Research And Application Of Multi-agent Behavior Simulation Technology In Railway Rescue Simulation Training System

With the rapid development of computer virtual simulation technology,simulation training system is gradually applied to the training of personnel in various industries because of its safety,efficiency,and repeatability.However,the focus of research and development of simulation training systems at home and abroad is to establish a physical simulation environment.Implementing the logic function of each device,there are few researches on the behavioral simulation of the agent in the application scenario.In the virtual environment,the role of the agent in the virtual environment is generally sluggish,the behavior is unreasonable,and the action is unreal and many other problems,which reduces the realism of the simulation training system.Affect training efficiency.This article through the in-depth study of artificial intelligence technology in the virtual environment,according to the railway rescue simulation training system for railway rescue site personnel behavior simulation needs,proposes a multi-agent behavior simulation system program,the intelligence system embedded in the railway rescue simulation training system In this way,it is possible to make the behavior of the agent more realistic and able to interact with the trainees,assist in completing the rescue work,and achieve a good simulation training effect.The main research work and research results of the paper are as follows:1.Design and implement a virtual agent's environmental information awareness model.The model consists of two methods: visual perception and polling awareness.It can ensure that the agent can obtain the required environmental information in real time in the virtual environment and provide the system with the intelligence.Accurate and reasonable information input.2.Design and implement a multi-agent decision model based on hierarchical finite state machines,which divides the standard railway rescue process into states,largely realizes state reuse,and can realize teamwork among different types of work and with trained drivers.Interact with them to help them complete their rescue work.3.Study the path planning method and partial collision avoidance strategy of the agent,use the A* algorithm based on the navigation grid to solve the path planning problem of the agent in the virtual railway rescue scenario,and use the RVO collision avoidance algorithm to realize the dynamics between the agents.Avoid collisions.4.Quantify the behavior of railway rescue site personnel based on research data,design character models and behavioral animations of different types of work,so thatthe behavior of the agent is real and credible,and enhance the realism of the virtual scene.5.The designed multi-agent behavior simulation system is embedded in the railway rescue simulation training system,and test adjustments and experimental analysis are performed.The system function reaches a predetermined target and the training efficiency of the railway rescue simulation training system is improved.