Estimation of servo-system parameters using maximum likelihood method

Abstract

The problem of estimation of servo system parameters that cannot be measured is solved. Unknown parameters are estimated using system response measurements. A special discrete, recursive and iterative form of maximum likelihood method is presented. The method is applied to measured time series of system input and output. It gives minimum variance unbiased estimates in the presence of noise. The use and efficiency of the given method are illustrated through two practical examples.

Keywords

• servo system
• system identification
• parameter estimation
• maximum likelihood