INTELLIGENT CLONE DETECTION AND CLASSIFICATION USING CAT SWARM OPTIMIZATION WITH DEEP LEARNING MODEL FOR WIRELESS SENSOR NETWORKS

Abstract

The progress of wireless sensor networks (WSN) technology wasobtaininggreatly enhance count of attention from researchers in recent decades. Its huge count of sensor nodes (SNs) is most feature which generatesit useful to technology. The sensors areconnectedtogether to network model. These SNs can be usually exploited for various applications like target tracking, health monitoring, pressure monitoring, fire recognition, and so on. But, the disadvantage is that WSNs canfrequentlyutilize in hostile, critical environments but it could not control physical access. The adversary can capture the legitimate SNs, take them out and thengather any sensitive data like node ID, keys and accomplish a replication attack. This study presents an Intelligent Clone Detection and classification using Cat Swarm Optimization with Deep Learning (ICDC-CSODL)technique for WSN. The main goal of the ICDC-CSODL system lies in the accurate identification and classification of clone nodes in the network. To accomplish this, the presented ICDC-CSODL technique follows a two-stage process. Initially, the ICDC-CSODL system utilizes attention-basedbi-directional long short-term memory (ABiLSTM) approach for clone node detection. Next, in the latter stage, the CSO system is used to adjust the hyperparameter values of the ABiLSTM approach. The simulation results of the ICDC-CSODL technique are tested on a series of experiments. A widespread simulation results analysis illustrated the improvement of the ICDC-CSODL technique in terms of different measures.