Nonlinear System Identification — A Continuous-Time Bilinear State Space Approach

摘要

A class of nonlinear systems can be approximated by a bilinear system to an arbitrary high order as needed. In this paper, an improved method is developed that uses single experiment data with multiple pulses for bilinear system identification. The main objective of this improved method is to identify a nonlinear system that is approximated by a bilinear model. An advantage of this improved method is that the system identification process may be done on line and in real time. Numerical examples such as the Euler’s equation describing the rotation of a rigid body are given to demonstrate the power of the bilinear approach for nonlinear system identification.