Directional-Spaces Trees for Efficient Robot Navigation in Dynamic Route Environments

This paper addresses the problem of path planning for robot navigation in dynamic route environments. Recently, EST （expansive-spaces trees） which is a single-query sampling-based path planner has been applied to robot navigation because of its advantages such as simple implementation and ability to solve the kinodynamic problem. In route environments, however, the performance of the EST-based robot navigation degenerates because the EST plans a path without considering the directionality to a goal. Moreover, if there are dynamic obstacles, its performance becomes worse because the EST plans a path without any consideration for dynamic obstacles. This paper proposes DST （directional-spaces trees）, which is a variant of the EST to improve the performance of robot navigation in dynamic route environments. The DST can plan a more efficient path based on the polar coordinate system to increase the directionality to the goal of robot navigation. Also, the DST can replan the path using the concept of VO （velocity obstacle） to avoid not only static obstacles but also dynamic obstacles. These two variations enable the robot navigation in dynamic route environments to be more efficient and safe. Simulation results reveal that the DST can plan a more efficient path than the EST and replan the more efficient path than other path replanning techniques in dynamic route environments.