Unity3D-based multi-agent collaborative integrated navigation virtual simulation system and method

Abstract

The application provides a multi-agent collaborative combined navigation virtual simulation system and method based on Unity3D, belongs to the technical field of combined navigation, and solves the problems that a traditional combined navigation experiment method is high in cost, high in risk, and difficult to simulate various complex actual working conditions, wherein the simulation system comprises a display layer, a business layer and an underlying service; the display layer provides an operation and monitoring platform for users; the business layer is used for the control logic, queue management, path planning and artificial intelligence training function of the agent in the simulation environment; and the underlying service is used for agent machine learning model training and project version management; the application provides a low-cost, high-safety virtual simulation scheme for agent motion control and collaborative operation efficiency optimization, integrates an improved A* search algorithm and an artificial intelligence training interface, and provides an economical and efficient virtual test platform for research and development of an agent control algorithm.

Description

Technical Field

[0001] This invention relates to a multi-agent collaborative integrated navigation virtual simulation system and method based on Unity3D, belonging to the field of integrated navigation technology. Background Technology

[0002] With the continuous advancement of technology, multi-agent systems are playing an increasingly important role in the fields of automation and intelligence. These systems typically consist of multiple interacting agents capable of cooperating to accomplish complex tasks. To improve the collaborative efficiency of these agents and ensure operational safety, it is necessary to conduct in-depth research and testing on their motion control and collaborative strategies. Traditional experimental methods are often costly, risky, and difficult to simulate various complex real-world working conditions. Therefore, developing a testing and training platform based on virtual simulation technology is particularly important.

[0003] In the field of virtual simulation, Unity3D, as a powerful game development engine, is widely used in simulation and training systems due to its high interactivity and realistic visual effects. Unity3D allows for the construction of complex virtual environments to simulate the behavior and responses of intelligent agents in various conditions. Furthermore, Unity3D supports multiple programming languages, particularly C#, enabling developers to flexibly implement the control logic and collaborative strategies of intelligent agents.

[0004] To enable intelligent agents to effectively avoid obstacles and work collaboratively in virtual environments, it is necessary to integrate advanced path planning and obstacle avoidance algorithms. The A* algorithm, as a classic path search algorithm, is widely used in various navigation and path planning problems due to its efficiency and accuracy. Optimizing the A* algorithm can further improve the obstacle avoidance capabilities of intelligent agents in complex environments.

[0005] Meanwhile, with the development of artificial intelligence technology, especially the application of reinforcement learning in multi-agent systems, agents are able to autonomously learn optimal control strategies through interaction with their environment. MLAgent, a machine learning framework based on Unity3D, provides an interface integrated with Tensorflow, allowing researchers to train and deploy AI models in virtual environments to achieve autonomous control of agents. Summary of the Invention

[0006] To address the problems of high cost, high risk, and difficulty in simulating various complex real-world conditions in traditional integrated navigation experimental methods, this invention proposes a multi-agent collaborative integrated navigation virtual simulation system and method based on Unity3D.

[0007] The technical solution adopted by this invention to solve the above problems is as follows: This invention proposes a multi-agent cooperative navigation virtual simulation system based on Unity3D, comprising:

[0008] Presentation layer, business layer, and underlying services;

[0009] The presentation layer provides users with an operation and monitoring platform;

[0010] The business layer is used for the control logic, queue management, path planning, and artificial intelligence training functions of intelligent agents in the simulation environment;

[0011] The underlying services are used for training machine learning models for intelligent agents and for version management of projects.

[0012] Optionally, the presentation layer includes an environment simulation module, a UnityUI interface, and a Unity Editor window;

[0013] The environment simulation module is used to construct the virtual environment required for intelligent agent interaction;

[0014] UnityUI provides users with an interface for operation.

[0015] The Unity Editor window provides users with a monitoring platform.

[0016] Optionally, the business layer includes an Agent control module, an Agent queue control module, a path planning module, and an AI environment module;

[0017] The Agent control module is responsible for the control of a single agent, which includes the management of power data, torque data, path data, and follow data, as well as mode control and algorithm selection.

