机器人对多运动障碍物环境中方向可变运动目标的跟踪

机器人要实现对动态环境中可变方向运动目标的跟踪,必须采用动态的规划算法.本文在快速随机搜索树算法的基础上,采用滚动时间帧的思想,周期性地采集环境信息与目标运动状况,来预测未来环境中障碍物的分布情况及运动目标位置.在每个周期内用B IAS_RRT来引导机器人行走,以适应障碍物与目标运动方向的变化.仿真结果表明,该算法能有效跟踪在多运动障碍物环境中方向可变的运动目标.

Changeable moving-goal tracking for robots in the environment of dynamic multi-obstacles

When a robot is required to track a moving object in dynamic environment,a dynamic algorithm must be taken.An algorithm called rolling timeframe biased rapidly-exploring random tree is proposed in this paper.Based on the analysis of the stochastic characteristics of rapidly-exploring random tree,a parameter called bias is introduced to speed up the search.Taking advantages of the rolling timeframe,robots collect the information of dynamic obstacles and object at the beginning of a period,and estimate their distributions in operation space of next timeframe.The robot plans local path using biased rapidly-exploring random tree algorithm.After many times of such local planning,the robot gets its object at last.Simulation results show that the algorithm can obtain good results in tracking the object with changing direction in dynamic environments.