Observer-Based Enhanced ANFIS Control for a Quadrotor UAV

Abstract

This paper introduces a new approach for optimizing the Adaptive Network Fuzzy Inference System (ANFIS) with the (metaheuristic) particle swarm optimization (PSO) algorithm to solve a trajectory tracking problem for a UAV quadrotor system. In this work, the input-output dataset is collected from the quadrotor's time response from a Proportional Integral Derivative (PID) controller. Then, before its use in the ANFIS algorithm, the collected dataset is optimized using the PSO algorithm. Moreover, an integral control action is integrated into the proposed ANFIS controller to ensure disturbance rejection. A suitable high gain observer is utilized in estimating unmeasured states for translational and rotational motions of the quadrotor system. Compared to a conventional ANFIS and traditional PID controllers, the results of two simulation tests demonstrate the efficiency and robustness of the proposed new output-feedback PSO-ANFIS controller.