Investigations on machine learning techniques for black hole sink hole and clone attack detection and isolation in health care wireless sensor networks

Abstract

newline Wireless Sensor Network (WSN) comprises of spatially distributed independent sensors to monitor the network and collect information about the surrounding environment. The biomedical sensor efficiently gathers the data securely in order to prevent the probable malicious user behavior (i.e. attack) in the network, but it is still being a difficult problem to identify the attack from normal sensor nodes. Therefore, the proposed works focus on developing means of efficient attack discovery for secured data packet broadcasting from source to the destination node in WSN.Position Verification Method with Message Verification and Passing (PVM-MVP) algorithm identified confidence nodes in WSN for secured communication. Though, PVM-MVP algorithm reduces the time consumption for message verification but does not prove to be cost effective. An efficient defense technique against Black Hole (BH) attack was implemented for secured data discovery with minimal routing overhead and delay in WSN. But, it cannot generate the solution for other attacks rather than BH attack. The first work is designed with the help of proposed Projected Independent Component Analysis (PICA) technique to find black hole attack during transmission. Initially, collecting the patient s information from biomedical sensors for healthcare applications with enhanced security is done and mishandling by malicious users is ignored in WSN. Besides, Independent Component Analysis (ICA) efficiently ensures security during transmission. Mutual Information (MI) is employed in ICA technique to identify the black hole node with lesser time according to the behavior analysis. newline newline