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

摘要

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.