U-model based adaptive internal model control for tracking of nonlinear dynamic plants

Abstract

Adaptive tracking of nonlinear dynamic plants is currently an important area of research. The main difficulty being felt b the research community is the lack of a general modelling framework that can facilitate synthesis of a simplistic control law, while being capable of providing accurate approximation of nonlinear systems. The aim of this study is to alleviate that problem by introducing a novel technique based on the control-oriented U-Model for the adaptive tracking of a wide range of stable nonlinear dynamic plants using only input-output data. The overall scheme is based on the robust internal model control (IMC) structure wherein different internal models, using nonlinear adaptive filtering and higher-order neural networks, are proposed. In each case, the U-Model equivalence of the internal model is developed and a simplistic control law based on polynomial root-solving is synthesized. Stability of the proposed adaptive scheme is analyzed with the help of various analytical tools. The effectiveness of the proposed adaptive schemes is demonstrated through simulations and real-time applications to a variety of nonlinear plants that include: Hammerstein model, CSTR model, DC motor and single-link robotic arm manipulator.