Cooperative Transmission of Energy-Constrained IoT Devices in Wireless-Powered Communication Networks

In Internet of Things (IoT) systems, a number of sensor devices monitor the physical system states and exchange information with each other. The main limitation is that the IoT devices are generally energy constrained since those are powered with batteries. To address this energy problem, we consider a cooperative wireless-powered communication network (WPCN), which consists of three phases: 1) downlink (DL) energy transfer from a multi-antenna access point (AP); 2) data sharing among IoT devices; and 3) uplink (UL) information transfer from single-antenna IoT devices. Based on the shared data and the harvested energy, the single-antenna IoT devices in the neighborhood cooperate to form a virtual antenna array in order to transmit their information simultaneously to the multi-antenna AP using a multiple-input multiple-output (MIMO) technique in the UL information transfer phase. In this study, the transmit covariance matrices (i.e., beamforming vectors and the corresponding transmit power allocation) used for both DL energy transfer and UL information transfer are jointly designed to maximize the UL capacity based on the Lagrangian method. Furthermore, the time allocation for each phase is optimized based on a stochastic gradient method. In the numerical results, it is shown that our proposed beamforming scheme and the stochastic time allocation can achieve near-optimal performance.

I. Introduction

With a growing demand for long service lifetime of devices in IoT [1], [2], supplying sufficient power continuously to the IoT devices is a crucial problem in practice. The lifetime of IoT devices can be prolonged by replacing or recharging batteries but the cost may be high or it is sometimes impossible under certain circumstances such as unreachable devices after installation. One of the practical approaches for prolonging the lifetime of IoT devices is to harvest energy at a device from external sources, such as solar, wind, geothermal, but which are usually neither controllable nor reliable [3]. As another approach, in ambient backscatter communication, the IoT device can transmit information without an internal power source by modulating ambient RF signals [4]–[6]. RF signals, which are man generated and thus more manageable, can be used for energy harvesting in a more active way by WPT technologies, i.e., directly transferring the energy to IoT devices via RF signals [7]–[9]. In the SWIPT system, a transmitter transmits RF signals for transferring information and power, and the receivers then perform information decoding and energy harvesting [10]–[18]. There are three different SWIPT techniques: 1) power splitting [10]–[14]; 2) time switching [10], [13], [15]; and 3) antenna switching [16], [17]. In the power-splitting approach, a receiver divides its received signal into two parts for energy harvesting and information decoding. In the time switching approach, each time slot is used for either energy harvesting or information decoding alternatively. In the antenna switching approach, a set of antennas are used for energy harvesting while the others are used for information decoding.