Distributed Resource Allocation for SWIPT-Based Cognitive Ad-Hoc Networks

Energy supplies, spectrum resources, and transmission overheads of wireless nodes are the bottlenecks in decentralized networks (e.g., device-to-device communication networks). In this paper, we study the total power minimization problem of secondary users (SUs) for a multiuser simultaneous wireless information and power transfer-based cognitive ad-hoc network by jointly optimizing the transmit power and power-splitting (PS) coefficients of SUs in a distributed way, where SUs can harvest ambient radio-frequency signals to prolong the lifetime of nodes via a PS-based scheme. Firstly, a resource allocation (RA) problem with perfect channel state information (CSI) is formulated under the quality of service constraints and the minimum energy-harvesting (EH) constraints. The non-convex problem is decomposed into a subproblem with a high EH threshold and a subproblem with a low EH threshold. The distributed closed-form solutions are obtained via Lagrange dual theory. Secondly, to overcome the impact of channel estimation errors, a robust RA problem with imperfect CSI is studied under bounded uncertainties. Both the feasible region and robust sensitivity are analyzed to give an insight into system performance. Simulation results demonstrate that the effectiveness of the proposed algorithm.