Wireless Information and Power Exchange for Energy-Constrained Device-to-Device Communications

Abstract

This paper studies device-to-device wireless communications, where two energy-constrained Internet-of-Things (IoT) nodes, which do not have constant power supplies, wish to exchange their information with each other. Because of small form factor, the IoT nodes are normally equipped with simple energy storages, which might suffer from a high self-discharging effect. Therefore, the energy stored in each node would not be available after a few time duration. In this system, we investigate power splitting (PS)-based energy exchange methods by exploiting radio frequency (RF) wireless energy transfer techniques, and propose a new concept called wireless information and power exchange (WIPE). In this WIPE protocol, each node operates either in a transmit mode and a receive mode at each time slot. First, a transmit node sends the information signal to a receive node which utilizes a PS circuit for information decoding and energy harvesting. Then, the harvested energy of the receive node is stored in the energy storage. At the consecutive time slot, two nodes switch their operations, i.e., the receive node in the previous time slot now operates in a transmit mode which transfers RF signals by using the harvested energy. This procedure continues by changing the operations of two nodes at each time slot. For the proposed WIPE protocol, we provide two different PS ratio optimization schemes which maximize the weighted sum throughput performance according to the level of channel state information (CSI) knowledge. For the ideal full CSI case where the CSI for all time slots is known in advance, the globally optimal PS algorithm is presented by applying convex optimization techniques. Also, for a practical scenario where only the causal CSI is available, we propose an efficient PS optimization method which achieves performance almost identical to the ideal full CSI case. Simulation results verify that the WIPE protocol with the proposed PS optimization techniques performs better than conventional schemes.