Research On Resource Allocation In Multicell Massive MIMO Systems

Energy efficiency(EE)has been considered as one of the key performance metrics for the fifth generation(5G)wireless communication systems.Re-cently various energy-efficient designs have been proposed to address the EE optimization problem from several aspects of the networks,ranging from ra-dio access techniques to network control.As one of the major advances in physical-layer,massive multiple-input multiple-out(MIMO)bearings great po-tentials to improve not only spectral efficiency but also EE.Equipped with tens to hundreds of antennas,massive MIMO is able to efficiently mitigate interfer-ence while focusing energy towards the target users by using relatively simple transceivers.Therefore,it has been attracting attention from both academia and industry.Designing massive MIMO for EE has been a recent focus towards 5G.Noting the influential interaction among the system parameters of the massive MIMO systems,the number of active base station antennas,the user scheduling and the transmission power allocation are considered as the three major opti-mization objectives in EE designs.In this thesis,the number of active antennas at each base station,the user scheduling and the transmission power allocation are optimized or jointly optimized to achieve high EE in the multicell scenario,and the main contributions are listed bellow.1.An asymptotic equivalent expression of the data rate for the considered scenario is derived.In the multicell massive MIMO scenario,an asymptotic equivalent expression of the data rate is derived,which includes the number of active base station antennas,user scheduling,channel estimation error,pilot contamination as well as transmission power allocation factors.Unlike the ex-isting works that assume an equal number of scheduled users,our derivation is more general and fits well to our considered EE optimization problem,where each cell may have distinct system configurations and user scheduling.2.Energy efficient resource allocation for multicell downlink massive MIMO systems with ZF precoding is investigated.An interference-aware re-source allocation scheme for multicell downlink massive MIMO systems is proposed,where the number of active antennas at each base station,the user scheduling and the transmission power allocation are jointly optimized to achieve high EE.In contrast to the conventional single-cell/multicell multi-objective op-timization frameworks,our scheme only collects and updates the interference information in a centralized fashion but enables parallel cell-wise interference-aware optimization.In this thesis,a two-step interference-aware cell-wise op-timization framework is proposed.Firstly,the interference-aware joint opti-mization of the number active BS antennas and user scheduling is performed,and then the interference-aware EE-oriented power allocation is performed.The interference-aware joint optimization achieves notable performance gain as compare to single-cell scenarios.3.An energy-efficient resource allocation of the number of active antennas at each base station with daily load profile(DLP)is proposed.In the multicell downlink massive MIMO systems with ZF procoding,each base station adap-tively adjusts the number of active antennas to reduce the waste of resources.The proposed scheme conducts adaptive energy-efficient optimization,where the number of active antennas is optimized according to the resource load or the requests from accessed users.The interference information is firstly ob-tained by the single-cell optimization,and then the interference-aware energy efficient optimization of the number of active base station antennas is proposed.The proposed scheme can effectively trade off the sum rate and the total power consumption,and then achieve high EE.