Fast fading channel estimator using DCT and simplified Kalman filter

J.F. SCHMIDT; J.E. COUSSEAU; R.WICHMAN; F.H. GREGORIO

Helsinki

Workshop; IEEE Signal Processing Advances in Wireless Communications Conference; 2007

IEEE

We present a novel channel estimator for fading channelsusing basis function approximations and Kalman filtering.The proposal results in a low complexity estimatorthat is robust to the shape of the channel Doppler spectrumand allows different Doppler spectra for each channelcoefficient in a frequency-selective scenario. The estimatoris based on the description of channel variationsvia truncated discrete cosine transform (DCT). We showthat DCT can be approximated by a set of linear narrowbandpassband filters that can be inserted into a Kalmanfilter formulation that track a set of decoupled parameters.This estimator attains close to perfect channel knowledgeperformance with only 1% training symbols in a dataframe with Doppler frequencies up to 0.8 times the framefrequency.