[0018] The Agent queue control module is used to manage agent queue data, target data, and queue control.

[0019] The path planning module is used to calculate the optimal path for the agent to reach the target location from its current location;

[0020] The AI ​​environment module is used to train intelligent agents with artificial intelligence and predict the best course of action based on environmental feedback.

[0021] Optionally, the Agent control module includes a single Agent control module and a multi-Agent control module;

[0022] The single-agent control module is used for agent dynamic control, target path planning, and action state control;

[0023] The multi-agent control module is used for queue mode switching, agent queue management, and queue movement simulation.

[0024] Optionally, the underlying services include the Tensorflow training interface service module and the Plastic SCM project version control module;

[0025] Tensorflow training interface service is used to convert data collected by the agent into training signals to guide the agent's behavior adjustment;

[0026] Plastic SCM Project Version Control is used for project version management.

[0027] A virtual simulation method for multi-agent cooperative navigation based on Unity3D includes:

[0028] Step 1: Construct the Unity3D virtual environment required for intelligent agent interaction based on the environment simulation module;

[0029] Step 2: Combining the single-agent control module, path planning module, and multi-agent control module, set the control strategy for the intelligent agent in the Unity3D virtual environment using C# scripts;

[0030] Step 3: Using the improved A* search algorithm and the RVO2 simulator concept, combined with the agent's control strategy, control multiple agents to avoid obstacles;

[0031] Step 4: Train and deploy the AI ​​model in the Unity3D virtual environment through the AI ​​environment module to control the agent. By connecting to the Tensorflow training interface service, use the MLAgent plugin to train the agent's control strategy so that the agent can successfully avoid obstacles and complete the multi-agent collaborative navigation task.

[0032] Optionally, step 1 specifically includes:

[0033] Step 1.1: Extract information about all objects in the navigation environment from multi-angle images, photos and video data of objects using computer vision technology, and calculate and construct 3D models of all objects using one of the following methods: structured light scanning, stereo vision and structure of motion reconstruction.

[0034] Step 1.2: Import the 3D models of all objects into the Unity3D environment, and optimize the 3D models of all objects during the import process. The optimization of the 3D models of all objects includes simplifying polygons, optimizing textures and materials, and adjusting the scale and proportions of the models.

[0035] Step 1.3: Place the optimized 3D models of all objects in their corresponding positions in the Unity3D environment, and set the lighting, weather effects, and physical properties in the Unity3D environment to complete the construction of the Unity3D virtual environment.

[0036] Optionally, step 2 specifically includes:

[0037] Step 2.1: Utilize Unity3D's sensors and physics engine to enable the agent to perceive changes in the surrounding environment, and collect data on the agent's position, velocity, and distance to surrounding objects through the integrated API;

[0038] Step 2.2: The path planning module uses the improved A* search algorithm to analyze the Unity3D virtual environment data, and combines the collected agent position, speed and distance to surrounding objects to calculate the optimal path for the agent to reach the target position from the current position.

[0039] Step 2.3: Through the single-agent control module and the multi-agent control module, multiple agents coordinate their actions to maintain the formation according to the preset formation and dynamic adjustment strategy. By monitoring the surrounding environment of the agents in real time, when a potential collision risk is detected, the optimal obstacle avoidance path is calculated and the speed and direction of the agents are adjusted to complete the setting of the agent control strategy.

[0040] Optionally, step 3 specifically includes:

[0041] Step 3.1: After receiving the path planning request, the agent activates the integrated control script of the A* algorithm module and the RVO2 simulator concept to calculate the optimal path from the current position to the target position;

[0042] Step 3.2: Calculate the repulsive force relationship between agent i and agent j in the x-axis direction;

[0043] Step 3.3: Combining with Step 3.2, calculate the repulsive force relationship between agent i and agent j in the y-axis and z-axis directions;

[0044] Step 3.4: Superimpose the repulsive force relationship between agent i and agent j in the xyz axis direction to obtain the desired velocity component of the i-th agent in the xyz axis direction in the inertial coordinate system at time t;

[0045] Step 3.5: Combine the data collected by Unity3D's sensors to dynamically update the search tree of the improved A* search algorithm, and adjust the maximum turning angular velocity and acceleration of the agent to ensure that the agent moves along the optimal obstacle avoidance path and avoids obstacles.

