Performance of OCDMA system with FEC based on interference statistical distribution analysis
Résumé
The optical code division multiple access (OCDMA) technique has been investigated for several years as a promising alternative to provide multiple access in high speed optical networks. However, due to the use of binary on off keying (OOK) modulation, an inherent main performance limitation is multiple access interference (MAI), following a specific distribution. In order to reduce errors induced by MAI, it is necessary to take into account this specific statistic when applying forward error correction (FEC) with soft decision decoding. In this paper, we analyse the MAI distribution in order to determine FEC soft decoding calculation rules. The study is performed for a low density parity check (LDPC) code scheme over an OCDMA system using two-dimensional (2-D) OCDMA spreading codes. The a priori information required for soft decoder initialisation is calculated according to the specific OCDMA channel statistic. The enhancement provided by the decoding algorithm adaptation confirms that it is important to take into account the specific MAI distribution as described in this paper. Finally, we show that, despite the loss of available data rate linked to the addition of redundancy, the data rates per user can significantly be increased thanks to this efficient error correction scheme. Copyright © 2009 John Wiley & Sons, Ltd.