Bit error probability for an adaptive diversity receiver in a Rayleigh fading channel

The bit error probability is determined for an adaptive diversity receiver using decision directed feedforward channel estimation. The channel is assumed to be a nonselective Rayleigh fading channel with AWGN. The analysis is valid for all feedforward estimation structures which use an FIR-type estimation filter. The analysis can be applied also to Rake receivers in a frequency selective Rayleigh fading channel. Numerical results are presented for an adaptive diversity receiver which uses either an FIR-type minimum mean-square error (MMSE) filter or a moving average (MA) filter as the estimation filter. The results show that the MA filter performs poorly compared to the MMSE filter at large signal-to-noise ratios and large Doppler spreads. However, at small signal-to-noise ratios and Doppler spreads the filters do not show any remarkable difference in the performance.<>