SSP'05 -- Proceedings

	IEEE/SP 13th workshop on Statistical Signal Processing July, 17-20, 2005 - Bordeaux - France

Information regarding the paper

Title

A Lower Bound to the AWGN Remote Rate-Distortion Function

Author(s)

Michael Gastpar

University of California, Berkeley

Get the paper in PDF format

To obtain Acrobat Reader (version 5 minimum required) necessary to his read.

Abstract

In the remote source coding problem, an underlying source is observed in noise. The noisy observations must be encoded into a bit stream in such a way as to enable the decoder to produce a good approximation to the original source sequence. The trade-off between the rate of the bit stream and the fidelity of the reconstructed source sequence is sometimes referred to as the remote rate-distortion function. This paper focuses on a special case of the remote source coding problem: The encoder obtains M noisy versions of each underlying source sample. The probability density function of the underlying source is arbitrary, but the observation noise is assumed to be Gaussian (hence the name ``AWGN remote rate-distortion function''). The goal is to reconstruct the underlying source sequence to within mean-squared error. For this scenario, a new lower bound to the rate-distortion function is presented.

The investigations are motivated by a study of the fundamental performance trade-offs in certain sensor network scenarios. The presented lower bound on the remote rate-distortion function is one of the building blocks for a cut-set argument that leads to an upper bound to the performance achievable in these sensor networks.