A Machine Learning Approach for the Underwater Investigation of Malicious Activity in a Wireless Sensor Network

Underwater Wireless Sensor Networks (UWSN) have drawn increased interest from researchers in recent years due to advancements in underwater tracking, the implementation of multiple uses, and marine monitoring. Both researchers and business are drawn to the UWSN field. It is also susceptible to hostile attacks and threats because of the harsh undersea surroundings, its own capabilities, and the open acoustic channel. These features make it simple for attackers to intercept information across the source and destination. Many studies have been done to secure UWSN from malicious attacks, but the majority of them rely on a threshold that is determined before deployment or a training set of data. Because there is no proven security foundation, it is difficult and challenging for UWSN to identify harmful attacks. To identify harmful attacks in a UWSN, we support vector machines in this research. SVM produces good results and requires significantly less time to train than neural networks. This study will help the researchers design the routing protocols to cover the known security threats and help industries manufacture the devices to observe these threats and security issues.