Equalizer adaptation algorithms for high-speed wireless communications

Adaptation algorithms for use in a decision feedback equalizer operating in an indoor wireless (multipath fading) environment are considered. The advantages and disadvantages of the LMS, RLS, fast Kalman, and lattice RLS algorithms when used in the above specified environment are discussed. Results are presented to compare the performance of these algorithms in terms of their convergence speed, achievable SNR, stability, and computational complexity. Results are also presented on the statistics of the number of iterations required to achieve a particular SNR and the achievable SNR when the training period is fixed. Finally, the effects of varying some of the equalizer design parameters are studied.