System Identification using Mixed Total Least Squares by Rotating the System of Coordinates

D. Pachner, J. Pekař, M. Cepák, and J. Štecha (Czech Republic)

Keywords

Identification, ARX model, Least Squares, Total Least Squares, Rotating system of coordinates

Abstract

: Real time system parameter estimation from the set of input-output data is usually solved by the quadratic norm minimization of system equations errors known as least squares (LS). But measurement errors are also in the data matrix and so it is necessary to use modification known as total least squares (TLS) or mixed LS and TLS. TLS or mixed LS and TLS problem is usually solved by the SVD decomposition of system matrix. So it is difficult to update the solution when the new data are obtained. In the article the new original method using iterative solution of the least squares based on the rotating system of coordinates [4] is described. The idea of the method is based on the fact that in some special cases the LS and TLS solutions are identical.

Important Links:



Go Back