Analysis of SISO front-end decoders in IM/DD optical channels with application to turbo code decoding

We analyze the log-likelihood ratio (LLR) of the received signal yielded by a soft input/soft output (SISO) front-end decoder, in intensity modulation direct detection (IM/DD) optical channels operating in the presence of amplified spontaneous emission (ASE) noise. We focus on long haul or metro links spanning several hundred kilometers of single mode fibers with optical amplifiers. Unlike in AWGN channels where the LLR is a scaled version of the received sample, our study shows that in IM/DD optical channels with dominant ASE noise the LLR is approximately proportional to the square root of the received signal. Moreover, although Gaussian modeling of the received samples in optical channels has been shown to be inaccurate (e.g., (G. Bosco, et al., 2003), (Y. Cai, et al., 2003)), in this work we show that the LLR yielded by a SISO front-end decoder can be accurately modeled as a Gaussian random variable. Building on this result, we develop an adaptive decoding scheme for block turbo codes (BTC) in IM/DD optical channels. The predictions of our analysis are verified by computer simulations of the bit error rate (BER) for several BTC. Our results also show that the new scheme performs better than previous proposals in transmissions over optical fiber channels