Neural-network-predictor-based control for an uncertain multiple launch rocket system with actuator delay

Abstract

Development of multiple launch rocket system (MLRS) has been restricted for several decades due to the poor dispersion characteristics of rockets, which is caused by the orientation of the MLRS departing from that intended. Hence, it is vital to maintain the angles of MLRS at a desired value via a proper control strategy. In this paper, a new neural network predictive control is developed for orienting control of the MLRS with actuator delay. First, the dynamic model of motor-mechanism coupling system is established using Lagrange method and field-oriented control theory. Then, for cancelling the effects of nonlinearities and uncertainties, the concept of feedback linearization and a dynamic recurrent neural network are introduced. In addition, a modified Smith predictor is employed to maintain the desirable orienting performance in the occurrence of actuator delay. For the stability analysis, Lyapunov’s method is utilized to ensure uniform ultimate boundedness of the closed-loop system. The simulated and experimental results demonstrate the effectiveness of the proposed controller.