PSO-based ANFIS for quadrotor system trajectory-tracking control

Abstract

This paper presents an application of particle swarm optimization algorithm (PSO) to tune the scaling gains of Adaptive Neuro-Fuzzy Inference System (ANFIS) controller with a multi-closed loop control applied to nonlinear quadrotor system. First, PID control approach is used to obtain the training data set that is necessary to ANFIS design. Precisely, only the error and the error rate inputs in addition to output are used in training data set vector. Afterwards, an adaptive Neuro-Fuzzy inference system controller is designed. Then, to tune the scaling gains associated with ANFIS controller, PSO algorithm is used. Finally, an integral control action is added to the ANFIS controller output for controlling each state of the MIMO quadrotor system. The simulation test show that the results obtained with ANFIS-PSO controller are better, when compared to those obtained using conventional ANFIS controller and traditional PID.