Data-based Formation Control for Underactuated Quadrotor Team via Reinforcement Learning
Hao Liu, Wanbing Zhao, Frank L. Lewis, Zhong‐Ping Jiang, Hamidreza Modares · 2020
In this paper, the formation problem of unknown multi-quadrotor systems with underactuation and nonlinearities is addressed. A formation controller including a position controller and an attitude controller is designed. The designed formation controller is based on hierarchical scheme and reinforcement learning method is used to learn the control weights of the formation controller. A simulation of formation of multiple quadrotor systems shows the effectiveness of the proposed controller.