Wireless channel tracking based on self-tuning second-order LMS algorithm

The use of the second-order LMS (SOLMS) algorithm for wireless channel tracking is explored. A frequency domain approach is presented to optimise the steady-state mean performance of SOLMS-based wireless channel tracking. It is shown that in comparison to the LMS algorithm, the benefit provided by the SOLMS algorithm is evident in tracking the wireless channel. However, such an optimisation is based on knowledge of the fading channel, which is usually not obtainable in practice. To deal with this problem, a self-tuning implementation of SOLMS-based wireless channel tracking is proposed. Performing online parameter optimisation using the technique of gradient descent, the proposed method can be used to implement high-performance wireless channel tracking without any a priori knowledge. These developments are supported by Monte Carlo simulations.