Robust Relay Beamforming in Device-to-Device Networks with Energy Harvesting Constraints

Motivated by the observation that energy harvesting (EH) from radio-frequency (RF) signal is subject to fluctuations, multiple EH-enabled relays are employed to collaboratively enhance data communications in a device-to-device (D2D) network underlying a cellular system. Each relay is equipped with a single antenna and unable to harvest energy and transmit data simultaneously. Thus, the D2D user equipment (DUE) needs to optimally schedule the channel time for the relays' EH and data transmissions, which depends on their EH capabilities and channel conditions. Considering that the relays' channel estimations are usually unreliable, we formulate a robust throughput maximization problem to optimize the relays' EH time and transmit power, subject to a probabilistic interference constraint at the cellular user equipment (CUE). We show that the proposed problem, though non-convex, can be tackled by exploiting its monotonicity structure. Specifically, we design a successive approximation algorithm that involves solving a sequence of semi-definite programs (SDPs) and show numerically that it always achieves the global optimum. This validates our analysis and demonstrates the efficacy of the proposed algorithm.