Obtaining diversity gain by block Markov superposition transmission

In time-varying wireless channels, the radio signals suffer from time-varying fading which may cause severe signal attenuation. In this paper, we present a new approach, called block Markov superposition transmission (BMST) over real field, to obtain diversity which increases the transmission reliability under the fading channels. The proposed scheme is similar to BMST over binary field, which has been used to construct big convolutional codes from short codes. However, unlike superposition in BMST over binary field, superposition in the proposed scheme is over real field. We also present an iterative sliding-window detection/decoding algorithm for the proposed transmission scheme. The performance of BMST over real field can be lower bounded by that of the equivalent system. Progressive edge-growth (PEG) constructed LDPC code of rate 0.5 with length 504 and bipolar basic mapping are chosen for simulation, and numerical results show that: 1) in additive white gaussian noise (AWGN) channels, no performance gain is obtained; 2) in fast Rayleigh fading channels, an extra diversity gain of 1.3 dB can be obtained by the proposed approach at bit-error-rate (BER) of 10^-5; 3) in block Rayleigh fading channels, about 22 dB performance gain are obtained by the proposed approach at BER of 10^-5.