Collaborative Edge Mining for predicting heat stress in dairy cattle

Kriti Bhargava, Stepan Ivanov

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Citations (Scopus)

Abstract

Edge Mining (EM), a novel Fog Computing technique, has been proposed to perform data analysis on sensor devices at the edge of Internet of Things (IoT). The approach, however, is limited to analysis conducted by each sensor node in isolation. In this paper, we propose Collaborative Edge Mining (CEM), an extension of the EM technique, wherein multiple sensor devices participate together in on-site data analysis and prediction. Our model detects contextually relevant events by integrating and analysing data arising from different sources and, thereby, lays the foundation of a sensor-based implementation of Apache Storm like framework. We have evaluated our approach with respect to the Linear Spanish Inquisition Protocol for a precision farming application. We illustrate CEM for the estimation of Temperature Humidity Index, an important metric to predict Heat Stress in dairy cattle, and compare its performance to EM. CEM performs well in most cases, especially, latency-sensitive scenarios.

Original languageEnglish
Title of host publication2016 Wireless Days, WD 2016
EditorsThierry Gayraud, Manuel P. Ricardo, Emmanuel Chaput, Samir Medjiah, Andre-Luc Beylot, Pascal Berthou
PublisherIEEE
ISBN (Electronic)9781509024940
DOIs
Publication statusPublished - 27 Apr 2016
EventWireless Days, WD 2016 - Toulouse, France
Duration: 23 Mar 201625 Mar 2016

Publication series

NameIFIP Wireless Days
Volume2016-April
ISSN (Print)2156-9711
ISSN (Electronic)2156-972X

Conference

ConferenceWireless Days, WD 2016
Country/TerritoryFrance
CityToulouse
Period23/03/201625/03/2016

Keywords

  • apache storm
  • edge mining
  • heat stress
  • precision farming
  • wireless sensor networks

Fingerprint

Dive into the research topics of 'Collaborative Edge Mining for predicting heat stress in dairy cattle'. Together they form a unique fingerprint.

Cite this