基于有限状态机的类人足球机器人决策系统设计

Design of Human Soccer Robot Decision System Based on Finite State Machine

RoboCup类人组(Humanoid League)机器人的决策系统是基于单目视觉自主决策的核心子系统,是决定比赛胜败的关键因素之一。本文以高效性为原则,设计了基于有限状态机的新决策系统,将复杂的机器人决策行为控制层划分为可以独立调试和优化的4个状态机模块,使机器人各状态之间的转换更加灵活,决策效率极大提高,决策算法更加易于继承和开发。将该决策系统应用于类人足球机器人平台,通过试验及比赛证明了其有效性和高效性。

The decision system of the RoboCup Humanoid League is an autonomous decision-making system based on independent monocular vision,which is the core subsystem of the robot and determines the success or failure of the game.Based on the principle of high efficiency,a new decision system based on Finite State Machine (FSM) is designed to replace the original decision system.The behavior control layer in complex robot decision-making is divided into four state machine modules,which can be debugged and optimized independently.The state transition between robots is more flexible,and the efficiency of the system is improved greatly.Moreover,the decision algorithm is inherited and developed easily.The decision-making system is applied to the humanoid soccer robot platform,and its effectiveness and efficiency have been proved by experiment and match results.