DECENTRALIZED ENCODING OF A REMOTE SOURCE

For decentralized estimation of a remote source under communication co nstraints, noisy observations taken by spatially separated sensors are encoded before transmission to a central decision maker. As the obser vations are not simultaneously available at a given site, the peripher al encoders are to operate disjointly. This problem cannot be cast as the classical one of encoding a remote source. Given that the peripher al encoders are scalar quantizers, we consider several criteria for pa rtitioning the space of observations and for reproducing the source wi th minimum distortion. They differ in regard to the knowledge about th e source-observations model, for example about the spatial correlation among the observations, and in regard to the complexity of the decodi ng operations, but all lead to an unifying design approach of cyclical optimization. For a linear Gaussian source-observations model, adopti ng the quadratic distortion measure, we carry out each design and eval uate the performance of the various schemes in terms of achieved disto rtion-rate pairs, allowing for some combined parametric variations of the model. We also derive the optimum theoretically achievable perform ance for the distributed and for the conventional (non-distributed) sc heme in terms of distortion-rate bounds. The results demonstrate that distributed encoding-decoding schemes, if properly designed, compare f avorably to the non-distributed schemes.