Abstract—System Identification is an important area of research in signal processing to design an unknown system. The identification task covers almost all the areas of engineering application such as problem of building models of systems. When insignificant prior information is available and system’s properties are known up to a few parameters, identification is most useful. This paper approaches the system identification problem using WLMS Algorithm (Wilcoxon based LMS) in presence of outliers. Also the result is compared with the conventional LMS. In addition to it, the error is analyzed for the deviation factor at the time of analysis. The result shows an excellent performance with minimum training samples.

Index Terms— Wilcoxon norm; Least Mean Square (LMS); WLMS algorithm; system identification.

Cite: Sidhartha Dash and Mihir Narayan Mohanty, "A Comparative Analysis for WLMS Algorithm in System Identification," Lecture Notes on Information Theory, Vol.1, No.3, pp. 95-98, Sept. 2013. doi: 10.12720/lnit.1.3.95-98