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

Authors

  • Shiyu Wang

DOI:

https://doi.org/10.54097/w85nvn58

Keywords:

UAV; reinforcement learning; PID control; obstacle avoidance.

Abstract

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.

Downloads

Download data is not yet available.

References

H. H. Zhang, X. S. Gan, Y. Mao, et al. A review of obstacle avoidance algorithms for UAVs[J]. Aviation Arms,2021,28(05):53-63Cheng Yibin,Tan Zhiyin. Analysis of UAV flight tracking control based on improved neural network PID control[J]. Journal of Lanzhou Institute of Technology,2018,25(05):69-72.

CHENG Yi-Bin,TAN Zhi-Yin. Analysis of UAV flight tracking control based on improved neural network PID control[J]. Journal of Lanzhou Institute of Technology,2018,25(05):69-72.

Yang Hui. A cooperative tracking algorithm for multiple moving targets based on remote sensing UAV[J]. Automation Applications,2022(08):1-3+8.DOI: 10.19769/j.zdhy.2022.08.001.

CHENG Yi-Bin,TAN Zhi-Yin. Analysis of UAV flight tracking control based on improved neural network PID control[J]. Journal of Lanzhou Institute of Technology,2018,25(05):69-72.

GUO Qianqian,ZHOU Xiaohui. Automatic tracking system of UAV antenna based on GPS[J]. Computer Application and Software,2017,34(06):85-90.

JORDAN M I, MITCHELL T M. Machine learning: trends, perspectives, and prospects [J]. science, 2015, 349(6245):255260

KOBER J, BAGNELL J A, PETERS J. Reinforce ment learning in robotics: A survey[J]. International Journal of Robotics Research, 2013, 32 (11): 12381274

Downloads

Published

26-06-2024