Abstract:
Wireless Sensor Networks (WSNs) play a pivotal role across various domains, such as environmental monitoring and industrial automation. Nevertheless, their decentralized and resource-constrained nature exposes them to security vulnerabilities, notably Denial of Service (DoS) attacks. Detecting and mitigating such threats in WSNs are imperative to uphold operational reliability. This study introduces an innovative methodology employing the Isolation Forest algorithm for DoS attack classification in WSNs. Trust metrics encompassing reliability, contact intimacy, cooperation, energy consumption, and throughput are gathered from sensor nodes to construct datasets. Through the application of the Isolation Forest algorithm on these datasets, anomalies indicative of DoS attacks are discerned. Leveraging the intrinsic characteristics of isolation trees, the algorithm effectively distinguishes between normal network behavior and malicious activities. The efficacy of the proposed approach is demonstrated through a mathematical model, substantiating its ability to detect and mitigate DoS attacks. Experimental findings further validate the effectiveness of our method in accurately identifying DoS attacks with minimal false positives. This method presents a promising avenue for bolstering WSN security and resilience against DoS attacks, ensuring uninterrupted operation and preserving data integrity in critical applications.
