Malicious Attack Classification Based-on Continuous Hidden Markov Models in Wireless Sensor Networks

Abstract

Alongside the rapid progress of Wireless sensor networks (WSNs) technologies, sensors and networks can rapidly be victim of distributed attacks. Attackers can perform intrusions to breakdown the network during the routing process, intercept gathered data by dropping or re-sharing them. To avoid the increasing of security issues, many attack identification models were proposed in WSNs in which detection systems are deployed to collect sensed data and categorize them using machine learning and stochastic binary-classification techniques. In this work, a new method is introduced to analyze and classify WSN dataset. We aim to design an anomaly identification approach to improve the sensor network security, it efficiency with high accuracy. To reach this goal, machine learning approaches are used to define a detection system which learn from routing dataset to identify network malicious entries. The proposed models is based on Hidden Markov Model (HMM), Gaussian Mixture Model (GMM) stochastic assumptions. Also, dimensionality reduction technique was deployed to select the most relevant features for training. The experimentation phase was realized on own made dataset reflecting different network situations of normal and attacked cases. The outcomes performances of the proposed method were obtained with classification accuracy of 92.18% using 2HMM/3GMMclassifier. This result demonstrate the quality of our proposed approach compared with existing literature and its usefulness to improve the security of WSN.