Object behavior simulation based on behavior tree and multi-agent model

摘要

Existing emergency plans and security deployment schemes of public safety events are too subjective and lack of practical testing. In this paper, behavior tree and multi-agent model in game artificial intelligence are applied to modeling the behavior characteristics of scene objects. Behavioral decision is made by the combination of behavior tree and scene object features. Based on that, a natural and intelligent mode of motion is designed by the autonomous multi-agent model. In the experiment, Unity3D engine is used to display the 3D dynamic visualization and the simulation results under the generated deployment scenarios are collected and analyzed. The experiments results show that our model can be used for reasoning and verification of the deployment schemes in advance, and it has a certain guiding role in optimizing the emergency plans and deployment schemes.