Publication Details
Overview
 
 
Evangelos Zimos, João Mota, Evangelia Tsiligianni, Miguel Rodrigues, Nikos Deligiannis
 

Chapter in Book/ Report/ Conference proceeding

Abstract 

Large-scale wireless sensor networks (WSNs) andInternet-of-Things (IoT) applications involve diverse sensingdevices collecting and transmitting massive amounts of heterogeneousdata. In this paper, we propose a novel compressivedata aggregation and recovery mechanism that reduces the globalcommunication cost without introducing computational overheadat the network nodes. Following the principles of compressivedemixing, each node of the network collects measurement readingsfrom multiple sources and mixes them with readings fromother nodes into a single low-dimensional measurement vector,which is then relayed to other nodes; the constituent signalsare recovered at the sink using convex optimization. Our designachieves significant reduction in the overall network data ratescompared to prior schemes based on (distributed) compressedsensing or compressed sensing with (multiple) side information.Experiments using real large-scale air-quality data demonstratethe superior performance of the proposed framework againststate-of-the-art solutions, with and without the presence ofmeasurement and transmission noise.

Reference 
 
 
DOI  scopus