AI and game-based efficient resource allocation and interference mitigation scheme for D2D communication

Abstract

Device-to-device (D2D) communication essentially means to communicate between two devices without the base station interfering. D2D communication is used in wireless 5G networks, the Internet of Things, Bluetooth and WiFi, and vehicular networks. D2D communication is useful as it increases spectral efficiency, network efficiency, user experience, and throughput, all of which are better than those available in cellular communication. Thus, D2D communication is a preferred mode of communication. However, D2D communication faces issues like mode selection, interference, resource allocation (RA), and security. The need for RA stands because multiple D2D users (DUs) compete for the same resources. Now, the goal should be to optimize RA to enhance the spectrum efficiency and network coverage. Still, the issues here are caused by eavesdroppers (ɛ), devices that can interfere with communication and encrypt/decrypt the messages. In our proposed scheme, we combined artificial intelligence (AI) and coalition game theory to resolve the issues of optimized RA and security. The coalition game is used for efficient RA, but it will take all the DUs that increase the overhead into the system. To mitigate this issue, we proposed an AI-based solution, which selects the best DUs based on their channel conditions. Our major evaluation parameters were accuracy, sum secrecy capacity and data rate.