Author(s)

Chong Cheng, Jiayuan Wang

Corresponding Author

Chong Cheng

Abstract

In response to the problems such as the failure of traditional two-dimensional obstacle avoidance planning in the complex urban building environment, based on theories such as reinforcement learning and deep reinforcement learning, the influencing factors of path planning and obstacle avoidance were analyzed, an energy consumption model for unmanned aerial vehicles was constructed, and an improved reinforcement learning algorithm was proposed, taking into account factors such as environmental wind and air resistance, a dynamic and energy consumption model for unmanned aerial vehicles was constructed, and an improved deep reinforcement learning algorithm was proposed. Through simulation and flight practice, it has been proved that the above algorithms can improve the intelligent obstacle avoidance trajectory planning level of unmanned aerial vehicles in complex urban scenarios.

Keywords

Unmanned Aerial Vehicle, Urban Complex Environment, Intelligent Obstacle Avoidance and Trajectory Planning