A New Robust Frequency-Domain Adaptive Filter with Colored Signal Input

A frequency-domain adaptive filter (FDAF) with parallel structure and reduced computational complexity for precise adaptive system identification has been proposed. However, the presence of additive noise degrades the performance of the FDAF, and the degradation is conspicuous in case of colored signal input. In this work, we propose a new FDAF which improves the performance of the conventional FDAF for system identification with the presence of additive noise and colored signal input.