Power Allocation Based LSTM-FCN in D2D Underlaying with Multi-Cell Cellular Network

Abstract

Device-to-Device (D2D) communication is one of the key technologies to achieving higher speeds, lower latency, and less energy. D2D communication is direct link communication between two communication devices, meaning that communication can occur without going through the base station. However, because communication occurs without going through the base station and D2D users do not have their resources, D2D users simultaneously use the resources owned by Cellular User Equipment (CUE) to communicate and cause interference. Power allocation is optimized to mitigate the interference between D2D users and CUEs and maximize the system's overall sum rate. The traditional power allocation scheme in D2D communication still has problems related to the efficiency of the allocation, coordination of interference, and limitations for operating in real-time systems. This work focuses on designing the Long Short Term Memory with Fully Convolutional Network (LSTM-FCN) algorithm suitable for the power control problem on a D2D underlay communication system with an uplink-side multi-cell scheme. The simulation results show that enhancement of CUE can increase the system's sum rate and energy efficiency. At the same time, enhancement of the D2D pair can also increase the sum rate but decrease energy efficiency. Both LSTM-FCN, LSTM, and FCN can approximate the performance of the conventional scheme (CA-based algorithm). Besides that, LSTM-FCN gets the smallest time complexity compared to the other two algorithms and gets the closest performance to CA in both scenarios above 97% accuracy.