Nonlinear identification of control systems

Nonlinear system identification of dynamic systems is a topic of growing interest. The ability of model-based systems analysis and design to attain the desired objectives depends to a large degree on the accuracy of the mathematical models derived and parameters used. To avoid deficiencies associated with model validation and order estimation, accurate mathematical models should be found by using the fundamental laws, and the unknown parameters have to be identified to perform analysis, design, and prototyping. This paper thoroughly studies the model-relevant identification concept for continuous-time nonlinear multi-input/multi-output systems in time-domain by using a nonlinear mapping-based identification framework. Discrete-time and hybrid systems play an important role due to microprocessor-, microcomputer-, and DSP-based hardware platforms applied to implement comprehensive control algorithms in order to attain the desired objectives, specifications, and capabilities. However, the majority of systems to be controlled are continuous-time, and analog control is a manageable and straightforward solution in many practical problems. The importance of the researched time-domain concept lies on the need to identify continuous-time systems to be used in the corresponding analysis, design, stabilization, and optimization. As an illustrative example, the unknown parameters of a high-performance aircraft are identified using the flight data