Pilot-Aided Adaptive Prediction Channel Estimation in a Frequency-Nonselective Fading Channel

Shinsuke TAKAOKA  Fumiyuki ADACHI  

IEICE TRANSACTIONS on Communications   Vol.E85-B   No.8   pp.1552-1560
Publication Date: 2002/08/01
Online ISSN: 
Print ISSN: 0916-8516
Type of Manuscript: PAPER
Category: Terrestrial Radio Communications
fading channel,  channel estimation,  pilot symbol,  adaptive prediction,  normalized LMS,  

Full Text: PDF(1.1MB)>>
Buy this Article

Pilot-aided adaptive prediction channel estimation is proposed for coherent detection in a frequency-nonselective fading channel. It is an extension of the conventional weighted multi-slot averaging (WMSA) channel estimation and consists of 3 steps. A block of Np pilot symbols is periodically transmitted, each pilot block being followed by Nd data symbols to form a data slot. In the first step, the instantaneous channel gain is estimated by coherent addition of Np pilot symbols. Using the K past and K future estimated instantaneous channel gains, the second step predicts the instantaneous channel gains at the end and beginning of data slot of interest by a forward predictor and a backward predictor, respectively. The tap-weights of forward prediction and backward prediction are adaptively updated using the normalized least mean square (NLMS) algorithm. Finally, in the third step, the instantaneous channel gain at each data symbol position within the data slot of interest is estimated by simple averaging or linear interpolation using the two adaptively predicted instantaneous channel gains. The computer simulation confirms that the proposed adaptive prediction channel estimation achieves better bit error rate (BER) performance than the conventional WMSA channel estimation in a fast fading channel and/or in the presence of frequency offset between a transmitter and a receiver.