Estimation of time and frequency selective channels in OFDM systems: a Kalman filter structure

A Kalman-filter based channel estimator of OFDM systems in a time-frequency-selective fading environment is studied. Based on an AR model of the dynamic channel, a vector Kalman filter of dimension equal to the number of subcarriers can be used to track the channel. It is proposed that a scalar Kalman filter be employed for each subcarrier to estimate the channel gain. The channel estimate is further refined with an MMSE combiner. This two-step structure, with the per-subchannel Kalman estimator explores the time-domain correlation of each subchannel, and with the MMSE combiner explores the frequency-domain correlations between the subchannels, offering a performance comparable to the much more complicated vector Kalman estimator.