Multi-agent cooperative unmanned planning method
By constructing a normalized target gravitational field and obstacle repulsive field and designing a smooth decision-making module, the safety avoidance problem in multi-agent cooperative planning was solved, efficient cooperation among multiple agents was achieved, and the safety and reliability of the autonomous driving system were improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- EAST CHINA JIAOTONG UNIVERSITY
- Filing Date
- 2026-04-24
- Publication Date
- 2026-05-29
AI Technical Summary
Existing autonomous driving planning methods are mainly geared towards single-agent scenarios and lack collaborative decision-making mechanisms among multiple agents. This makes it difficult to achieve safe avoidance and non-interference in highly dynamic scenarios such as multiple agents traveling together or crossing intersections, which can easily lead to collision accidents.
A normalized target gravitational field and a normalized obstacle repulsive field are constructed. A smooth decision-making module is designed. Through kinematic modeling and the triangle interior point method, the expected forward linear velocity and yaw angular velocity of the agents are planned for static and dynamic obstacle scenarios, respectively, to achieve collaborative planning of multiple agents.
It significantly improves the dynamic adaptability and collaborative coherence of multi-agent collaborative work, enhances operational safety and reliability, improves the adaptability and stability of path planning, and meets the safety and reliability requirements of multi-machine collaborative operations.
Smart Images
Figure CN122108159A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of autonomous driving technology, and more specifically to an autonomous driving planning method based on multi-agent cooperation. Background Technology
[0002] With the deepening application of unmanned technology, the application scenarios of driverless vehicles have expanded from open roads to diverse scenarios such as factory production, agricultural management, and biomedicine. In these scenarios, driverless vehicles no longer operate in isolation but exist in a collaborative mode of multi-agent clusters. When a large number of agents (such as drones, autonomous vehicles, and unmanned boats) operate simultaneously in the same space, their trajectories intertwine. How to ensure safe avoidance and non-interference between them through effective collaborative planning, and avoid chain accidents such as collisions and rollovers, has become a core challenge that urgently needs to be overcome.
[0003] However, existing autonomous driving planning methods are mainly geared towards single-agent scenarios, focusing on individual path planning and local obstacle avoidance, and lacking collaborative decision-making mechanisms among multiple agents. In highly dynamic scenarios such as multi-agent traffic and intersection crossings, existing methods struggle to achieve intent sharing and real-time conflict resolution among multiple agents, easily leading to mutual interference or even collisions, and failing to meet the safety and reliability requirements of multi-machine collaborative operations.
[0004] Therefore, providing a multi-agent collaborative autonomous driving planning method is an important problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0005] The purpose of this invention is to provide a multi-agent collaborative autonomous driving planning method to better meet the safety and reliability requirements of multi-machine collaborative operations.
[0006] A multi-agent cooperative autonomous driving planning method includes: Step S1: Define the obstacle coordinate system, the agent coordinate system for avoiding static obstacles, the agent coordinate system for avoiding dynamic obstacles, and key parameters. The key parameters include the type of agent, which is divided into... THAT Intelligent agents and ? B Intelligent agents, among which, THAT A class of intelligent agents is an intelligent agent controlled by this planning method. ? B A class of intelligent agents is an intelligent agent that is not controlled by this planning method. THAT Intelligent agents can perceive through their own installed sensors. ? B Upper and lower limits of the speed of intelligent agents; Step S2: Based on the defined coordinate system and key parameters, design a multi-agent cooperative planning strategy to complete the kinematic modeling of the forward linear velocity and yaw rate of each agent. Step S3: When in a smart body coordinate system scenario that avoids static obstacles, based on kinematic modeling, a normalized target gravitational field and a normalized obstacle repulsive field based on the triangle interior point method are constructed for the target point and the static obstacle, respectively. Then, the normalized target gravitational field and the normalized obstacle repulsive field are fused, a smooth decision module is designed, and finally, the output of the smart body coordinate system scenario for avoiding static obstacles is provided. THAT The expected forward linear velocity and yaw rate of the intelligent agent; Step S4: When in the intelligent body coordinate system scene of avoiding dynamic obstacles, based on kinematic modeling, in THAT The intelligent agent is surrounded by four regions. Based on these four regions, obstacle avoidance trajectory planning is performed, and the intelligent agent coordinate system for avoiding dynamic obstacles is output. THAT The expected forward linear velocity and yaw rate of the intelligent agent.
