Neural network based adaptive nonlinear model inversion control of a twin rotor system in real time

Abstract

A real time adaptive nonlinear model inversion controller is implemented on a twin rotor multi-input-multi-output system (TRMS) using artificial neural networks. The TRMS is an experimental aerodynamic test bed representing the control challenges of unmanned aerial vehicles. In this study a nonlinear dynamic model of the system is considered and a nonlinear inverse model is used for the pitch channel of the system. The feedback control system and an adaptive neural network element are integrated to compensate the possible model inversion errors. The proposed online learning algorithm updates the weights and biases of the neural network using the error between the setpoint and the real output. Square and sinusoidal reference command signals are used to test the performance of the controller. It is noted that a reasonable tracking response is exhibited in the presence of inversion errors caused by model uncertainty. This paper is the experimental study of the previous work by the same authors in which the proposed method has only been verified by simulation results.