|Correlation Analysis-based Stochastic Gradient and Least Squares Identification Methods for Errors-in-variables Systems Using the Multi-innovation
Shujun Fan, Ling Xu, Feng Ding*, Ahmed Alsaedi, and Tasawar Hayat
International Journal of Control, Automation, and Systems, vol. 19, no. 1, pp.289-300, 2021
Abstract : This paper deals with the identification problem of discrete-time linear time-invariant errors-in-variables systems for the case of the colored output noise. Based on the correlation analysis, the multi-innovation theory is introduced to the errors-in-variables systems where both input and output data are noisy. A correlation analysisbased multi-innovation stochastic gradient algorithm and a correlation analysis-based multi-innovation least squares algorithm are proposed by means of the multi-innovation theory in order to improve the parameter accuracy. The simulation results confirm that these two algorithms are effective.
Correlation analysis, EIV system, gradient method, least squares, multi-innovation identification, parameter estimation