[0007] The multi-agent cooperative autonomous driving planning method provided by the present invention has the following beneficial effects: (1) In the collaborative work of multiple agents, the present invention constructs a normalized target gravitational field and a normalized obstacle repulsive field, and then integrates the normalized target gravitational field and the normalized obstacle repulsive field to design a smooth decision module, which significantly improves the dynamic adaptability and collaborative coherence among agents, and effectively enhances the operational safety and overall reliability of collaborative work.
[0008] (2) The present invention constructs a normalized obstacle repulsion field based on the triangle interior point method, which effectively improves the adaptability and stability of path planning in complex environments.
[0009] (3) To address the differentiated needs of different collaborative work environments, this invention constructs intelligent body coordinate systems for avoiding static obstacles and intelligent body coordinate systems for avoiding dynamic obstacles, and inputs corresponding coordinate system scenarios according to different coordinate system scenarios. THAT The expected forward velocity and yaw rate of the intelligent agent can achieve adaptive coverage and precise optimization for different scenarios, thereby better meeting the safety and reliability requirements of multi-machine collaborative operations. Attached Figure Description
[0010] Figure 1 This is a flowchart illustrating the multi-agent collaborative autonomous driving planning method provided by the present invention. Figure 2 This is a schematic diagram of the obstacle coordinate system; Figure 3 A schematic diagram of the intelligent agent's coordinate system for avoiding static obstacles; Figure 4 A schematic diagram of the intelligent agent's coordinate system for avoiding dynamic obstacles; Figure 5A schematic diagram illustrating the definition of a normalized obstacle repulsive field using the triangle interior point method; Figure 6 Simulation diagram of existing technologies for avoiding static obstacles; Figure 7 This is a simulation diagram illustrating how the present invention avoids static obstacles. Figure 8 Simulation diagrams illustrating how existing technologies can avoid dynamic obstacles; Figure 9 This is a simulation diagram illustrating how the present invention avoids dynamic obstacles. Detailed Implementation
[0011] To facilitate understanding of the present invention, a more complete description will be given below with reference to various embodiments. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
[0012] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.
[0013] Please see Figure 1 The embodiments of the present invention provide a multi-agent cooperative autonomous driving planning method, including steps S1-S4: Step S1: Define the obstacle coordinate system, the intelligent agent coordinate system for avoiding static obstacles, the intelligent agent coordinate system for avoiding dynamic obstacles, and key parameters.
[0014] Among them, the key parameters include the type of agent, which is divided into: THAT Intelligent agents and ? B Intelligent agents.
[0015] in, THAT Intelligent agents are those controlled by this planning method, such as self-driving cars.
[0016] ? B Intelligent agents are those that are not controlled by this planning method, such as pedestrians and animals.
[0017] THAT Intelligent agents can perceive through their own installed sensors. ? B Upper and lower limits of the speed of the intelligent agent.
[0018] In this embodiment, for the first indivual THAT intelligent agents It has a position vector State vector intelligent agent It is constructed as a circular rigid body model, with a set positive constant R as the radius, and each... THAT All intelligent agents have a target point they need to reach, the first indivual THAT The target point of the intelligent agent position vector .
[0019] For the indivual ? B intelligent agents It has a position vector State vector intelligent agent It is constructed as a circular rigid body model, with a radius set as a positive constant R.
[0020] Key parameters also include: in the first indivual THAT Neighborhood of intelligent agents Inner adjacent THAT A collection of intelligent agents In the indivual THAT Neighborhood of intelligent agents Inner adjacent ? B A collection of intelligent agents Neighborhood Indicates the first indivual THAT A circular sensory region resembling an intelligent agent, this region being centered around the intelligent agent. position vector Centered on the positive constant Rc, with the radius set as the radius.
