Tracking performance of momentum LMS algorithm for a chirped sinusoidal signal

In this paper we study the tracking performance of the momentum LMS (MLMS) algorithm in adaptive prediction for a time-varying chirped sinusoidal signal. The momentum term of the algorithm not only helps to speed up the convergence rate, but also improves the tracking capability for a nonstationary signal. We compare the simulation results of the MLMS with the conventional LMS to highlight the tracking performance. The simulation results show that the MLMS algorithm with an additional momentum term has a better tracking capability in an 8-tap adaptive predictor under noise-free conditions, especially when the filter tracks a fast time-variant signal. However, the MLMS does not have significant improvement when tracking a noise corrupted chirp signal. The normalised MLMS (NMLMS) algorithm has similar simulation results as the ordinary MLMS algorithm for the tracking performance.