the distribution of agents, the stronger the illumination intensity. For two grid points with the same local agent distribution, the greater the deviation between the direction of the line connecting the agent and the grid point and the gradient direction, the stronger the illumination intensity.

[0108] The artificial light field is the light intensity of each point within the agent's perception range. Each agent can calculate the light intensity of the grid points within its perception range based on the state of its neighboring agents, as follows: i All its neighboring agents are considered as point light sources, and the sum of the light intensity emitted by all point light sources at the grid point is the light intensity of the grid point within the agent's perception range. Among them, the light intensity of the point light source decreases as the propagation distance increases.

[0109] Agent i Grid points within sensing range k Light intensity The expression is as follows:

[0110]

[0111] in, is a constant and , is a positive constant, the attenuation of light intensity is proportional to the distance Inversely proportional to the power.

[0112] Step S4: each agent selects a target grid point set according to the light intensity.

[0113] Determine the target grid point set of each agent based on whether there are grid points within the perception range of each agent .

[0114] ;

[0115] when , i.e., the intelligent agent i When the agent is outside the target formation and there is no grid point within the range, it selects the nearest unoccupied grid point from the grid points as the target grid point. This strategy can drive the agent outside the target formation to move into the target formation and finally enter the formation. Q Indicates all n A collection of grid points.

[0116] when , i.e., the intelligent agent i When the agent is outside the target formation and there are grid points within the range, there may be multiple target grid points to be selected. The purpose of this multi-target selection is to eliminate the motion oscillation caused by switching back and forth between single target grid points. In addition, according to the construction of the artificial light field, the fewer task points the agent has, the smaller its light intensity is. Therefore, this strategy essentially allows the agent to actively explore those unoccupied grid points. In addition, is a constant greater than zero, used to eliminate the motion oscillation caused by switching back and forth between single target points. express The elements in .

[0117] The above target grid point set selection strategy gives the agent the ability to actively explore unoccupied grid points. This exploration ability is a necessary component for the emergence of swarm intelligence and is crucial for large-scale homogeneous swarm formation control.

[0118] Step S5, determining the target speed of each agent based on the current position, current speed of each agent, the current position of the neighboring agent, the current speed of the neighboring agent, and the target grid point.

[0119] Target speed The expression is as follows:

[0120]

[0121] in, represents the target formation speed term, Represents the target formation interaction speed term.

[0122] Agent i Determine the target formation speed item based on the distance between the current position and the target grid point .

[0123]

[0124] in, Representing an Agent i Current location, is a constant coefficient greater than zero, is the weight coefficient,

[0125]

[0126] in, Target grid point set The elements in .

[0127] If the agent i There are no unoccupied grid points within the sensing range, so it uses the nearest unoccupied grid point as its target grid point. In particular, this point has no light field intensity, so is set to 1. In addition, the weight coefficient of each grid point decreases as the light intensity increases. i When there are multiple grid points to choose from, the agent i will tend to go to the grid point with the lowest light intensity.

[0128] The agent determines the target formation interaction speed item based on the current position, current speed, the current position of the neighboring agent, and the current speed of the neighboring agent. .

[0129]

[0130] in, is a constant coefficient greater than zero, is the weight coefficient used to quantify the agent repulsion effect, Acts as a damper to ensure velocity alignment by reducing the relative error between the velocities of the agents. Representing an Agent j The current speed, Representing an Agent i The current speed.

[0131]

[0132] in, Represents the agent's obstacle avoidance range.

[0133] Target formation speed item Interaction speed item with target formation The vector sum is the target speed .

[0134] Step S6: Control the movement of each agent according to the target speed of each agent to achieve distributed formation control of the multi-agent cluster.

[0135] By repeatedly executing the above steps S2 to S6, distributed formation control of multi-agent clusters can be achieved.

