Aiming at the autonomous obstacle avoidance problem of UAV in multi-obstacle map environment, a UAV obstacle avoidance algorithm based on the improved Q learning method is proposed. By analyzing the UAV dynamics principle, the UAV kinematic model is built, and the Markov jump system model is further obtained. Considering the safe distance from the obstacle and the position of the target point, an improved immediate reward function is presented, and a Q learning algorithm of UAV obstacle avoidance is proposed by adopting the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varepsilon $$\end{document}-greedy strategy, which can improve the learning efficiency, realize autonomous obstacle avoidance and optimize the route to the target position. In the simulation experiment, the UAV can track with down different environments and the accumulative rewards are compared and analyzed, which show the effectiveness and advantages of the UAV self-learning algorithm proposed in this paper.
Multi-obstacle Avoidance of UAV Based on Improved Q Learning Algorithm
Lect. Notes Electrical Eng.
International Conference on Autonomous Unmanned Systems ; 2022 ; Xi'an, China September 23, 2022 - September 25, 2022
Proceedings of 2022 International Conference on Autonomous Unmanned Systems (ICAUS 2022) ; Chapter : 6 ; 59-71
2023-03-10
13 pages
Article/Chapter (Book)
Electronic Resource
English
UAV obstacle avoidance , Q learning , <inline-formula><alternatives><math xmlns="http://www.w3.org/1998/Math/MathML"><mi>ε</mi></math><tex-math>\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\varepsilon $$\end{document}</tex-math><inline-graphic></inline-graphic></alternatives></inline-formula>-greedy strategy Engineering , Control, Robotics, Mechatronics , Computational Intelligence , Communications Engineering, Networks , Artificial Intelligence
Neural Q Learning Algorithm based UAV Obstacle Avoidance
IEEE | 2018
|Multi-UAV Cooperative Obstacle Avoidance Trajectory Planning Method Based on Improved RRT Algorithm
Springer Verlag | 2025
|European Patent Office | 2021
|