IMDEA Networks Institute Publications Repository

RAZOR: A Compression and Classification Solution for the Internet of Things

Danieletto, Matteo and Bui, Nicola and Zorzi, Michele (2013) RAZOR: A Compression and Classification Solution for the Internet of Things. [Journal Articles]

[img]
Preview
PDF (RAZOR: A Compression and Classification Solution for the Internet of Things) - Published Version
Download (440Kb) | Preview

Abstract

The Internet of Things is expected to increase the amount of data produced and exchanged in the network, due to the huge number of smart objects that will interact with one another. The related information management and transmission costs are increasing and becoming an almost unbearable burden, due to the unprecedented number of data sources and the intrinsic vastness and variety of the datasets. In this paper, we propose RAZOR, a novel lightweight algorithm for data compression and classification, which is expected to alleviate both aspects by leveraging the advantages offered by data mining methods for optimizing communications and by enhancing information transmission to simplify data classification. In particular, RAZOR leverages the concept of motifs, recurrent features used for signal categorization, in order to compress data streams: in such a way, it is possible to achieve compression levels of up to an order of magnitude, while maintaining the signal distortion within acceptable bounds and allowing for simple lightweight distributed classification. In addition, RAZOR is designed to keep the computational complexity low, in order to allow its implementation in the most constrained devices. The paper provides results about the algorithm configuration and a performance comparison against state-of-the-art signal processing techniques.

Item Type: Journal Articles
Uncontrolled Keywords: Signal processing; motif; compression; classification; computational complexity; Internet of Things.
Subjects: UNSPECIFIED
Divisions: UNSPECIFIED
Depositing User: Nicola Bui
Date Deposited: 26 Mar 2014 15:12
Last Modified: 24 Apr 2014 08:23
URI: http://eprints.networks.imdea.org/id/eprint/737

Actions (login required)

View Item View Item