Toward a Robust Sparse Data Representation for Wireless Sensor Networks

TL;DR

This paper proposes a novel method to transform sensor data into sparse representations using an unsupervised neural network, improving efficiency for wireless sensor networks.

Contribution

It introduces a new approach combining sparsity extraction, a guarantee scheme, and a customized learning algorithm for better sparse data representation.

Findings

Outperforms conventional sparsity methods on real data

Effectively extracts intrinsic sparse coding from sensor data

Provides a guarantee scheme for sparsity ratio

Abstract

Compressive sensing has been successfully used for optimized operations in wireless sensor networks. However, raw data collected by sensors may be neither originally sparse nor easily transformed into a sparse data representation. This paper addresses the problem of transforming source data collected by sensor nodes into a sparse representation with a few nonzero elements. Our contributions that address three major issues include: 1) an effective method that extracts population sparsity of the data, 2) a sparsity ratio guarantee scheme, and 3) a customized learning algorithm of the sparsifying dictionary. We introduce an unsupervised neural network to extract an intrinsic sparse coding of the data. The sparse codes are generated at the activation of the hidden layer using a sparsity nomination constraint and a shrinking mechanism. Our analysis using real data samples shows that the proposed method outperforms conventional sparsity-inducing methods.