Real-time path planning of controllable UAV by subgoals using goal-conditioned reinforcement learning

GyeongTaek Lee, Kang Jin Kim, Jaeyeon Jang

Abstract: Highlights•We propose a real-time autonomous path planning of UAVs in unknown environments.•We aim to train the UAV agent to reach various goals that have never been visited in the training.•The UAV agent is trained to learn basic maneuvers using goal-conditioned reinforcement learning.•We assume the specific scenarios that the UAV agent should perform difficult missions in unknown environments.•Experimental results show that the UAV agent can perform difficult maneuvers in real time in unknown environments.