IJCAS

The current study presents a data-driven integral sliding mode predictive control method for a category of discrete-time repetitive nonlinear systems. At first, a compact form of iterative dynamic linearization (IDL) technology is utilized to establish an IDL data model. Then, considering the time and iterative domain simultaneously, an iterative integral sliding mode surface is constructed to establish an iterative integral sliding mode controller. The stability of the presented control strategy is then demonstrated through a precise mathematical analysis. Furthermore, to further reduce the control error, an iterative integral sliding mode predictive control strategy is established using the model predictive control. Since the proposed method is a data-driven control scheme, it only employs the online I/O data for parameter estimation and controller design. The effectiveness and monotonic convergence of the proposed schemes are evaluated through simulations. Comparative results with the data-driven optimal iterative learning controller (DDOILC) and the enhanced DDOILC indicate that the presented controller can provide a faster convergence and less tracking error.

Keywords: Dynamic linearization, integral sliding mode, iterative learning control, nonlinear system.