Adaptive system identification using the normalized least mean fourth algorithm

In this work we propose a novel scheme for adaptive system identification. This scheme is based on a normalized version of the least-mean fourth (LMF) algorithm. In contrast to the LMF algorithm, this new normalized version of the LMF algorithm is found to be independent of the input sequence autocorrelation matrix. It is also found that it converges faster than the normalized least mean square (NLMS) algorithm for the lowest steady-state error reached by the NLMS algorithm. Simulation results confirm the superior performance of the new algorithm.