Design And Implementation Of Artificial Intelligence Decision Model Driven By Behavior Tree

In recent years,related software and platforms are developing rapidly with the greatly improved computer processing capabilities.Technologies such as video games and three-dimensional simulation have developed to relatively high levels,especially images and sound effects which are close to the real.At the same time,as an artificial intelligence technology that enhances the appeal of the game and enhances the system experience,it is also placed in an equally important position.The system proposed in this paper is based on the actual game prototype designed.It uses multiple algorithms to handle the interaction between the agent and the person,and the environment.It is driven by behavior trees so that the agent can make relatively reasonable decisions and can generate continuous feedback and context-related re-decisions,and ultimately agent autonomy.This paper studies the problem of agent decision-making under the conditions of voice intervention of external devices in the actual prototype system,and implements the main program based on Unity.The analysis explored the choice of existing engines or components and how to integrate them with the Unity project and designed a system architecture and process flow.Use a related engine to implement a speech recognition plug-in for cross-language calls.Set reasonable parameters to convert external information,including voice,into a linear text format.For the result of speech recognition,we let the program communicate with the server through the local network,use the method of text machine learning,convert the sentence into a list of dense vectors,and then perform similarity calculation and word matching to analyze the system-related implications.Eventually,the agent can make relatively reasonable decisions driven by the behavior tree.