Turbo-Like Codes for Transmission of Correlated Sources over Noisy Channels

In this article we focus on distributed coding of correlated sources that are transmitted through either separated noisy channels (one per source) or over a multiple access channel (MAC). For simplicity, we consider only two sources, but the proposed approach can be easily extended to the case of more sources. The two sources are encoded independently of each other (i.e., for a given source neither the realization from the other source nor the correlation model are available at the encoder site) and transmitted through the channel. Then, the correlation between the sources is exploited at a common receiver, aiming at the reconstruction of the two sources. Although this problem has many practical applications e.g., in the case of sensor networks (where correlated data has to be communicated between the different nodes) and video/image applications, we will focus in the key ideas without discussing specific applications. For simplicity, we will assume simple correlation models and binary sources.