In this paper, we implemented a system that arrives at a target by avoiding complex installations with only the sensors mounted on the drone in an environment in which the drone operates in urban. The system is designed to operate autonomously using proximal policy optimization (PPO) reinforcement learning, and the simulation performance of autonomous drone operation according to various sensor conditions and various terrain complexity was analyzed.
ACKNOWLEDGMENT This work was supported by Institute of Information communications Technology Planning Evaluation (IITP) grant funded by the Korea government(MSIT) (No. 2021-0-00794, Development of 3D Spatial Mobile Communication Technology).