Quadrotor motion control using deep reinforcement learning
Authors:
- 1 December 2021
- journal article
- research article
- Published by Canadian Science Publishing in Journal of Unmanned Vehicle Systems
- Vol. 9 (4) , 234-251
Abstract
We present a deep neural-net-based controller trained by a model-free reinforcement learning (RL) algorithm to achieve hover stabilization for a quadrotor unmanned aerial vehicle (UAV). With RL, two neural nets are trained. One neural net is used as a stochastic controller, which gives the distribution of control inputs. The other maps the UAV state to a scalar, which estimates the reward of the controller. A proximal policy optimization (PPO) method, which is an actor–critic policy gradient approach, is used to train the neural nets. Simulation results show that the trained controller achieves a comparable level of performance to a manually tuned proportional-derivative (PD) controller, despite not depending on any model information. The paper considers different choices of reward function and their influence on controller performance.Keywords
This publication has 19 references indexed in Scilit:
- Incremental model based online dual heuristic programming for nonlinear adaptive controlControl Engineering Practice, 2018
- Mastering the game of Go without human knowledgeNature, 2017
- Control of a Quadrotor With Reinforcement LearningIEEE Robotics and Automation Letters, 2017
- Inner-outer loop control with constraints for rotary-wing UAVs2015
- Handbook of Unmanned Aerial Vehicles2015
- High-performance trajectory tracking control of a quadrotor with disturbance observerSensors and Actuators A: Physical, 2014
- Reinforcement learning in robotics: A surveyThe International Journal of Robotics Research, 2013
- Introduction to feedback control of underactuated VTOLvehicles: A review of basic control design ideas and principlesIEEE Control Systems, 2013
- Autonomous Helicopter Aerobatics through Apprenticeship LearningThe International Journal of Robotics Research, 2010
- OnActor-Critic AlgorithmsSIAM Journal on Control and Optimization, 2003