Data-Driven Model-Free Adaptive Predictive Control and its Stability Analysis

Abstract: In this paper, a novel full form dynamic linearization (FFDL) data-driven model-free adaptive predictive control (MFAPC) method is proposed for a class of discrete-time single-input single-output nonlinear systems. The novelty of MFAPC is that preliminary physical model and the Lyapunov stability theory are not required for the controller design and theoretical analysis. Instead, the proposed MFAPC only uses a new dynamic linearization method called pseudo-gradient (PG) vector, which is merely related to the input/output (I/O) measurement data. The main contributions of this paper are: First, a novel MFAPC with adjustable parameters is proposed; Second, we have proved the bounded-input bounded-output stability, the monotonic convergence of the tracking error, and the internal stability of the proposed method. Third, the proposed MFAPC can be considered as an elegant extension of the current MFAC. The simulations have been carried out to verify the effectiveness of the proposed MFAPC.