Optimal Rate Allocation in Mismatched Multiterminal Source Coding

TL;DR

This paper studies how to optimally allocate rates among remote sensors in a multiterminal source coding setup to minimize overall distortion, considering mismatched local and central distortion measures.

Contribution

It introduces an optimal rate allocation framework for mismatched multiterminal source coding with practical Gaussian and binary models.

Findings

More remote encoders can reduce distortion under certain conditions.

Fewer high-quality estimates outperform many low-quality ones.

Optimal rate allocation depends on source and noise characteristics.

Abstract

We consider a multiterminal source coding problem in which a source is estimated at a central processing unit from lossy-compressed remote observations. Each lossy-encoded observation is produced by a remote sensor which obtains a noisy version of the source and compresses this observation minimizing a local distortion measure which depends only on the marginal distribution of its observation. The central node, on the other hand, has knowledge of the joint distribution of the source and all the observations and produces the source estimate which minimizes a different distortion measure between the source and its reconstruction. In this correspondence, we investigate the problem of optimally choosing the rate of each lossy-compressed remote estimate so as to minimize the distortion at the central processing unit subject to a bound on the overall communication rate between the remote sensors and the central unit. We focus, in particular, on two models of practical relevance: the case of a Gaussian source observed in additive Gaussian noise and reconstructed under quadratic distortion, and the case of a binary source observed in bit-flipping noise and reconstructed under Hamming distortion. In both scenarios we show that there exist regimes under which having more remote encoders does reduce the source distortion: in other words, having fewer, high-quality remote estimates provides a smaller distortion than having more, lower-quality estimates.