Dynamical linearization based PLS modeling and model-free adaptive control

In this paper, a new model free adaptive control (MFAC) strategy based on partial least squares (PLS) framework is proposed to achieve trajectory tracking for multivariable nonlinear processes. The nonlinear dynamic characteristics of the multivariable systems are addressed by a dynamic linearization method and a linear PLS inner data model is obtained conse-quently including an unknown pseudo-partial derivative (PPD) parameter. Under the PLS framework, the multivariable system can be decomposed into multiple single-loop systems to facilitate the controller design. The controller design only depends on the measured input and output data. Simulation results demonstrate the effectiveness of the proposed method.