Control design for multivariable nonlinear systems has cultivated into a mature research area. A variety of control design methodologies have been well established. In most of the approaches, however, it is implicitly assumed that an analytically mathematical model can be available. This is not always feasible since in many practical control problems, system dynamics may be represented by nonanalytical modules, such as look-up tables, c or Fortran codes, etc. As a consequence, conventional methods can only be performed after analytical models are obtained. In this paper, an approach is proposed where the nonanalytical modules can be handled directly. Important results are obtained on optimal control laws and their implementation; robust control is achieved using a new online tuning method; input saturation and stability issues are also discussed. A numerical study is provided to validate the effectiveness of the proposed method.
NARX Prediction-Based Parameters Online Tuning Method of Intelligent PID System