An Adaptive Channel Estimator for Frequency-Selective Fading Channels

A channel estimator for extracting the discrete-time channel coefficients (“tap-gains”) of a time-varying frequency-selective channel is investigated. A Kalman filter is integrated into a least mean squares (LMS) estimator such that a-priori knowledge about the statistics of the channel is taken into account. Assuming a state-space model of the channel is given, parameter settings and the stationary error variance of each single tap-gain are analytically derived for Rayleigh-fading channels. The performance curves of the modified algorithm are shown to be superior to that of the conventional LMS estimator, and are close to the theoretical lower bounds which are also investigated.