基于极限环的多机器人协作围捕算法

A distributed multi-robot cooperative hunting algorithm based on limit-cycle

多机器人协作围捕问题是检测多智能体系统有效性的重要手段,针对该问题介绍了一种基于极限环的多机器人协作围捕算法,算法采用分布式结构,单个机器人独立控制自身的围捕行为,机器人团队则利用机器人之间的相互作用来达到整体的协调行为。算法一方面利用人工势场完成多机器人的队形控制,另一方面,利用极限环保证机器人团队向目标的逼近,相比于其它方法,基于极限环的算法不仅能够实现围捕机器人对目标的包围,还能使围捕机器人在形成包围圈后沿着以目标为圆心的圆环做等间距的环绕运动,这种包围方式将更有利于防止目标的逃走。仿真实验验证了该算法对于解决多机器人协作围捕问题的有效性和可靠性。

Multi-robot cooperative hunting problem is an important way to validate multi-agent systems.This paper introduces an algorithm for predator prey pursuit problem based on limit-cycle.The algorithm is distributed,in which each robot is controlled locally according to sensor information,and achieves concerted behavior by mutual effect among the team.On the one hand,the algorithm uses artificial potential field for multi-robot formation control.On the other hand,limit-cycle based method guarantee the robot team to move toward the target.Using such limit-cycle based method,predators will move around the prey and keep a given distance between each other after encircling,so as to avoid escape of the prey more effectively and efficiently.In addition,the paper designs a heading direction control law for obstacle avoidance,and solves the problem of robot freezing caused by counteraction between repulsion and attraction.Simulation results show the effectiveness and reliability of the proposed method in cooperative hunting control.