Effective Event Exposure Classifier (E3C) in Wireless Sensor Network through SVM

Abstract

A Wireless Sensor Network (WSN) is composed of distributed nodes designed for environmental monitoring and event detection. To optimize energy usage during correlated data gathering, the Correlated Data Gathering (CDG) scheme utilizes the Adaptive Routing algorithm. While this approach effectively conserves energy, it introduces a significant energy-delay tradeoff when securing data within the sensor network. In response to this challenge, the E3C-SVM system leverages Support Vector Machine (SVM) classification with minimal energy consumption to assess event significance. E3C-SVM also employs the Doppler Effect method for efficient sensing data event recovery, particularly in identifying periodic events caused by moving objects. By reducing classification time, E3C-SVM mitigates the energy-delay tradeoff. Furthermore, E3C-SVM incorporates a mechanism for generating event-specific keys, reducing energy consumption during key generation, and enhancing security when broadcasting notifications to sensor nodes. This feature elevates the security level of object collection. Experimental evaluations primarily assess classifier performance, security levels, and energy consumption rates.

Index Terms: Correlated Data Gathering, Wireless Sensor Network, Support Vector Machine, Classification time, Adaptive Routing.