[0021] Define the obstacle coordinate system, such as Figure 2 As shown, intelligent agent The forward linear velocity is The yaw rate is intelligent agent The angle between the direction of motion and the positive x-axis is .
[0022] Define the agent coordinate system for avoiding static obstacles, such as Figure 3 As shown, the agent coordinate system for avoiding static obstacles has an upper boundary in its neighborhood. Neighborhood lower boundary On the boundary of the neighborhood Within this area, each agent exerts a certain influence. At the lower boundary of the neighborhood... Within this boundary, the limit to which intelligent agents can be mutually safe is that if the distance is less than this boundary, they are considered unsafe. Furthermore... , For the first indivual THAT The radius of the intelligent agent, This is a predefined positive integer. (The agent...) The expected forward velocity is and yaw rate , No. indivual THAT The angle between the direction of motion of the intelligent agent and the positive x-axis is... .
[0023] Define the agent coordinate system for avoiding dynamic obstacles, such as Figure 4 As shown, the intelligent agent coordinate system for avoiding dynamic obstacles has an upper boundary in the neighborhood along the positive direction of motion. Upper boundary of the neighborhood in the opposite direction of motion The lower boundary of the neighborhood in the positive direction of motion The lower boundary of the neighborhood in the opposite direction of motion ,in, , , , , For the set coefficient, For the first indivual THAT The expected forward velocity of the intelligent agent is the discrete time step.
[0024] Step S2: Based on the defined coordinate system and key parameters, design a multi-agent collaborative planning strategy to complete the kinematic modeling of the forward linear velocity and yaw rate of each agent.
[0025] Specifically, step S2 includes: When the indivual ? B When the intelligent agent is a static obstacle, the kinematic modeling expression is:
[0026]
[0027] in, For the first indivual ? B The forward linear velocity of an intelligent agent, For the first indivual ? B The yaw rate of the intelligent agent; When the indivual ? BWhen the intelligent agent is a dynamic obstacle, the kinematic modeling expression is:
[0028]
[0029] in, For the first indivual ? B The maximum forward linear velocity of the intelligent agent. For the first indivual ? B The maximum value of the yaw rate of the intelligent agent. and A random number within a specified range.
[0030] Step S3: When in a smart body coordinate system scenario that avoids static obstacles, based on kinematic modeling, a normalized target gravitational field and a normalized obstacle repulsive field based on the triangle interior point method are constructed for the target point and the static obstacle, respectively. Then, the normalized target gravitational field and the normalized obstacle repulsive field are fused, a smooth decision module is designed, and finally, the output of the smart body coordinate system scenario for avoiding static obstacles is provided. THAT The expected forward linear velocity and yaw rate of the intelligent agent.
[0031] Wherein, let the target point state vector =0, For the first indivual THAT The polar angle of the line connecting the target point of the intelligent agent and the origin gives the normalized target gravitational field. The expression is:
[0032]
[0033]
[0034] in, and They are respectively Vector components in the x-axis and y-axis directions, For the first indivual THAT The distance between the agent-like entity and the target point. , for Vector components in the x-axis direction, for Vector components in the y-axis direction; The normalized obstacle repulsive field is defined using the triangle interior point method. .
[0035] First, please combine Figure 5 Define some key points, namely the first... indivual THAT intelligent agents , No. indivual THAT The target point of the intelligent agent , No. indivual ? B intelligent agents The first central point Second central point , No. indivual THAT Temporary target point of intelligent agents When intelligent agents Entering the intelligent agent When the detection range is limited, in order to avoid intelligent agents intelligent agent The target point is Switch to When intelligent agents Leaving the intelligent agent After the detection range is reached, the point coordinates of the target point are restored to... .
[0036] Setting points and The line in question is a straight line. ,by With the center of the circle, A circle with radius is a circle At this point, the straight line With circle If two points intersect, choose the one closest to the other. The point is the first central point. .
