BCJR vs SOVA for a practical coherent turbo coded OFDM system

Vineel Kumar Veludandi
DOI: https://doi.org/10.1109/icccnt45670.2019.8944894
2019-07-01
Abstract:In this paper, we compare the performances of Bahl Cocke Jelinek Raviv (BCJR) and soft output Viterbi algorithm (SOVA)-based decoders for a practical coherent turbo coded orthogonal frequency division multiplexed (OFDM) system. The rate-1/2 turbo coded quadrature phase-shift keyed (QPSK) signals are transmitted through a time variant distorting channel. The magnitude of the channel taps is Rayleigh distributed. The output of the channel is rotated by an unknown phase offset and further distorted by carrier frequency offset and additive white Gaussian noise (AWGN). Synchronization and channel estimation are carried out using the algorithm proposed in [1]. Since noise variance estimate is required for the BCJR algorithm for its branch metric calculation, the BCJR-based turbo decoder's performance depends on the length of the training sequence. However, SOVA does not require noise-variance estimate. A random data interleaver is used to decorrelate the channel to improve the turbo decoder's performance. Simulation results show that the bit-error-rate (BER) performance of BCJR-based decoder is just 0.2 dB better than SOVA-based decoder, for the practical coherent turbo coded OFDM system.
What problem does this paper attempt to address?