Deep Learning Based Transmit Power Control in Underlaid Device-to-Device Communication

Abstract

In this paper, a means of transmit power control for underlaid device-to-device (D2D) comm proposed based on deep learning technology. In the proposed scheme, the transmit power of D2D user equipment (DUE) is autonomously learned via a deep neural network such that the weighted stun rate (WSR) of DUEs can be maximized by considering the interference from cellular user equipment. Unlike conventional transmit power control schemes in which complex optimization problems have to be solved in an iterative manner which possibly requires long c imitation time, in our proposed scheme the transmit power can be determined with a relatively low computation time. Through simulations, we confirm that the proposed scheme achieves a sufficiently high WSR with a sufficiently low computation time.