Wireless sensor networks have been widely applied in many areas for real-time event monitoring. It is designed and deployed in the field using both mobile and static sensor nodes. These nodes susceptible to different types of attacks such as Distributed Denial of service attack, Sybil attack and eavesdropping attack. These attacks lead to the limited energy of SNs and causes loss of data packets within a network. The DDoS attack launches a coordinated attack by flooding the target nodes with bogus requests, thus exhausting their resources, and forcing them to deny service to legitimate member nodes. Sybil attacks forge one or multiple identities to decrease the localization accuracy, or sometimes to collapse the whole localization systems. Eavesdropping attacks intercepts the data under transmission. In order to avoid these attacks, many supervised techniques have been employed to mitigate it but nowadays different attack characteristics have been injected to bypass the employed solution. In this paper, a novel unsupervised spatio temporal characteristic (USTC) protocol is provided to each node of the sensors to withstand against the dynamic attacks. The unsupervised learning imposed on the node is used to examine spatio temporal attack characteristics of intermediate node before data transmission based on selected path. It implicitly provides the data traffic balancing and energy consumption by reducing the overhead of the network with reference to cache of data in the routing table which contain monitored information of the node and network through node density calculation. Alternate path is selected on detection of attack in the predicted path. Analytical and simulation results on the proposed model proves that USTC protocol provides better performance terms of packet delivery ratio , Network overhead, transmission delay, computation time and packet loss.
Volume 11 | 05-Special Issue
Pages: 2126-2135