
For FullText PDF, please login, if you are a member of IEICE,
or go to Pay Per View on menu list, if you are a nonmember of IEICE.

Optimal Power Splitting and Power Allocation in EHEnabled MultiLink MultiAntenna Relay Networks
Shengyu LI Wenjun XU Zhihui LIU Junyi WANG Jiaru LIN
Publication
IEICE TRANSACTIONS on Communications
Vol.E100B
No.8
pp.14801488 Publication Date: 2017/08/01 Publicized: 2017/02/09 Online ISSN: 17451345
DOI: 10.1587/transcom.2016EBP3255 Type of Manuscript: PAPER Category: Wireless Communication Technologies Keyword: multilink multiantenna relay, energy harvesting, independent power splitting, cooperative power allocation, channel state information,
Full Text: PDF>>
Summary:
This paper studies the multilink multiantenna amplifyandforward (AF) relay system, in which multiple sourcedestination pairs communicate with the aid of an energy harvesting (EH)enabled relay and the relay utilizes the power splitting (PS) protocol to accomplish simultaneous EH and information forwarding (IF). Specifically, independent PS, i.e., allow each antenna to have an individual PS factor, and cooperative power allocation (PA) i.e., adaptively allocate the harvested energy to each channel, are proposed to increase the signal processing degrees of freedom and energy utilization. Our objective is to maximize the minimum rate of all sourcedestination pairs, i.e., the maxmin rate, by jointly optimizing the PS and PA strategies. The optimization problem is first established for the ideal channel state information (CSI) model. To solve the formulated nonconvex problem, the optimal forwarding matrix is derived and an auxiliary variable is introduced to remove the coupling of transmission rates in two slots, following which a bilevel iteration algorithm is proposed to determine the optimal PS and PA strategy by jointly utilizing the bisection and golden section methods. The proposal is then extended into the partial CSI model, and the final transmission rate for each sourcedestination pair is modified by treating the CSI error as random noise. With a similar analysis, it is proved that the proposed bilevel algorithm can also solve the joint PS and PA optimization problem in the partial CSI model. Simulation results show that the proposed algorithm works well in both ideal CSI and partial CSI models, and by means of independent PS and cooperative PA, the achieved maxmin rate is greatly improved over existing nonEHenabled and EHenabled relay schemes, especially when the signal processing noise at the relay is large and the sources use quite different transmit powers.

