Abstract
A computer-aided method is presented in this paper for optimum discretization of continuous data control systems using a time domain optimization technique.
Digital controllers with different orders can be assumed and search for optimum values for their parameters is carried out by minimizing the value of the ISE (integral squared error criterion) between the output step responses of the continuous control system and that of the discretized control system. The minimization procedure adopted makes use of both Rosenbrock's automatic hill climbing technique (which optimizes the direction of search using Gram-Schmidt procedure) and the Fibanacci search method for step size optimization.
The method is illustrated with numerical examples. Results show that step responses obtained using this method of discretization were very close to the desired responses as compared to other methods when one uses as the basis for comparison the integral squared error between step responses.