PARAMETER ESTIMATION OF CONTINUOUS-TIME LINEAR SYSTEMS USING GENETIC ALGORITHMS

摘要

A method for estimating a continuous-time linear system using a genetic algorithm is provided to effectively apply to selection of a continuous time model, which is necessary to design a controller for the continuous-time linear system, by solving economical and technical problems caused from signal measurement. The technique transforms the continuous time model into a discrete time model(30) and transforms the parameter of the continuous time model into the parameter of the discrete time model by using a parameter relational expression(40). The technique makes the genetic algorithm(50) optimally adjust the parameter of the continuous time model to make an operation property of the system(10) similar to the operation property of the discrete time model. The parameter of low pass filters(20), which remove noise generated when an I/O(Input/Output) signal of the system is measured, is estimated simultaneously with the parameter of the model by using the genetic algorithm.