[0046] The expression for the repulsive force relationship between agents i and j along the x-axis is:

[0047]

[0048] In formula (1), η is the repulsive force scale factor, l ij Rdetect is the distance between agents i and j, and Rrecept is the radius of influence of the repulsive field.

[0049] The formula for calculating the desired velocity components of the i-th agent at time t in the xyz axis direction in the inertial coordinate system is:

[0050]

[0051] In formulas (3)-(5), Let be the desired velocity component of the i-th agent in the x-axis direction at time t. Let be the desired velocity component of the i-th agent in the y-axis direction at time t. Let Δt be the desired velocity component of the i-th agent in the z-axis direction at time t, and let Δt be the interval for system state updates.

[0052] Optionally, step 4 specifically includes:

[0053] Step 4.1: In the Unity3D virtual environment, receive real-time data from the sensors through the MLAgent plugin. The real-time data includes the agent's position, velocity, and distance to surrounding obstacles. Convert the real-time data into training signals.

[0054] Step 4.2: By connecting to the Tensorflow training interface service, the AI ​​model is trained with the assistance of the MLAgent plugin. Based on sensor data feedback, it predicts the best course of action and completes the multi-agent collaborative navigation task.

[0055] The beneficial effects of this invention are:

[0056] This invention provides a Unity3D-based multi-agent collaborative navigation virtual simulation scheme, achieving low-cost, high-safety agent motion control and optimized collaborative operation efficiency. It also integrates advanced A* search algorithms and artificial intelligence training interfaces, enabling agents to effectively train for obstacle avoidance and path planning in a realistic virtual environment. This improves the agents' adaptability and intelligence level, providing an economical and efficient virtual testing platform for the research and development of agent control algorithms. Attached Figure Description

[0057] Figure 1 This invention provides a schematic diagram of the structure of a multi-agent cooperative navigation virtual simulation system based on Unity3D.

[0058] Figure 2 This is a functional block diagram of the multi-agent cooperative navigation virtual simulation system provided by the present invention;

[0059] Figure 3 A flowchart illustrating a multi-agent cooperative navigation virtual simulation method based on Unity3D provided by this invention;

[0060] Figure 4 This is a schematic diagram of intelligent agent collision avoidance provided by the present invention;

[0061] Figure 5 AI training data flow diagram provided by this invention;

[0062] Figure 6 The Agent lifecycle flowchart provided by this invention. Detailed Implementation

[0063] Specific implementation method one: Combining Figure 1 and Figure 2 This implementation method is described as follows: Figure 1 As shown in the figure, the structure of a multi-agent cooperative navigation virtual simulation system based on Unity3D described in this embodiment includes: a presentation layer, a business layer, and underlying services;

[0064] The presentation layer includes the HDRP virtual simulation environment, the UnityUI interface, and the Unity Editor window, which together provide users with an intuitive operation and monitoring platform. The business layer is the core of the simulation system, handling key business functions such as agent control logic, queue management, path planning, and AI training. This layer is further subdivided into two modules: the Agent control module and the Agent queue control module. The Agent control module is responsible for the control of individual agents, including the management of power data, torque data, path data, and follower data, as well as mode control and algorithm selection. The Agent queue control module manages agent queue data, target data, and queue control, ensuring that multiple agents can work collaboratively. The underlying services provide the necessary low-level support for the simulation system, including Tensorflow training interface services and Plastic SCM project version control, ensuring the training of agent machine learning models and the version management of the project.

