Coding performance for signal dependent channels in visible light communication system

Avalanche photo-diodes (APD) are widely employed in visible light communication (VLC) systems. Experiments show that under official illuminance APD shot noise can not be ignored due to the multiplication process. Under VLC dual-purpose illumination, symbols on different constellation points inevitably experience different levels of noise. In such a signal dependent noise channel, traditional log likely-hood ratio (LLR) calculation for decoder assuming signal independent noise will cause performance loss. In this work, we first investigate the general signal dependent Gaussian channel. The optimal LLR calculation for decoder is then derived. We further analyze the extrinsic information transfer chat to find the decoding thresholds. Finally a modified minimum sum algorithm is suggested with reduced complexity. Numerical simulations show that, with a regular (3,6) low-density parity check (LDPC) code of block length 20000, our proposed scheme can achieve 0.7dB gain over the scheme designed for signal independent noise.