Non-linear model inversion control for air vehicle system using neural networks and particle swarm optimization

Abstract

This paper investigates the development of a non-linear model inversion controller for air vehicle system utilizing artificial neural networks and particle swarm optimization (PSO). An adaptive neural network element is integrated with feedback control system to compensate for model inversion errors. Two control structures: direct inverse model and internal model control are considered in this work. A twin rotor multi-input multi-output system (TRMS) is considered as a test rig for air vehicle system. A non-linear inverse model is developed for the pitch movement of TRMS utilizing artificial neural network (ANN) which is used as one of the main components of control system for input tracking. A relatively new population-based, self-adaptive optimization, PSO, is used to train the inverse model in order to avoid premature convergence to local minima. The control scheme shows good tracking capability with satisfactory level of rise time, settling time and steady state error.