Path-tracking Control of Quadrotor Using Adaptive Critic-based Neurofuzzy Controller

The objective of this paper is to propose Adaptive Critic-based Neurofuzzy Controller for quadrotor’s path-tracking. This controller is based on Reinforcement Learning and its advantage is that the parameters are tuned automatically by a critic agent. The controller consists of two parts, a Takagi-Sugeno-Kang fuzzy controller and a PD critic agent. The performance of the controller was then compared with Sliding Mode Control by conducting two simulation tests. The simulation results depict that the Adaptive Critic-based Neurofuzzy Controller can follow the desired trajectory and stabilized the system with presence of disturbance.