Steady-state performance analyses for sliding window max-correlation matching adaptive algorithms

Abstract

This paper presents the sliding exponential window max-correlation matching (SEWMCM) adaptive algorithm and the sliding rectangular window max-correlation matching (SRWMCM) adaptive algorithm for finding the maximum correlation of two different signal vectors. A unified approach to the steady-state excess mean square error (MSE) performance analyses for proposed algorithms is developed, including several general close-form analytical expressions based on the non-stationary system identification model. It is conclusively shown by numerical simulations that the SEWMCM algorithm converges faster than the SRWMCM algorithm, whereas the estimation accuracy and the steady-state performance of the SRWMCM outperform those of the SEWMCM and the conventional exponentially-weighted RLS (EWRLS).