Optimal Scheme for the Detection and Classification of Clone Node Attack in WSN Using TAIGBRFCNIA

Abstract

Wireless Sensor Network is an isolated set of low cost hardware sensor nodes that supply nodes, memory and processing resources to limit battery life for environmental (or physical) monitoring conditions. WSN is used in unsafe and unforeseen circumstances and thus prone to multiple forms of attack. A replication or clone attack, in which the attacker can instantly grab a network node and get data from the node confiscated, represents a physical threat. Then reprogram to create a node replica. These copies will then be distributed through all networking areas and named true networkers as a replicated node cannot be found. Where centralized clone attack detection methods can be used, the WSN can be either static. In the proposed work, the effective technique for the detection and classification of clone node is presented. Initially, the input data is preprocessed and the preprocessed data is normalized for removing unwanted data. From this, the best features are selected for the classification process. So as to attain the optimal data the preprocessed data is optimized with the use of Modified Particle Swarm Optimization technique (MPSO). The data is then clustered by means of K-means clustering process. The training is done by means of this MPSO and the classifier Modified Artificial Neural Network (MANN). The optimized and trained features are then classified with the utilization of MANN technique to detect and classify the clone attack as normal or malicious. Additionally, the Clone node identification can be done using the Trust Aware Intense Grade Boosting Random Forest Clone Node Identification Algorithm. Finally, the performance analysis is performed and the proposed and the existing techniques are analyzed to demonstrate the efficiency of the scheme.