
Drones, Journal Year: 2024, Volume and Issue: 8(9), P. 516 - 516
Published: Sept. 23, 2024
Autonomous navigation of Unmanned Aerial Vehicles (UAVs) based on deep reinforcement learning (DRL) has made great progress. However, most studies assume relatively simple task scenarios and do not consider the impact complex UAV flight performance. This paper proposes a DRL-based autonomous algorithm for UAVs, which enables path planning UAVs in high-density highly dynamic environments. state space representation method that contains position information angle by analyzing changes performance In addition, reward function is constructed non-sparse to balance agent’s conservative behavior exploratory during model training process. The results multiple comparative experiments show proposed only best but also optimal efficiency
Language: Английский