Performance of turbo codes with two-branch receive diversity and correlated fading

We analyze the effects of fading correlation on the performance of turbo codes with two-branch receive diversity combining schemes. We use transfer function bounding techniques to obtain upper bounds on the bit error rate (BER) for maximum-likelihood (ML) decoding of turbo codes constructed with uniform interleavers. We apply the techniques to parallel concatenated coding (PCC) schemes with two-branch receive diversity. We present the average bounds on the BER for two cases: (a) fully interleaved channels with channel state information (CSI) and (b) Rayleigh fading channels with fading perfectly correlated in time. We show that as the fades in the two antennas become more correlated, the BER improvement obtained by increasing the block size N becomes larger. We also show that the effects of fading correlation for maximal ratio combining (MRC) and for equal gain combining (EGC) are almost identical, while those for selection combining (SC) are much more pronounced than for these two techniques.