Abstract:

Aiming at the problems of local deadlock and path redundancy of the mobile robots running in the complex and unknown environments, a local path planning method based on prediction and fuzzy control is proposed. The decision-making strategy is also given to solve the problems that existed in the planning process. Firstly, the common obstacle avoidance behavior and the walking along the wall behavior are integrated into a fuzzy controller, and the behavior alternation are realized through internal rules. Secondly, aiming at the deadlock problem of multi-U obstacles existed in fuzzy control rules, a judgment method of cumulative corner sum is proposed to help the robot escape from the deadlock state. Then, the trap prediction mechanism is added to overcome the limitation of the sensor measurement range to a certain extent, and predict the feasible path in front of the robot for making decisions to reduce redundant sections. In addition, a finite state machine based on artificial potential field method is designed to solve the path redundancy caused by the symmetry of obstacles. Finally, the simulation tests are carried out on the MATLAB platform to verify the feasibility and validity of the designed methods.

Key words: mobile robot, local path planning, fuzzy control;trap prediction mechanism, multiple U-shaped obstacles, finite state machine

摘要：

针对移动机器人在复杂未知环境下运行时易出现局部死锁和路径冗余的问题，提出了一种结合预测和模糊控制的局部路径规划方法，并对规划过程中出现的问题提出了解决策略。将普通避障行为与沿墙走行为融合为一个模糊控制器，通过内部规则实现行为交替。针对模糊控制方法存在多U型障碍物的死锁问题，提出了累加转角和的判断方法，帮助机器人逃脱死锁状态。加入陷阱预测机制，使机器人能在一定程度上克服传感器测量范围的局限性，预测前方是否可行并做出决策，减少冗余路段。设计了基于人工势场法的有限状态机，解决了因障碍物对称而无法确定方向所产生的路径冗余。在MATLAB平台中进行了仿真测试，验证了所设计方法的可行性和有效性。

关键词: 移动机器人, 局部路径规划, 模糊控制, 陷阱预测机制, 多U型障碍物, 有限状态机