Distributed Source Coding using non-binary LDPC codes for sensor network applications

Recent revolution in sensor networks has resulted in the development of many enabling techniques, one of which being distributed source coding (DSC). DSC allows a low-complexity encoder, shifting the high complexity to the decoder, making this scheme appropriate for sensor networks. In this paper, we have extended the non-binary low-density parity-check (LDPC) codes, developed for channel coding, to be used for distributed source coding of correlated non-binary sources. The system consists of an equiprobable memoryless non-binary source with side information at the decoder. The correlation between the sources is modelled as a virtual channel. The results obtained through simulations demonstrate that for rates 1/2 and 3/4, the non-binary compression scheme performs better the equivalent binary compression scheme. This scheme makes it extremely suitable for sensor networks.