Accelerating the convergence of the widely linear LMS algorithm for channel equalization

In this paper, we investigate the use of some classical techniques for accelerating the convergence of the LMS algorithm in widely linear (WL) adaptive processing. The application of channel equalizing is considered. Simulation results permit to verify that the normalized LMS strategy presents the best performance as far as a reduction of the trade-off between the convergence rate and steady-state error is concerned.