[0037] Take point , , Find the coordinates of the incenter of the triangle formed by the triangle; this point is the second pivot point. .
[0038] Setting points and The line in question is a straight line. Regarding the straight line Do something The symmetrical point, which is the temporary target point. .
[0039] Ultimately, the normalized obstacle repulsive field The expression is:
[0040]
[0041]
[0042]
[0043]
[0044]
[0045]
[0046] in, and They are respectively Vector components in the x-axis and y-axis directions, For the first indivual THAT intelligent agents Location coordinates, For the first indivual THAT Temporary target point of intelligent agents Location coordinates, For the first indivual THAT The target point of the intelligent agent Location coordinates, The first central point Location coordinates, The second central point Location coordinates, , , , , , These are process parameters.
[0047] In step S3, the smooth decision module satisfies the following equation:
[0048] in, For the comprehensive potential field, For the first indivual THAT Intelligent agents and the first indivual ? B The distance between intelligent agents . This indicates the case where there are no adjacent static obstacles. This indicates the presence of adjacent static obstacles.
[0049] In a scenario where an intelligent agent coordinate system avoids static obstacles, the first... indivual THAT Expected forward velocity of the intelligent agent and yaw rate They are respectively:
[0050]
[0051]
[0052] in, For the first indivual THAT The maximum forward linear velocity of the intelligent agent. For the first indivual THAT The maximum yaw rate of the intelligent agent. For the first indivual THAT The angle between the direction of motion of the intelligent agent and the positive x-axis. The Gaussian error function is a function whose values range from ( A smooth saturated nonlinear function between 1 and 1) can make the control input or weighting coefficients transition smoothly and avoid abrupt changes; Indicates the orientation of the overall potential field. It is the arctangent function. and They are respectively Vector components in the x-axis and y-axis directions.
[0053] Step S4: When in the intelligent body coordinate system scene of avoiding dynamic obstacles, based on kinematic modeling, in THAT The intelligent agent is surrounded by four regions. Based on these four regions, obstacle avoidance trajectory planning is performed, and the intelligent agent coordinate system for avoiding dynamic obstacles is output. THAT The expected forward linear velocity and yaw rate of the intelligent agent.
[0054] Specifically, step S4 includes: When in an agent coordinate system scenario that avoids dynamic obstacles, first define binary values. ,when At that time, the first indivual THAT intelligent agents in Adjacent THAT No intelligent agent has been found to be related to ? B If the intelligent agents are adjacent, then ;when At that time, the first indivual THAT intelligent agents in Adjacent THAT Intelligent agents, at least one of which is similar to ? B If the intelligent agents are adjacent, then ; Then based on the first indivual THAT The forward linear velocity direction and yaw angular velocity direction of the intelligent agent are in the first... indivual THAT The intelligent agent is surrounded by four regions, namely , , , ,like Figure 4 As shown.
[0055] when ,and At that time, the first indivual THAT Expected forward velocity of the intelligent agent and yaw rate They are respectively:
[0056]
[0057] in, For the first indivual THAT The position vector of the agent-like entity. This indicates taking the minimum value. Indicates the first indivual THAT Intelligent agents and the first indivual THAT The distance between intelligent agents For the first indivual THAT Intelligent agents relative to the first indivual THAT The critical forward velocity of an intelligent agent. For the first indivual THAT The agent-like entity relative to the preceding neighbor THAT The forward linear velocity of an intelligent agent, For the first indivual THAT intelligent agents relative to their rear neighbors THAT The forward linear velocity of an intelligent agent, , , , For coefficients, , , For the first indivual THAT The polar angle of the line connecting the target point of the intelligent agent and the origin; when ,and At that time, the first indivual THAT Expected forward velocity of the intelligent agent and yaw rate They are respectively:
[0058]
[0059] in, Indicates the first indivual THAT Intelligent agents and the first indivual ? B The distance between intelligent agents For the first indivual THAT Intelligent agents relative to the first indivual ? B The critical forward velocity of an intelligent agent. For the first indivual THAT The agent-like entity relative to the preceding neighbor ? B The forward linear velocity of an intelligent agent, For the first indivual THAT intelligent agents relative to their rear neighbors ? B The forward linear velocity of an intelligent agent, , , , For coefficients, , , Indicates time, express The differential, These are preset positive integers; when ,and At that time, the first indivual THAT Expected forward velocity of the intelligent agent and yaw rate They are respectively:
[0060]
[0061] when ,and At that time, the first indivual THAT Expected forward velocity of the intelligent agent and yaw rate They are respectively:
[0062]
[0063] in, Indicates the first indivual THAT Neighborhood of intelligent agents Inner adjacent THAT Intelligent agents and neighbors ? B A collection of intelligent agents, for The elements in Indicates the first indivual THAT The agent-like entity relative to the preceding neighbor THAT intelligent agents or those in front ? B The forward linear velocity of an intelligent agent, , Indicates the first indivual THAT intelligent agents relative to their rear neighbors THAT intelligent agents or their rear neighbors ? B The forward linear velocity of an intelligent agent, .
