提出一种基于极坐标空间的、以机器人期望运动方向角为路径优化指标的动态不确定环境下移动机器人的在线实时路径规划方法.该法通过机器人的传感器系统,实时探测局部环境信息,在每一采样时刻,机器人首先对"视野"内的动态障碍物的位置进行采样,然后根据所采样的位置信息,利用自回归模型预测出下一采样时刻动态障碍物的位置,再将预测位置上的动态障碍物当作静态障碍物来处理,然后对其规划避碰路径,从而将动态路径规划转化为静态路径规划.仿真和实验结果验证了该方法有效可行,具有实时规划性和良好的避障能力.%This paper presents an on-line real-time path planning method for mobile robots in dynamic uncertain environment. Based on polar coordinates space, this method uses expected movement direction angle of robots as the index of path optimization. It explores information of local environment in time through sensor system of robots. At each sampling time, the robot firstly takes the sample of positions of dynamic obstacles within its ”visual field”, and then predicts the positions of dynamic obstacles at next sampling time with autoregressive model according to sampled positions information, and deals with dynamic obstacles at predicting positions as the static ones afterwards. Then it plans collision avoidance path for them so as to have transformed the dynamic path planning into static path planning. The method is available and feasible, it has real-time planning property and preferably good collision avoidance capacity, all of these are demonstrated by the simulations and experimental results.
展开▼
