D-Star-Based Optimized Trajectory Planner for Mobile Robots Operating in Dense Environments

This paper proposes the path planner module designed for the robots operating in the dense dynamic environments. It contains description of the novel ROS module that implements adopted version of the known and well described D-star (D*) algorithm accompanied with customized heuristic, and optimizer pipeline for trajectory decompaction, and postprocessing engine for holding a safety distance from the obstacles. The module removes the known problems that D-star experiences on high-resolution maps, providing sparse trajectory useful for the industrial unmanned ground vehicles (UGVs) with simple actuation trackers installed. Test results of the algorithm are given for the dense industrial environment containing closed loops.