[0064] To verify the obstacle avoidance and navigation performance of the autonomous driving planning method proposed in this invention, a comparative simulation was conducted using existing technologies, with the Rapid Random Search Tree (RRT) path planning method being used as the comparison.
[0065] (1) Simulation comparison of avoiding static obstacles The simulation scene is set to contain eight ? B Intelligent agent, single THAT A planar environment for intelligent agents.
[0066] This invention shows significant differences compared to existing technologies for simulating static obstacle avoidance. Figure 6 It is known that, using existing technology, THAT Intelligent agents in evasion ? B When the intelligent agent is not smooth enough, the turning action is abrupt, the path planning efficiency is insufficient, and there is a significant detour in the process of reaching the target point. Figure 7 It can be seen that by using this invention, THAT Intelligent agents can sense in advance ? B The intelligent agent plans a smooth, continuous optimal obstacle avoidance path, and works with the system throughout the process. ? BThe intelligent agent maintains a safe distance and ultimately reaches the target point efficiently and smoothly.
[0067] The comparative results show that the method proposed in this invention provides a smoother path, higher safety, and better navigation efficiency in scenarios where static obstacles are avoided, effectively improving the motion performance of intelligent agents in complex static environments.
[0068] (2) Simulation comparison of avoiding dynamic obstacles The simulation scenario is set to include four ? B Three intelligent agents THAT A planar environment for intelligent agents.
[0069] Depend on Figure 8 It can be seen that the present invention differs significantly from existing technologies in simulating dynamic obstacle avoidance. Using existing technologies, each... THAT Intelligent agents in evasion ? B The path continuity of the intelligent agent is poor, the turning action is abrupt and stiff, and the movement trajectory lacks rationality and realism. Figure 9 It can be seen that, by employing this invention, each THAT Intelligent agents can predict in advance ? B Based on the movement trends of the intelligent agent, a smooth and continuous obstacle avoidance path is planned, and THAT Intelligent agents can avoid each other and prevent collisions.
[0070] The comparative results show that the method proposed in this invention is superior to existing solutions in terms of real-time path performance, reasonable prediction, and multi-agent collaborative obstacle avoidance in dynamic obstacle avoidance scenarios, and the overall navigation is more stable and efficient.
[0071] In summary, the multi-agent cooperative autonomous driving planning method described above has the following beneficial effects: (1) In the collaborative work of multiple agents, the present invention constructs a normalized target gravitational field and a normalized obstacle repulsive field, and then integrates the normalized target gravitational field and the normalized obstacle repulsive field to design a smooth decision module, which significantly improves the dynamic adaptability and collaborative coherence among agents, and effectively enhances the operational safety and overall reliability of collaborative work.
[0072] (2) The present invention constructs a normalized obstacle repulsion field based on the triangle interior point method, which effectively improves the adaptability and stability of path planning in complex environments.
