Intrusion Detection in Wireless Sensor Networks using Optics Algorithm
N Dharini, J Sowndharya, P Sudha · 2022 International Conference on Applied Artificial Intelligence and Computing (ICAAIC) · 2022
Security is a major issue in Wireless Sensor Networks (WSNs). Wireless sensor networks are commonly utilized for real-time monitoring in a variety of real-world applications. Meanwhile, WSNs are vulnerable to both insider and outsider attacks; and detecting and protecting against insider attacks are very difficult. A Gray hole attack, is one of the insider attack is also known as Selective Forwarding attack (SFA) is more prevalent in WSN. Here the attackers choose some received data packets to drop and frighten the clustered WSNs. To overcome this problem, this paper proposes an intrusion detection mechanism using OPTICS (Ordering Points to Identify the Cluster Structure) which is kind of density - based clustering algorithm to defend against selective forwarding attack. The intruder node is identified by using two distinct features such as Energy Consumption and Received packet count. To identify outliers, OPTICS Algorithm uses its parameters core distance and reachability distance. Using NS2 (Network Simulator Version 2) – Mannasim, simulations are carried out. The energy consumption and received packet count data analysis are carried out using WEKA (Waikato Environment for Knowledge Analysis) simulation tool. The detection rate and elapsed time are better with the proposed OPTICS based intrusion detection algorithm when compared with its peers. Also, the performance analysis of the network was enhanced in terms of PDR (Packet Delivery Ratio) and throughput after the deletion of identified malicious nodes. Simulation outcomes prove that our proposed work give the mean accuracy of 100% detection rate and 0% of missed detection rate.