Denial-of-Service Attack Detection in Wireless Sensor Networks: A Simple Artificial Intelligence Method

When compared to more conventional networks, wireless sensor networks (WSNs) stand out. Classical networks' networking techniques, in contrast to WSN, can only regulate incoming packets according to their Internet Protocol (IP) address translation. The efficient gathering and transmission of data from such a network is an essential need for any application. A Denial of Service (DoS) assault occurs when malicious actors intentionally try to disrupt the data supply process, making the anticipated services unavailable. Because denial-of-service attacks aim to disrupt regular services, they may often severely limit the use of wireless sensor networks. When malicious nodes gain access to a network, they may launch a denial-of-service (DoS) assault. Using a trust-based method, this study aims to trace these nodes that are colluding. One component of the trust-based system is a way to quantify the level of confidence in each node. Nodes that have a low level of trust are considered to be malevolent nodes. All packets sent within the given time window are considered a measure of trust. The node that forwards the most packets while leveraging the fewest assistance will be marked as malicious. By outperforming RF, XGBoost, and KNN classifiers while using less training data, the suggested method has shown strong performance. Its accuracy rate is 99.5%. Our suggested method considerably beats these classifiers in regard to processing time, since it only uses 9.7 percent, 13.1 percent, and 2% of the time that FR, XGBoost, and KNN use, respectively. Notably, RF did yield slightly better accuracy; nevertheless, the suggested method substantially outperformed RF by requiring only 9.7 percent of RF processing time, a crucial component in fulfilling WSN restrictions. We used certain performance indicators to evaluate the results after deploying the emerging architecture in NS3. The total number of packets sent by each node during their allocated time window was used to determine trust. The most spiteful node was determined to be the one that sends out the most packets while using almost minimal resources. The NS3 program is used to explain the method and analyze the outcomes using certain metrics in this task implementation.