Key points are not available for this paper at this time.
Motion planning in dynamic environments is an important issue for autonomous mobile robot, and has been extensively studied for two decades. However, most of the existing methods repeat planning in short time, only looking around the current position of obstacles in the environment. In other words, they do not take into account the dynamic feature of environment. In our approach, we consider that the environmental changes are predictable in a short time interval, and regard the motion planning problem in a known dynamic environment as temporal spatial path planning problem. StRRT makes it possible to find a practically reasonable path for the robot with the limits of velocity and acceleration. It is also possible to choose the best among the paths by incrementally building a search tree in the allocated time. We demonstrate, using simulations, that StRRT can generate better trajectories in dynamic environments in comparison with RRT and a potential field approach. We also analyze the function of StRRT, considering perception and action constraints.
Sakahara et al. (Mon,) studied this question.