Hokyung Hwang

Date of Graduation


Document Type



In this research, an integrated method of system identification and control via neural networks and feedback linearization theories was developed. Two different approaches (network and mathematical) are integrated to identify and control nonlinear dynamic systems. The neural network is used as a nonlinear system identifier to extract system equations/parameters with assistance from mathematical concepts such as multi-dimensional curve fitting techniques. Nonlinear feedback controller design based on differential geometry and linear state feedback control theories are then used to design a controller for the identified system. This approach has been investigated on an academic problem, a brushless dc motor, Chua's chaotic circuit, and a dc series motor. The algorithm was successfully tested on a dc series motor in a laboratory setting. Generally, the applicability of this methodology can be extended to a class of dynamic systems that can be described by a set of nonlinear ordinary differential equations in a form of {dollar}\\underline{lcub}\\dot x{rcub} = f(\\underline{lcub}x{rcub},\\underline{lcub}u{rcub}{dollar}), for which measurements sufficient to identify them are available.