[0065] like Figure 2As shown, the functional modules of the multi-agent collaborative navigation virtual simulation system include an environment simulation module, a single-agent control module, a multi-agent control module, a path planning module, and an AI environment module. The environment simulation module is used to construct the virtual environment required for agent interaction. The single-agent control module is used for agent dynamic control, target path planning, and action state control. The multi-agent control module is used for queue mode switching, agent queue management, and queue movement simulation. The path planning module is used to calculate the optimal path for the agent to reach the target location from the current position. The AI ​​environment module is used to train the agent in artificial intelligence and predict the best action plan based on environmental feedback.

[0066] Specific Implementation Method Two: Combining Figure 3-6 This embodiment will be described as follows: Figure 3 As shown in this embodiment, the steps of a multi-agent cooperative navigation virtual simulation method based on Unity3D include:

[0067] S1: Virtual environment construction;

[0068] S101: The first step in building a virtual environment in Unity3D is to obtain 3D models of all objects that the agent will interact with in the scene. This acquisition can be automated using advanced computer vision techniques that extract information from multi-angle images, photographs, or video data of objects, and use methods such as structured light scanning, stereo vision, or structure-of-motion (SFM) to calculate and construct the 3D models of the objects. For models that cannot be obtained automatically, this implementation relies on professional 3D modeling software such as AutoCAD, 3ds Max, and Maya, where professionals manually construct the models based on the object's design blueprints or reference photographs.

[0069] S102: This implementation imports these meticulously crafted 3D models into the Unity3D environment. During the import process, necessary optimization of the models is crucial. This includes simplifying polygons, optimizing textures and materials, and adjusting the model's scale and proportions to ensure they maintain high-quality visuals while efficiently simulating physics and interacting within the virtual environment. This model optimization is essential for ensuring smooth interaction of the agent within the virtual environment.

[0070] S103: This implementation method builds a scene layout in Unity3D, placing the optimized 3D model in appropriate positions within the simulated space to simulate the distribution and relationships of objects in the real world. Furthermore, this implementation method also sets up lighting, weather effects, and physical properties such as gravity and friction in the environment to enhance its realism. Through these meticulous construction steps, this implementation method provides the intelligent agent with a visually realistic and physically accurate operating space, laying a solid foundation for subsequent simulations of collaborative intelligent agent operations.

[0071] S2: Control logic implementation;

[0072] S201: In the Unity3D environment, the control logic of an intelligent agent is precisely implemented through C# scripts, such as... Figure 3 As shown, this process encompasses several key aspects of the agent's interaction with the virtual environment. First, the information acquisition module utilizes Unity3D's sensors and physics engine to enable the agent to perceive changes in its surroundings, such as terrain undulations, weather conditions, and the effects of various sources of resistance. Through integrated APIs, the agent can collect data about its position, velocity, and distances to surrounding objects; this data forms the basis for the agent's decision-making.

[0073] S202: The path planning module employs advanced algorithms, such as the A* search algorithm, to calculate the optimal path for the agent to reach the target location from its current position. This involves analyzing environmental data, including terrain accessibility, obstacle locations, and potential dynamic changes. The path planning algorithm generates a path that takes all these factors into account, ensuring the agent can navigate to the target in the most efficient way while avoiding unnecessary energy consumption and potential collision risks.

[0074] S203: Combining a single-agent control module, a path planning module, and a multi-agent control module, formation control and automatic obstacle avoidance functions are implemented through cooperative algorithms and environmental perception technology. The formation control module enables multiple agents to coordinate their actions to maintain the integrity and efficiency of the formation based on a preset formation or dynamic adjustment strategy. The automatic obstacle avoidance function monitors the agent's surrounding environment in real time. When a potential collision risk is detected, it can quickly calculate an obstacle avoidance path or adjust speed and direction to ensure the agent's safety. The implementation of these control logics provides solid technical support for the efficient and safe operation of agents in virtual environments.

