Path planning combining improved rapidly-exploring random trees with dynamic window approach in ROS

Abstract

Path planning is a fundamental research area in robotics. Sampling-based methods have been extended further away from basic robot planning into further difficult scenarios and diverse applications for their efficient solution. Issues occurred where they only offer a path rather than velocity commands; the changing situation of environment is hard to tackle. This paper presents a novel integrated approach of creating path for robot navigation with dynamic constrains. In the proposed algorithm, a biased Rapidly-Exploring Random Trees (RRT) is utilized to find a global path in the configuration space. Then, the Dynamic Window Approach (DWA) is performed over the path to calculate translational and rotational velocity commands for the robot. Performance of the proposed method is tested and validated using Robot Operating System (ROS). Simulations show that the proposed methodology achieves efficient and smooth path for the robot under the dynamic constrains.