| Abstract: | Device-to-device (D2D) communication has captured the researchers attention working in data-intensive applications. It has various benefits, such as low communication latency, load balancing, high spectral efficiency, and many more. However, despite these benefits, it has significant issues like efficient resource allocation, device discovery, and interference mitigation. Various solutions have given by the researchers to tackle these issues and the research community accepts them well. Here, we are targeting the issues associated with the device discovery, i.e., the base station assisted discovery. The initial step for D2D communication is the device discovery that the base station can perform. But, if the channel indicator parameters of the base station are not good, then the device discovery and further data sharing will be affected. Thus, there is a need for the best base station selection that improves the efficiency of the overall network. Many network selection solutions (for cellular networks) are available in the literature, but none of it talked about in the D2D communication scenario. So, motivated by this, this paper proposes an AI-based intelligent and efficient network selection scheme for D2D users to improve the device discovery experience and overall system’s sum rate. We then evaluate the performance of the proposed scheme using various evaluation metrics, such as accuracy, precision, recall, receiver operating curve (ROC), computation time, and sum rate. |