[0075] S3: Obstacle avoidance algorithm integration;

[0076] The integration of obstacle avoidance algorithms is one of the key technologies in intelligent agent virtual simulation systems, such as... Figure 4The core idea of ​​obstacle avoidance and collision avoidance is to define a conflict domain and keep it as far away as possible. This ensures that the agent can effectively navigate autonomously when facing complex environments and diverse sources of resistance. In the Unity3D environment, this implementation uses the A* search algorithm as the core obstacle avoidance mechanism. By adjusting the heuristic function and cost evaluation strategy, the algorithm is optimized to adapt to the dynamic obstacle avoidance needs of the agent. This includes optimizing the preprocessing stage of the algorithm, such as identifying potential obstacles in advance and assessing the passability of the environment, thereby reducing the computational burden on the agent in the simulation environment.

[0077] When adjusting the heuristic function and cost evaluation strategy, to further avoid collisions between multiple agents, consider introducing a repulsive field from the artificial potential field. The repulsive force frep can represent the gradient of Urep, and the repulsive relationship between agent i and agent j in the x-axis direction can be expressed by the following formula.

[0078]

[0079] In formula (1), η is the repulsive force scale factor, which is taken as 10 in the simulation. ij Let Ri be the distance between agents i and j, and Rdetect be the radius of influence of the repulsive field. When the agents are far apart, there is no repulsive influence between them. Similarly, the repulsive relationship between multiple agents in the y-axis direction can be obtained. Finally, the desired velocity components in the xyz-axis direction of the i-th (i>0) agent at time t in the inertial coordinate system are expressed by the following formula:

[0080]

[0081] In formulas (3)-(4), Let be the desired velocity component of the i-th agent in the x-axis direction at time t. Let be the desired velocity component of the i-th agent in the y-axis direction at time t. Let Δt be the desired velocity component of the i-th agent in the z-axis direction at time t, and let Δt be the interval for system state updates.

[0082] For two-dimensional motion, only the xy-axis is written. In this embodiment, the z-axis calculation is written in accordance with formulas (3) and (4), as follows:

[0083]

[0084] The pathfinding process of the A* search algorithm is refined. By integrating the agent's sensor data into Unity3D's physics engine, the algorithm can acquire real-time information on the agent's relative position and distance to surrounding obstacles. This data is used to dynamically update the A* search tree, enabling the agent to quickly respond to environmental changes and adjust its path to avoid newly appearing obstacles during movement. Furthermore, the algorithm considers the agent's dynamic characteristics, such as maximum turning angular velocity and acceleration limits, ensuring that the generated obstacle avoidance path is not only safe but also feasible.

[0085] At the Unity3D implementation level, the A* search algorithm is tightly integrated with the agent's control scripts. Upon receiving a path planning request, the A* algorithm module is activated and begins calculating the optimal path from the current position to the target position. The algorithm's output is a path containing key turning points and velocity suggestions. The agent adjusts its motion state based on this information, achieving smooth and efficient obstacle avoidance navigation. In this way, the agent's actions in the virtual environment more closely resemble real-world behavior patterns, improving the overall realism and reliability of the simulation system.

[0086] S4: Integration of AI training interface;

[0087] The integration of an AI training interface is an innovation in virtual simulation systems, providing agents with the ability to learn and self-optimize. In the Unity3D environment, this implementation uses the MLAgent plugin to implement a flexible AI training interface, allowing researchers to design and implement complex training scenarios. Figure 5 In the data flow diagram, MLAgent acts as an intermediary layer. It receives real-time data from the Unity environment, such as the agent's position, velocity, and distance to surrounding obstacles, and transforms this data into training signals to guide the agent's behavior adjustment.

[0088] Through TensorFlow integration, researchers can design deep learning models to process the vast amounts of data collected by agents. These models are trained with the assistance of MLAgents, learning how to predict the optimal course of action based on environmental feedback. For example, when an agent encounters resistance, the AI ​​model can identify obstacle avoidance patterns through prior training and automatically adjust the agent's path planning strategy to avoid collisions and maintain smooth task execution. This training mechanism allows the agent to gradually adapt to the uncertainty and complexity of virtual environments, improving its autonomous decision-making capabilities.

