受决策过程神经学机制启发的机器人行为协调

Robot Behavior Coordination Inspired by Decision-Making Neurophysiological Mechanism

行为协调是基于行为的机器人控制系统中的一个主要问题。受决策过程的神经学机制启发,将一个或多个传感器的信息以及不同时间点上的信息组合,并累积得到似然比来表示行为选择的相对可能性,然后根据行为的可能性排序来决定机器人行为的取舍。通过一个机器人推棒子的仿真实验验证了该方法的可行性。

Behavior coordination is important to the design of behavior-based robot. Inspired by neurophysiological mechanism of decision-making, the relative probability of behaviors is expressed by likelihood ratio, which accumulates information coming from both a variety of sensors and different times. Then, the robot＇s behavior is chosen according to the sequence of behaviors＇ relative probability. The simulation of pushing stick shows the validity of the proposed method.