Neural network NARMA-L2 control of a Twin Rotor MIMO System

Abstract

This work presents the design of a nonlinear autoregressive-moving average controller (NARLA-L2) applied to a TWIN ROTOR MIMO SYSTEM (TRMS). This system simulates the dynamic of a helicopter, which is a kind of UAVs; they are the interest of nowadays researches, in particular the control of this system. Because of the difficulty presented in modeling of such non linear systems, a non model control technique is used -to avoid this hard and not exact task-based on neural network structure.So this method is applicable for the control of this type of complex nonlinear system, and its strength is that it is not based on physical modeling that is often far from the real system, thus the ease of synthesizing the corrector with good performances in trajectory tracking and disturbances rejection.