Discrete-time sliding mode control and its application to an autonomous underwater vehicle

Abstract

This thesis presents a discrete-time quasi-sliding mode controller of SIMO systems with parameter uncertainties and long sampling interval based on variable structure system. Although conventional sliding mode control techniques are robust to system uncertainties, the sliding control system reveals chattering phenomenon and even makes the system unstable when the system has long sampling interval. The discrete quasi-sliding mode control system is designed to track the desired path on the basis of a Lyapunov function, and sufficient conditions of the switching gain are given to make the discrete system stable. Each component of the switching gain can be determined separately one another. The switching region is variable according to the product of the bounds of uncertainty and the amplitude of the trajectory. The controller is robust to the uncertainties, and the reaching condition of the control system is satisfied for any initial condition. This control law is a generalized form of the discrete quasi-sliding mode control and reduces the chattering problem considerably. This thesis also points out that the VSS type self-tuning controller of Furuta``s method can be improved by modifying the sector with separate gains and with equivalent control algorithm. The control system is stable in the outside of the sector and ultimately bounded in the inside of the sector. The proposed VSS type self-tuning controller prevents the estimated parameters from the fluctuation which can occur during the parameter adaptation. This controller with reduced switching sector is actually applicable to the tracking control of discrete systems in the presence of parameter uncertainties. The dynamic characteristic of an autonomous underwater vehicle is introduced and a model is derived for the AUV``s motion control in vertical plane. Depth keeping control and contouring control are performed for a numerical model of AUV with full nonlinear equations of motion to verify the effectiven...