UAV Path Planning System Based on 3D Informed RRT* for Dynamic Obstacle Avoidance

Abstract

A path planning system based on the Informed RRT* path planner was developed to enable an unmanned aerial vehicle (UAV) to avoid moving obstacles in a cluttered 3D environment. For congested environments such as a construction site, path planning systems that help a UAV to safely manoeuvre around dynamic objects and potential co-workers operating within the same workspace is needed. Instead of using a general RRT* path planner approach which will generate a sinuous path, we proposed a flexible approach to increase the convergence of our path planner by re-defining the search space based on 2D Informed RRT* path planner. General RRT* has a relatively low convergence speed to optimize its original solution. By using motion tracking cameras, we obtained real-time feedback of the UAVs pose as well as map structuring and obstacle positions. With this setup, the performance of our proposed path planning approach was assessed using a set of diverse scenarios to compare against general RRT* in convergence rate, quality of solution and ability to handle multiple obstacle situation.