[0089] Intelligent agents in such Figure 6Throughout its lifecycle, the agent's behavior is continuously optimized through interaction with the virtual environment. MLAgent provides a rich set of sensor and actuator APIs, enabling the agent to perceive and react to environmental changes. Researchers can set reward mechanisms to guide the agent's learning process, such as rewarding the agent for successfully avoiding obstacles or efficiently completing designated tasks. As training progresses, the agent's control strategies gradually mature, enabling it to handle unknown situations more intelligently and exhibiting human-like learning abilities and adaptability. This integrated AI training interface greatly enhances the agent's navigation and operational capabilities in complex environments, providing a solid technical foundation for further research and applications of intelligent agents.

[0090] In summary, this invention provides a low-cost, high-security virtual simulation scheme for optimizing the motion control and collaborative operation efficiency of intelligent agents. It integrates an improved A* search algorithm and an artificial intelligence training interface, providing an economical and efficient virtual testing platform for the research and development of intelligent agent control algorithms.

[0091] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent substitutions, and improvements made to the above embodiments without departing from the scope of the present invention, based on the technical essence of the present invention and within the spirit and principles of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A virtual simulation method for multi-agent cooperative navigation based on Unity3D, characterized in that, include: Step 1: Construct the Unity3D virtual environment required for intelligent agent interaction based on the environment simulation module; Step 2: Combining the single-agent control module, path planning module, and multi-agent control module, set the control strategy for the intelligent agent in the Unity3D virtual environment using C# scripts; Step 3: Using the improved A* search algorithm and the RVO2 simulator concept, combined with the agent's control strategy, control multiple agents to avoid obstacles; Step 3 specifically includes: Step 3.1: After receiving the path planning request, the agent activates the integrated control script of the A* algorithm module and the RVO2 simulator concept to calculate the optimal path from the current position to the target position; Step 3.2: Calculate in x Intelligent agents along the axis i With intelligent agents j The repulsive relationship between them; Step 3.3: Combining with Step 3.2, calculate... y shaft and z Intelligent agents along the axis i With intelligent agents j The repulsive relationship between them; Step 3.4: [The text appears to be incomplete and contains several grammatical errors. A more accurate translation would require xyz Intelligent agents along the axis i With intelligent agents j The repulsive relationship between them is superimposed to obtain the first... i A smart agent t Always in the inertial coordinate system xyz The desired velocity component along the axial direction; Step 3.5: Combine the data collected by Unity3D's sensors to dynamically update the search tree of the improved A* search algorithm, and adjust the maximum turning angular velocity and acceleration of the agent to ensure that the agent moves along the optimal obstacle avoidance path and avoids obstacles. x Intelligent agents along the axis i With intelligent agents j The expression for the repulsive relationship between them is: （1）； In formula (1), The repulsive force scale factor, ij For intelligent agents i and j The distance between them The radius of influence of the repulsive field; No. i An intelligent agent at time t is in an inertial coordinate system. xyz The formula for calculating the desired velocity component along the axial direction is: （3）； （4）； （5）； In formulas (3)-(5), For the first i The expected velocity component of an agent in the x-axis direction at time t. For the first i The desired velocity component of an agent in the y-axis direction at time t. For the first i The desired velocity component of an agent in the z-axis direction at time t. The interval for system status updates; Step 4: Train and deploy the AI ​​model in the Unity3D virtual environment through the AI ​​environment module to control the intelligent agent. By connecting to the Tensorflow training interface service module, use the MLAgent plugin to train the control strategy of the intelligent agent, so that the intelligent agent can successfully avoid obstacles and complete the multi-agent collaborative navigation task.

2. The virtual simulation method for multi-agent cooperative navigation based on Unity3D according to claim 1, characterized in that, Step 1 specifically includes: Step 1.1: Extract information about all objects in the navigation environment from multi-angle images, photos and video data of objects using computer vision technology, and calculate and construct 3D models of all objects using one of the following methods: structured light scanning, stereo vision and structure of motion reconstruction. Step 1.2: Import the 3D models of all objects into the Unity3D environment, and optimize the 3D models of all objects during the import process. The optimization of the 3D models of all objects includes simplifying polygons, optimizing textures and materials, and adjusting the scale and proportions of the models. Step 1.3: Place the optimized 3D models of all objects in their corresponding positions in the Unity3D environment, and set the lighting, weather effects, and physical properties in the Unity3D environment to complete the construction of the Unity3D virtual environment.

