Enhanced Extreme Learning Machine for Energy Efficient Vampire Attack Detection in Wireless Sensor Networks

Abstract

A sensor node fulfils a specific function inside a wireless sensor network (WSN). WSNs are characterised by a lower tolerance for errors or failures, but they are also more essential to the success of businesses and the well-being of individuals because of the inherent risks involved. Consequently, the battery life of a node would diminish, rendering it nonoperational, which is considered the most severe kind of denial of service assault. Vampire assaults, a kind of denial of service attack, may cause damage to a network, resulting in increased difficulty in detection and unnecessary energy consumption. This research proposes a new method for detecting and preventing vampire attacks by predicting energy consumption in the data path. The method uses the Extreme Learning Machine with Sleep Scheduling Algorithm (ELM_SSA), which has a fast learning speed and is well-suited for resource-limited environments such as WSNs. The sleep scheduling algorithm determines when the nodes should be active and when they can enter sleep mode to conserve energy. Nodes may be scheduled to wake up periodically to perform energy consumption measurements and collect data for anomaly detection.