Research On Continuous Obstacle Avoidance Control of UAV Tracking Target Based on PID-DDPG Control Method

UAV tracking and obstacle avoidance are important functions of UAVs, and at present, the control system of mainstream UAVs adopts PID control system. However, the traditional PID control can no longer meet the current requirements for faster and more accurate control of UAVs. In this paper, we aim to combine the traditional PID control with deep learning of artificial intelligence to realize the fast tracking and precise obstacle avoidance of the target of UAV in a complex environment. This paper designs a PID-DDPG control model, theoretically, the UAV will prioritize the use of PID control system for fast approach to the selected target, and then switch to the DDPG algorithm to accurately track the target and achieve obstacle avoidance to the environmental obstacles, and ultimately achieve fast tracking and accurate obstacle avoidance to the target in complex environments. Compared with the traditional PID control, the PID-DDPG control system can realize more precise control and faster response to the UAV. At the same time, compared with the pure DDPG deep learning method, this system only needs less learning times and a smaller computation amount.