[0073] (3) To address the differentiated needs of different collaborative work environments, this invention constructs intelligent body coordinate systems for avoiding static obstacles and intelligent body coordinate systems for avoiding dynamic obstacles, and inputs corresponding coordinate system scenarios according to different coordinate system scenarios. THATThe expected forward velocity and yaw rate of the intelligent agent can achieve adaptive coverage and precise optimization for different scenarios, thereby better meeting the safety and reliability requirements of multi-machine collaborative operations.
Claims
1. A multi-agent cooperative autonomous driving planning method, characterized in that, include: Step S1: Define the obstacle coordinate system, the agent coordinate system for avoiding static obstacles, the agent coordinate system for avoiding dynamic obstacles, and key parameters. The key parameters include the type of agent, which is divided into... CA Intelligent agents and CB Intelligent agents, among which, CA A class of intelligent agents is an intelligent agent controlled by this planning method. CB A class of intelligent agents is an intelligent agent that is not controlled by this planning method. CA Intelligent agents can perceive through their own installed sensors. CB Upper and lower limits of the speed of intelligent agents; Step S2: Based on the defined coordinate system and key parameters, design a multi-agent cooperative planning strategy to complete the kinematic modeling of the forward linear velocity and yaw rate of each agent. Step S3: When in a smart body coordinate system scenario that avoids static obstacles, based on kinematic modeling, a normalized target gravitational field and a normalized obstacle repulsive field based on the triangle interior point method are constructed for the target point and the static obstacle, respectively. Then, the normalized target gravitational field and the normalized obstacle repulsive field are fused, a smooth decision module is designed, and finally, the output of the smart body coordinate system scenario for avoiding static obstacles is provided. CA The expected forward linear velocity and yaw rate of the intelligent agent; Step S4: When in the intelligent body coordinate system scene of avoiding dynamic obstacles, based on kinematic modeling, in CA The intelligent agent is surrounded by four regions. Based on these four regions, obstacle avoidance trajectory planning is performed, and the intelligent agent coordinate system for avoiding dynamic obstacles is output. CA The expected forward linear velocity and yaw rate of the intelligent agent.
2. The multi-agent cooperative autonomous driving planning method according to claim 1, characterized in that, In step S1, the key parameters also include the first... indivual CA intelligent agents position vector , No. indivual CA Target point of intelligent agent position vector , No. indivual CB Position vector of an intelligent agent In the indivual CA Neighborhood of intelligent agents Inner adjacent CA A collection of intelligent agents In the indivual CA Neighborhood of intelligent agents Inner adjacent CB A collection of intelligent agents .
3. The multi-agent cooperative autonomous driving planning method according to claim 2, characterized in that, In step S1, the intelligent agent coordinate system for avoiding static obstacles has an upper boundary in its neighborhood. Neighborhood lower boundary ,and , For the first indivual CA The radius of the intelligent agent, For set positive numbers; The agent coordinate system for avoiding dynamic obstacles has an upper boundary in the neighborhood along the positive direction of motion. Upper boundary of the neighborhood in the opposite direction of motion The lower boundary of the neighborhood in the positive direction of motion The lower boundary of the neighborhood in the opposite direction of motion ,in, , , , , For the set coefficient, For the first indivual CA The expected forward velocity of the intelligent agent is the discrete time step.
4. The multi-agent cooperative autonomous driving planning method according to claim 3, characterized in that, Step S2 specifically includes: When the indivual CB When the intelligent agent is a static obstacle, the kinematic modeling expression is: in, For the first indivual CB The forward linear velocity of an intelligent agent, For the first indivual CB The yaw rate of the intelligent agent; When the indivual CB When the intelligent agent is a dynamic obstacle, the kinematic modeling expression is: in, For the first indivual CB The maximum forward linear velocity of the intelligent agent. For the first indivual CB The maximum value of the yaw rate of the intelligent agent. and A random number within a specified range.
