Spectrum Allocation Based on Simultaneous Wireless Information and Power Transfer in Cognitive Radio Networks

Abstract

This paper proposes an incentive simultaneous wireless information and power transfer (SWIPT) scheme for cognitive radio networks, wherein the secondary user (SU) provides wireless power transfer for the primary user (PU) in the exchange for partial bandwidth of the latter. To enable the proposed scheme, both the secondary transmitter (ST) and the primary receiver (PR) are equipped with multiple antennas. Specifically, SU helps to charge PU via adjusting its beamforming vector and PR harvests energy with a subset of its antennas; as a reward, PU allocates part of its bandwidth to SU. Our goal is to jointly optimize the beamforming vector of SU, the bandwidth allocation and covariance matrix of PU, such that the transmission rate of SU is maximized and meanwhile PU’s requirements in transmission rate and energy are satisfied. This optimization problem is non-convex. To handle this non-convex problem, we decompose it into two independent subproblems. One is to optimize the covariance matrix, the other is to optimize the beamforming vector of SU as well as the bandwidth allocation of PU, which is divided into a two-layer optimization problem. In the inner subproblem, we obtain the optimal beamforming vector for any given fixed bandwidth allocation, while in the outer subproblem we propose a gradient based algorithm and obtain the optimum bandwidth allocation. Numerical results are given to show the achievable system performance with varying parameters and also the convergence of the proposed algorithm.