Research On Energy Efficient Cooperative Transmission Techniques In Ultra-dense Cloud Radio Access Networks

The continuous innovation of the mobile Internet,along with the contin?uous evolution of the mobile application ecosystem,have created ubiquitous access demand for wireless networks,resulting in exponential growth of wire-less traffic.Ultra-dense deployment of lightweight access nodes enabled by the Cloud Radio Access Networks(Cloud-RAN),emerges as an important way for operators to cope with the huge traffic demands.However,severe interference and high energy consumption,which are two inherent problems of network den-sification,seriously degrade the energy efficiency of ultra-dense Cloud-RAN.Cooperative transmission is a promising method that can effectively suppress interference and improve network throughput and energy efficiency.Therefore,with the prominent of environmental issues and the deteriorated imbalances be-tween the Operators' revenue and expenditure,designing the cooperative trans-mission strategies for the ultra-dense Cloud-RAN,reducing interference and controlling the energy consumption,fully reaping the benefits of network den-sification,is very significant.There are three main research direction for the energy efficient cooperative transmission in the ultra-dense Cloud-RAN.Firstly,although large-scale coop-erative transmission can greatly reduce interference and increase throughput,it also leads to a proliferation of complexity and energy consumption,which in-stead limits the energy efficiency gains.Therefore,cooperation transmissions need to be elastically implemented in the form of small-scale clusters,in order to achieve the optimal balance between energy consumption and throughput gain.Secondly,the operating energy consumption generated by the hardware infras-tructures such as access nodes is the main part of the total energy consumption,while the traditional sleeping mechanism has drawbacks in maintaining cov-erage and suppressing interference.Therefore,it is necessary to combine the sleeping mechanism with cooperative transmission to transform the hardware infrastructures into flexible network resources that can be dynamically occu-pied and released according to the traffic load,which can greatly improve the network energy efficiency.Finally,cooperative transmission brings new prob-lems while improving the energy efficiency of ultra-dense Cloud-RAN,leading to great increase of data volume on the fronthaul,also making energy consump-tion on the fronthaul become an important part of the total energy consumption.Therefore,cooperative transmission needs to be implemented together with the innovative fronthauling technologies such as fronthaul compression,in order to fully reap the energy efficient benefits brought by network densification with the limited fronthaul capacity.This thesis focuses on the above three problems and the main contents and novelties are as follows:1.Virtual cell based energy efficient cooperative transmission:The virtual cell is an extension of the traditional cell concept.It has the ad-vantages of low scheduling complexity,high compatibility,and relatively low fronthaul capacity requirements.The existing researches seldom study the opti-mal trade-offf between the benefits and the extra energy consumption brought by virtual cell based cooperative transmission.This thesis models the virtual cell based cooperative transmission in the ultra-dense Cloud-RAN as a combined non-convex optimization problem.In order to reduce the solving complexity,a heuristic iterative optimization framework is proposed.The original problem is first decomposed into two sub problems,that are energy efficient beamforming and energy-efficient virtual cell clustering problems.Fraction programming,continuous lower bound tight convex approximation,block coordinate descent,together with the Lagrangian duality method,coalition formation game method are utilized to transform and solved these two sub problems.Through the cross-iterative solving of the two sub-problems,the optimal virtual cell clustering and beamforming scheme that makes the network energy efficiency maximal are obtained.The simulation results show that the proposed strategy has fast con-vergence speed,and can significantly improve the network energy efficiency,meanwhile achieve the optimal trade-off between throughput gain and energy consumption brought by the cooperative transmission.2.Energy efficiency oriented,overlapping clustering and beamforming design:The overlapping clustering breaks down the cell-based service model of cellular networks.As a new form of transmission mode,this type of cooper-ative transmission takes a full consideration of user's quality of service(QoS)requirements and can achieve the highest cooperation gain.Existing researches seldom study the energy efficiency maximization through the overlapping clus-tering.In this thesis,the implicit relationship between the clustering informa-tion and the global beamforming vector,along with the channel uncertainty are considered when modeling the optimization problems.The overlapping cluster-ing based cooperative transmission is modeled as a sparse optimization problem under the QoS requirements and transmission power constraints,with the goal of maximization the energy efficiency.The fractional programming,continu-ous lower bound tight convex approximation and iterative weighted l1-norm approximation are utilized to transform the original problem into a tractable sparse optimization problem.The problem is then solved by the Lagrangian duality method.Finally,we obtain the specific association threshold which can be used to directly obtain the clustering pattern,and also the closed-form of optimal beamforming coefficients.Simulation results show that the proposed strategy can balance the throughput gain and energy consumption brought by cooperative transmission,can provide user-centric QoS guarantee,and has sig-nificant advantages in improving network energy efficiency.3.Traffic-aware,cooperative transmission aided two-tier sleeping strat-egy:The traffic-aware sleeping mechanism is the most efficient way to improve the energy efficiency of ultra-dense Cloud-RAN.The existing sleeping mecha-nisms usually only operates on one time scale and cannot adapt to the differen-tiated characteristics of traffic on different time scales.In order to provide QoS guarantee with the least active network resources,this thesis utilizes the advan-tage of Cloud-RAN on the node management,and propose a two-tier sleeping strategies which includes deep sleep mode and opportunistic sleep mode,ac-cording to self-similarity on the large time scale and the short-term randomness of wireless traffic.Two sleeping mode determination algorithms are further proposed.Firstly,utilizing the traffic prediction mechanism and the capability of coverage expansion provided by cooperative transmission,we determine the access nodes that need to enter deep sleep mode for a long time period.Then,in each short time slot,the opportunistic sleep mode determination and the co-operative beamforming are jointly optimized.Through solving this energy effi-ciency maximization oriented joint optimization problem with the Lagrangian duality method,the unnecessary antenna elements that should be shut down and enter opportunistic sleep mode are selected.Simulation results show that the proposed strategy can realize the hierarchical sleeping of network hardware resources,significantly improve the energy efficiency.4.Joint optimization of fronthaul compression and cooperative beamform-ing for energy efficiency maximization:Although cooperative transmission can greatly improve the energy effi-ciency of ultra-dense Cloud-RAN,it also leads to a rapid increase the data volume on the fronthaul and hence the fronthaul energy consumption.It si necessary to combine the cooperative transmission with fronthaul compression when the fronthaul capacity is limited.However,fronthaul compression has a negative impact on the performance of cooperative beamforming and de-grades the energy efficiency gain.In this thesis,the joint optimization of fron-thaul compression and cooperative beamforming for energy efficiency maxi-mization is studied.Considering the limitation of fronthaul capacity,some of users' QoS requirement may not be guaranteed,a two-step optimization frame-work,which includes the access control stage and the joint optimization stage,is proposed.Introducing sparse variables,the access control problem is mod-eled as a sparse optimization problem.Smooth lp-norm approximation and Majorization-Minimization algorithm are used to solve this problem and obtain the maximal users set that can be supported through joint fronthaul compression and cooperative beamforming.Through this way,the throughput is maximized.Then at the second stage,we perform joint fronthaul compression and coopera-tive beamforming optimization to reduce the energy consumption and maximize energy efficiency based on the results from the first stage.Simulation results show that the proposed strategy is adaptive and can flexibly select the largest set of users that can be successfully accessed under multiple constraints,mean-while maximize the energy efficiency through the joint optimization fronthaul compression and cooperative beamforming.