[0136] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A distributed formation control method for a multi-agent cluster, characterized in that: The following steps are involved: Step S1: Obtain the target formation and represent it as a set of grid points; Step S2: Each agent receives the target formation, and each agent negotiates the target formation parameters with its neighboring agents to obtain unified target formation parameters; Step S3: Based on the hop count algorithm of distributed network propagation, the light intensity coefficient of the grid points within the perception range of each intelligent agent is calculated, and the light intensity of the grid points within the perception range of each intelligent agent is determined; Step S4: each agent selects a target grid point set according to the light intensity; Step S5: Determine the target speed of each agent according to the current position, current speed, current position and current speed of neighboring agents and the target grid point of each agent; Step S6: Control the movement of each agent according to the target speed of each agent to achieve distributed formation control of multi-agent cluster; The step S3 specifically includes: Step S3-1: Determine whether there are unoccupied grid points within the perception range of each intelligent agent, and calculate the hop count of each intelligent agent using a hop count algorithm of distributed network propagation according to the determination result; Step S3-2: determining the illumination intensity coefficient of the grid points within the sensing range of each agent based on the hop count of each agent; Step S3-3: Each agent regards all of its neighboring agents as point light sources, and the sum of the light intensities emitted by all point light sources at the grid point is the light intensity of the grid point within the agent's perception range; In step S3-2, the agent i Grid points within sensing range k Light intensity coefficient The method of determining is: If the agent i is the seed agent, ,otherwise, in, Represents a grid point k location, Representing an Agent i Current location, Representing an Agent i The gradient direction is calculated as follows: in, For intelligent agents i Gradient Agent location, The calculation method is as follows: In step S3-3, the agent i Grid points within sensing range k Light intensity The expression is as follows: in, is a constant and , is a normal number, Representing an Agent i Neighbor Agent j current location.

2. The multi-agent cluster distributed formation control method according to claim 1, characterized in that: The step S1 specifically includes: Step S1-1: Generate a target formation image manually or automatically through a program; Step S1-2: converting the target formation image into a binary image, wherein black pixels represent the target formation; Step S1-3: Calculate the number of black pixels in the binary image m and the number of agents n The ratio between r ; Step S1-3: Use n Agent coverage m black pixels, that is, each agent covers r black pixels, thus we get n Grid points.

3. The multi-agent cluster distributed formation control method according to claim 2, characterized in that: The step S2 specifically includes: Step S2-1: Each agent receives a target formation and determines the target formation parameters of each agent, wherein the target formation parameters include the target formation position, the target formation size and the target formation direction; Step S2-2: Each agent negotiates the target formation parameters with its neighboring agents, and the negotiation protocol includes the target formation position change rate, the target formation size change rate, and the target formation direction change rate; Step S2-3: Obtain unified target formation parameters.

4. The multi-agent cluster distributed formation control method according to claim 3, characterized in that: The step S4 specifically includes: Determine the target grid point set of each agent based on whether there are grid points within the perception range of each agent : (1) When the agent is outside the target formation and there is no grid point within the range, it selects the grid point closest to itself as the target grid point; (2) When the agent is within the target formation and there are grid points within the range, the target grid point is selected based on the light intensity of the grid points within the agent's perception range.

5. The multi-agent cluster distributed formation control method according to claim 4, characterized in that: The step S5 specifically includes: Step S5-1: the agent determines the target formation speed item according to the distance between the current position and the target grid point; Step S5-2: the agent determines the target formation interaction speed item according to the current position, current speed, current position of the neighboring agent, and current speed of the neighboring agent; Step S5-3: The vector sum of the target formation speed term and the target formation interaction speed term is the target speed.

6. The multi-agent cluster distributed formation control method according to claim 5, characterized in that: In step S2-1, the agent i Receive target formation and determine target formation parameters , and ,in, Representing an Agent i The target formation position, Representing an Agent i The target formation size, Representing an Agent i The direction of the target formation; The target formation position change rate in step S2-2 The expression is as follows: ； Target formation size change rate The expression is as follows: ； Target formation direction change rate The expression is as follows: ； in, , , , , , are all constant coefficients greater than 0. Representing an Agent i The set of neighbor agents, Representing an Agent i Neighbor Agent j The target formation position, represents the time-varying formation position, represents the time-varying formation speed, Representing an Agent i Neighbor Agent j The target formation size, The size of the formation changes with time. is the rate of change of the time-varying formation size, Representing an Agent i Neighbor Agent j The target formation direction, represents the time-varying formation direction, Represents the rate of change of the direction of the time-varying formation.

7. The multi-agent cluster distributed formation control method according to claim 6, characterized in that: The step S3-1 specifically includes: (1) Agent i By Euclidean distance Determine whether there are unoccupied grid points within the sensing range, where: Representing an Agent i All grid points within the sensing range, Indicates the size radius of the agent; (2) If there are unoccupied grid points, the agent i is the seed agent, whose hop count is 0; otherwise, ,in Representing an Agent i Neighbor Agent j The number of hops.

8. The multi-agent cluster distributed formation control method according to claim 6, characterized in that: The target speed in step S5 The expression is as follows: in, represents the target formation speed term, represents the target formation interaction speed term; in, is a constant coefficient greater than zero, is the weight coefficient, in, Represents the target grid point set The elements in in, is a constant coefficient greater than zero, is the weight coefficient, Representing an Agent j The current speed, Representing an Agent i The current speed; in, Represents the agent's obstacle avoidance range.

Images