Optimal Relay Node Selection Using Machine Learning to Extend Coverage in Disaster Area Network

In times of both natural and human-made disasters, the establishment of an emergency wireless network plays a vital role in saving human lives. Such a network becomes crucial for facilitating communication and coordination among rescue teams and survivors. Device-to-Device (D2D) communication can play an important role in enabling wireless communication networks in the event of a vulnerable network, high traffic, or network congestion. Recently, there has been significant attention directed toward improving the coverage and reliability of D2D communication by implementing relay node selection techniques. In this paper, we consider a Base Station (BS) located in the Functional Area (FA) that allows communication to User Equipment (UE) located in the disaster-affected region through the Relay Node (RN). We propose a relay node selection method using machine learning, namely an optimal relay node selection based on the nearest neighbor voting method. This work evaluates the optimal relay node's performance to improve the coverage of disaster area network, minimize the outage probability, improve energy efficiency in post-disaster scenarios, and investigate the network capacity.