Artificial Potential Field Based Real-time Obstacle-free Motion Planning for Unmanned Aerial Vehicle
Date of Award
Doctor of Philosophy (PhD)
Mechanical and Aerospace Engineering
APF, Motion Planning, Obstacle-free, Real-time, RL, UAV
Unmanned Aerial Vehicles (UAVs) are widely used worldwide for a board range of civil and military applications. A smart UAV system should be capable of fully controlling itself autonomously without human pilot interference. Moreover, it also needs to control the generation and maintenance of the fleet formation for a swarm of UAVs. The real-time obstacle-free motion planning problem is another important issue for UAV autonomous control. Due to limited onboard computing power, the UAV path planning algorithm needs to be concise but powerful to handle the vehicle’s collision-free flight in uncertain environments. Above all, the autonomous control system should also guarantee the safety of the UAV.
The Artificial Potential Field (APF) approach provides a simple and effective motion planning method for autonomous UAV control. However, the Artificial Potential Field approach has a major problem; the UAV may be trapped at local minimums (due to the presence of several potential fields generated by obstacles) before reaching its goal. In this dissertation, the Reinforcement Learning (RL) based Dynamic Artificial Potential Field (DAPF) method and Modified Formation Potential Field (MFPF) are proposed and implemented to address this issue, and to enable onboard real-time obstacle-free motion planning for a small UAV system.
The RL based DAPF algorithm has been tested in different simulation environments. The tests demonstrate high efficiency and reliability of the RL based DAPF algorithm. The proposed MFPF has also been tested in both 2D and 3D environments. It has been shown that the MFPF algorithm is capable of guiding multiple UAVs through an unknown or a partially known environment, avoiding internal and external collisions, while generating and maintaining a given formation.
SURFACE provides description only. Full text may be available to ProQuest subscribers. Please ask your Librarian for assistance.
Yin, Hang, "Artificial Potential Field Based Real-time Obstacle-free Motion Planning for Unmanned Aerial Vehicle" (2020). Dissertations - ALL. 1292.