The design of UAV controller has always been a hot topic in related fields, but the traditional controllers have their own limitations: For example, the most classic PID controller has insufficient adjustment ability. For this reason, fuzzy PID controller has been researched, whose anti-interference ability is stronger, but the rule base has many parameters and it is difficult to select. At the same time, reinforcement learning has emerged in some control fields. This paper uses Gymfc as the simulation environment, and takes the fuzzy PID controller as the framework for the mission of UAV to achieve the goal altitude. Through reinforcement learning algorithm A2C, it realizes the self-tuning of the rule base, thus further completing the self-tuning of PID parameters.
Fuzzy PID Controller for UAV Based on Reinforcement Learning
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 : 160 ; 1724-1732
2023-03-10
9 pages
Article/Chapter (Book)
Electronic Resource
English
Improve PID controller through reinforcement learning
IEEE | 2018
|Interpreting data of reinforcement learning agent controller
European Patent Office | 2023
|