5. The multi-agent cooperative autonomous driving planning method according to claim 4, characterized in that, In step S3, the target gravitational field is normalized. The expression is: in, and They are respectively Vector components in the x-axis and y-axis directions, For the first indivual CA The distance between the agent and the target point , for Vector components in the x-axis direction, for Vector components in the y-axis direction; Normalized obstacle repulsive field The expression is: in, and They are respectively Vector components in the x-axis and y-axis directions, For the first indivual CA intelligent agents Location coordinates, For the first indivual CA Temporary target point of intelligent agents Location coordinates, For the first indivual CA Target point of intelligent agent Location coordinates, The first central point Location coordinates, The second central point Location coordinates, , , , , , These are process parameters.
6. The multi-agent cooperative autonomous driving planning method according to claim 5, characterized in that, In step S3, the smooth decision module satisfies the following equation: in, For the comprehensive potential field, For the first indivual CA Intelligent agents and the first indivual CB The distance between intelligent agents .
7. The multi-agent cooperative autonomous driving planning method according to claim 6, characterized in that, In step S3, in the intelligent agent coordinate system scenario for avoiding static obstacles, the first... indivual CA Expected forward velocity of the intelligent agent and yaw rate They are respectively: in, For the first indivual CA The maximum forward linear velocity of the intelligent agent. For the first indivual CA The maximum yaw rate of the intelligent agent. For the first indivual CA The angle between the direction of motion of the intelligent agent and the positive x-axis. The Gaussian error function is... Indicates the orientation of the overall potential field. It is the arctangent function. and They are respectively Vector components in the x-axis and y-axis directions.
8. The multi-agent cooperative autonomous driving planning method according to claim 7, characterized in that, Step S4 specifically includes: When in an agent coordinate system scenario that avoids dynamic obstacles, first define binary values. ,when At that time, the first indivual CA intelligent agents in Adjacent CA No intelligent agent has been found to be related to CB If the intelligent agents are adjacent, then ;when At that time, the first indivual CA intelligent agents in Adjacent CA Intelligent agents, at least one of which is similar to CB If the intelligent agents are adjacent, then ; Then based on the first indivual CA The forward linear velocity direction and yaw angular velocity direction of the intelligent agent are in the first... indivual CA The intelligent agent is surrounded by four regions, namely , , , ; when ,and At that time, the first indivual CA Expected forward velocity of the intelligent agent and yaw rate They are respectively: in, For the first indivual CA The position vector of the agent-like entity. This indicates taking the minimum value. Indicates the first indivual CA Intelligent agents and the first indivual CA The distance between intelligent agents For the first indivual CA Intelligent agents relative to the first indivual CA The critical forward velocity of an intelligent agent. For the first indivual CA The agent-like entity relative to the preceding neighbor CA The forward linear velocity of an intelligent agent, For the first indivual CA intelligent agents relative to their rear neighbors CA The forward linear velocity of an intelligent agent, , , , For coefficients, , , For the first indivual CA The polar angle of the line connecting the target point of the intelligent agent and the origin; when ,and At that time, the first indivual CA Expected forward velocity of the intelligent agent and yaw rate They are respectively: in, Indicates the first indivual CA Intelligent agents and the first indivual CB The distance between intelligent agents For the first indivual CA Intelligent agents relative to the first indivual CB The critical forward velocity of an intelligent agent. For the first indivual CA The agent-like entity relative to the preceding neighbor CB The forward linear velocity of an intelligent agent, For the first indivual CA intelligent agents relative to their rear neighbors CB The forward linear velocity of an intelligent agent, , , , For coefficients, , , Indicates time, express The differential, These are preset positive integers; when ,and At that time, the first indivual CA Expected forward velocity of the intelligent agent and yaw rate They are respectively: when ,and At that time, the first indivual CA Expected forward velocity of the intelligent agent and yaw rate They are respectively: in, Indicates the first indivual CA Neighborhood of intelligent agents Inner adjacent CA Intelligent agents and neighbors CB A collection of intelligent agents, for The elements in Indicates the first indivual CA The agent-like entity relative to the preceding neighbor CA intelligent agents or those in front CB The forward linear velocity of an intelligent agent, , Indicates the first indivual CA intelligent agents relative to their rear neighbors CA intelligent agents or their rear neighbors CB The forward linear velocity of an intelligent agent, .