Compressive Sensing Based on Homomorphic Encryption and Attack Classification using Machine Learning Algorithm in WSN Security

Abstract

Data protection is essential for sensitive applications using Wireless Sensor Networks like health monitoring or video surveillance. WSN are deployed generally in harsh environment making them vulnerable for attacks thus it's important to secure data while being transferred from the sensor until the base station. This article is a proposal for a methodology which enable attack detection and classification on WSN. Compressive sensing is used to optimize the size of data exchange and hence optimize energy consumption. Homomorphic encryption allows to reduce encryption complexity by applying arithmetic operations on cypher text. Machine learning is applied to classify the attacks quickly.