Joint Optimization for IRS-Assisted Self-Powered IoT in 5G mmWave Networks

Harvesting energy from ambient energy source is the key technology for self-powered Internet of Things (IoT) devices to maintain continuous operation without an external power supply. Motivated by the expansion and popularity of 5G networks, we propose a novel solution for IoT devices which are self-powered via harvesting energy from the millimeter-wave (mmWave) communications in 5G mmWave networks. For overcoming the high path loss in mmWave communications, directional narrow-beam transmission is adopted to provide sufficient link budget between transceivers through beamforming technology, which however makes IoT devices difficult to scavenge energy from the mmWave signals. Hence, we employ multiple intelligent reflecting surfaces (IRSs) to assist in energy harvesting at the IoT devices and data transmission at the 5G users. Considering beam codebook design for 5G mmWave networks, this paper jointly optimizes the Discrete Fourier transform (DFT) codebook-based transmit codevectors at the 5G base station (BS) and the phase shifts of IRS's reflective elements for minimizing BS's transmit power, while satisfying the Signal to Interference plus Noise Ratio (SINR) constraints at users and energy harvesting constraints of IoT devices. Nevertheless, owing to the intricate coupling of variables and discrete constraints, this joint optimization problem is extremely non-convex and non-linear. To address such challenges, we propose a penalty dual-decomposition (PDD)-based algorithm which combines the penalty-based augmented Lagrangian method and block coordinate descent method. It explores the structure of the mmWave channel and performs a double iterations in which the joint optimization problem is decomposed into several simplified subproblems. Simulation results reveal that the above algorithm enhances the energy efficiency as compared to other algorithms.