Capacity approaching coded modulation in optical communications

Forward error correction (FEC) coding, modulation formats like differential phase shift keying (DPSK) and adaptive electronic equalization to mitigate, e.g. polarization mode dispersion (PMD) are some of the methods applied in current direct-detection optical communication systems, which are limited to a spectral efficiency of 1 bit/s/Hz. Coherent single-carrier QPSK systems with polarization multiplexing (PolMUX) allows to attain an spectral efficiency of up to 4 bit/s/Hz. However, to guarantee a bit error rate (BER) <; 10^-15 adaptive multiple-input multiple-output (MIMO) equalization as well as outer FEC coding has to be included which results in less than 4 bit/s/Hz in commercial 40 Gbit/s systems. From wireline and wireless systems it is well known that larger symbol alphabets in combination with FEC coding (denoted as coded modulation) and orthogonal frequency division multiplexing (OFDM) can significantly enhance the spectral efficiency. However, in order to exploit the channel capacity of the fiber the modulation itself needs to be capacity achieving. In this contribution we apply capacity achieving constellations in an optical OFDM system and jointly design mapping and low-density parity-check (LDPC) codes for an iterative decoding scheme, namely bit-interleaved coded modulation with iterative decoding (BICM-ID). The authors expect this coding scheme to be most appropriate for future optical transponders to utilize long haul communications close to channel capacity.