基于模糊逻辑的多移动机器人自适应协作围捕

Adaptive cooperative hunting for multiple mobile robots based on fuzzy logic

提出一种在未知动态环境下实现多移动机器人自适应协作围捕运动目标的整体方案,为成功实现围捕,设计了基于模糊逻辑的钳型夹击策略,模糊规则通过遗传算法学习获得.同时为躲避围捕过程中遇到的动态随机障碍,提出了基于碰撞风险的随机避障策略.围捕机器人的综合行为通过融合避障行为、合围行为和抓捕行为获得.在MRS仿真环境下进行了模拟实验,获得的不同条件下的围捕结果证明了围捕策略的有效性和鲁棒性.

A general scheme of adaptive cooperative hunting for a moving target by multiple mobile robots in unknown dynamic environments was presented. To realize successful hunting, pincer attack strategy was proposed and its behavior module was described by fuzzy logic, which was learned by a genetic algorithm. Simultaneously, obsta- cle avoidance scheme based on collision risk was used to avoid random obstacles during hunting. The synthesized behavior was obtained by combining avoidance behavior, formation behavior, and approach behavior. Simulated hunting experiments were conducted under different conditions in the Microsoft robotics studio and simulation results demonstrate the effectiveness and robustness of the proposed scheme.