Confluence of pattern recognition and signal processing: application of Al-Alaoui pattern recognition algorithm to digital filters design

A weighted mean square error (WMSE) approach to optimising digital filters is delineated. It is applied in the current work to optimising the classical Al-Alaoui IIR differentiators to obtain new improved wideband differentiators of varying orders. These can be directly used for analog to digital conversion and in many digital signal processing applications. The weighted MSE approach is motivated by the Al-Alaoui WMSE approach to pattern recognition. In addition, the differentiators are converted to integrators of similar orders that are further optimised. Various examples and comparisons are presented to demonstrate the viability of the proposed approach.