This paper presents a neural network-based digital redesign approach to design digital PID controllers for continuous-time noise-free nonlinear multivariable systems with known state dimension but unknown structures and parameters. Important features of the approach are: (i) it generalizes the existing optimal linearization approach to the models which are nonlinear in both the state and the input; (ii) it develops a neural network-based optimal linear state-space model for unknown nonlinear systems; (iii) it develops an inverse digital redesign approach for indirectly estimating an analog PID control law from a fast-rate optimal digital PID control law, without directly utilizing the analog models. This analog control law is then converted to a slow-rate digital PID control law via the prediction-based digital redesign method; (iv) it develops a linear time-varying piecewise-constant low-gain tracker which can be implemented using microprocessors. A nonlinear synchronous motor is utilized as a simulation example to demonstrate the effectiveness of the proposed methodology.
|Number of pages||32|
|Journal||Dynamics of Continuous, Discrete and Impulsive Systems Series B: Applications and Algorithms|
|Publication status||Published - 2007 Jun 1|
All Science Journal Classification (ASJC) codes
- Discrete Mathematics and Combinatorics
- Applied Mathematics