Low-Complexity Coding and Source-Optimized Clustering for Large-Scale Sensor Networks

TL;DR

This paper introduces a scalable distributed source coding method for large sensor networks, combining low-complexity quantization, hierarchical clustering, and efficient decoding to handle hundreds of sensors effectively.

Contribution

It proposes a novel scalable approach integrating distortion-optimized quantization, source-optimized clustering, and sum-product decoding for large-scale sensor networks.

Findings

Effective handling of hundreds of sensors

Reduced complexity compared to Turbo and LDPC codes

Maintains rate-distortion performance

Abstract

We consider the distributed source coding problem in which correlated data picked up by scattered sensors has to be encoded separately and transmitted to a common receiver, subject to a rate-distortion constraint. Although near-tooptimal solutions based on Turbo and LDPC codes exist for this problem, in most cases the proposed techniques do not scale to networks of hundreds of sensors. We present a scalable solution based on the following key elements: (a) distortion-optimized index assignments for low-complexity distributed quantization, (b) source-optimized hierarchical clustering based on the Kullback-Leibler distance and (c) sum-product decoding on specific factor graphs exploiting the correlation of the data.