Dimension Reduction and Expansion: Distributed Source Coding in a Noisy Environment

We studied the problem of distributed coding and transmission of inter-correlated sources with memory. Different from the conventional distributed source coding structure which relies on design of effective channel codes to model the inter-correlation and quantizer, the proposed system utilizes distributed compressed sensing [1] for signal dimension reduction through linear matrix operations and dimension expansion for protection against channel noise through a hybrid scalar quantizer linear coder [2]. The proposed system is optimized for minimum end-to-end distortion under a transmission energy constraint. Its performance is verified through simulation and can serve as a good starting point for designing similar analogue based dimension reduction- expansion schemes for applications in sensor networks.