Teacher-Student Learning based Low Complexity Relay Selection in Wireless Powered Communications

• DOI: 10.48550/arxiv.2402.02254
• Publication Date: 2024-02-03
• Authors: Aysun Gurur Önalan; Berkay Köprü; Sinem Coleri

Radio Frequency Energy Harvesting (RF-EH) networks are key enablers of massive Internet-of-things by providing controllable and long-distance energy transfer to energy-limited devices. Relays, helping either or information transfer, have been demonstrated significantly improve the performance these networks. This paper studies joint relay selection, scheduling, power control problem in multiple-source-multiple-relay RF-EH under nonlinear EH conditions. We first obtain optimal solution scheduling for given selection. Then, selection is formulated as a classification problem, which two convolutional neural network (CNN) based architectures proposed. While architecture employs conventional 2D convolution blocks benefits from skip connections between layers; second replaces them with inception blocks, decrease trainable parameter size without sacrificing accuracy memory-constrained applications. To runtime complexity further, teacher-student learning employed such that teacher larger, student smaller CNN-based distilling teacher's knowledge. A novel dichotomous search-based algorithm determine best network. Our simulation results demonstrate proposed solutions provide lower than state-of-art iterative approaches compromising optimality.