3. The virtual simulation method for multi-agent cooperative navigation based on Unity3D according to claim 1, characterized in that, Step 2 specifically includes: Step 2.1: Utilize Unity3D's sensors and physics engine to enable the agent to perceive changes in the surrounding environment, and collect data on the agent's position, velocity, and distance to surrounding objects through the integrated API; Step 2.2: The path planning module uses the improved A* search algorithm to analyze the Unity3D virtual environment data, and combines the collected agent position, speed and distance to surrounding objects to calculate the optimal path for the agent to reach the target position from the current position. Step 2.3: Through the single-agent control module and the multi-agent control module, multiple agents coordinate their actions to maintain the formation according to the preset formation and dynamic adjustment strategy. By monitoring the surrounding environment of the agents in real time, when a potential collision risk is detected, the optimal obstacle avoidance path is calculated and the speed and direction of the agents are adjusted to complete the setting of the agent control strategy.

4. The virtual simulation method for multi-agent cooperative navigation based on Unity3D according to claim 1, characterized in that, Step 4 specifically includes: Step 4.1: In the Unity3D virtual environment, receive real-time data from the sensors through the MLAgent plugin. The real-time data includes the agent's position, velocity, and distance to surrounding obstacles. Convert the real-time data into training signals. Step 4.2: By connecting to the Tensorflow training interface service module, the AI ​​model is trained with the assistance of the MLAgent plugin. Based on sensor data feedback, it predicts the best course of action and completes the multi-agent collaborative navigation task.

5. A Unity3D-based multi-agent cooperative navigation virtual simulation system, applied to the Unity3D-based multi-agent cooperative navigation virtual simulation method described in any one of claims 1-4, characterized in that, include: Presentation layer, business layer, and underlying services; The presentation layer provides users with an operation and monitoring platform; The business layer is used to implement the control logic, queue management, path planning and artificial intelligence training functions of intelligent agents in the simulation environment; The underlying services are used for training intelligent agent machine learning models and managing project versions.

6. A multi-agent cooperative navigation virtual simulation system based on Unity3D according to claim 5, characterized in that, The presentation layer includes an environment simulation module, a UnityUI interface, and a Unity Editor window; The environment simulation module is used to construct the virtual environment required for intelligent agent interaction. The UnityUI interface provides users with an operation page; The Unity Editor window provides a monitoring platform for users.

7. A multi-agent cooperative navigation virtual simulation system based on Unity3D according to claim 5, characterized in that, The business layer includes an Agent control module, an Agent queue control module, a path planning module, and an AI environment module; The Agent control module is responsible for the control of a single intelligent agent, which includes the management of power data, torque data, path data, and follow data, as well as mode control and algorithm selection. The Agent queue control module is used to manage the agent queue data, target data, and queue control. The path planning module is used to calculate the optimal path for the agent to reach the target location from the current location; The AI ​​environment module is used to train the intelligent agent in artificial intelligence and predict the best course of action based on environmental feedback.

8. A multi-agent cooperative navigation virtual simulation system based on Unity3D according to claim 7, characterized in that, The Agent control module includes a single Agent control module and a multi-Agent control module; The single-Agent control module is used for agent dynamic control, target path planning, and action status control. The multi-Agent control module is used for queue mode switching, agent queue management, and queue movement simulation.

9. A multi-agent cooperative navigation virtual simulation system based on Unity3D according to claim 5, characterized in that, The underlying services include the Tensorflow training interface service module and the Plastic SCM project version control module; The Tensorflow training interface service module is used to convert sensor-collected data into training signals to guide the behavior adjustment of the agent. The Plastic SCM project version control module is used for project version management.

Images