Research On Base Station Activation With Stable Access Relationship Guarantee In C-RAN

With the explosive growth of mobile date traffic,the traditional wireless access network can not meet the new requirements of future wireless communication because of its low coverage,serious interference,and difficulty in maintenance and upgrade.A cloud-based wireless access network(C-RAN)came into being.In C-RAN,RRH(Remote radio Heads)can be deployed in an intensive manner and increase signal quality and edge coverage through multi-point collaboration.The number of RRH participating in cooperative communication is often limited due to backhual capacity,interference suppression,and scheduling complexity.In the traditional RRH activation study,RRH activation/dormancy is generally based on channel state information,considering energy consumption,channel capacity and other factors to choose.In this channel-base selection mode,the system often operates in a high-performance state because of matching channel changes.But when the channel changes faster,RRH state will change rapidly,which will increase the difficulty of scheduling,accelerate the wastage of equipment,and even lead to service interruption.In this thesis,the author focus on the base station activation algorithm based on stable access relation in C-RAN network,and aim to find a scheduling strategy which can balance the flexibility and stability of the system.(1)A kind of mathematical model of joint optimizing resource scheduling and beamforming design based on stable access relationship between users and base station is proposed.Different from the traditional network model,this model considers a better balance between the flexibility and stability of the BS-user association,it is not limited to the performance of a certain timeslot,but focuses on the performance of a long-term resource scheduling strategy to achieve the purpose of reducing energy consumption to enhance capacity.(2)Based on the non-causal channel,an offline RRH activation strategy is designed,and two online activation strategies based on causal channel are designed.The long-term model of the system is different from the instantaneous system model,and its solution is often subjected to the causality of the channel.Based on the non-causal channel.In this thesis,the author explores a centralized RRH activation algorithm based on the non-causal channel,and then designs two RRH activation strategies for causal channel: dynamic activation strategy and training-base activation strategy.The simulation results show that compared with the traditional non-switched control activation algorithm,the three RRH activation algorithms proposed in this thesis can improve the stability of the system and the dynamic activation algorithm has higher flexibility in two kinds of online algorithms.(3)A distributed solution scheme based on Alternating Direction Method of Multipliers(ADMM)is designed for three strategies.In the centralized solution method,the algorithm complexity is very high because the large variable dimension and the variable coupling.In order to solve this problem,a distributed computing method based on ADMM is proposed.The ADMM algorithm can effectively separate the coupled variables,and divide the original problem into many subproblems that can be solved independently.This solution not only reduces the computational complexity,but also conforms to the network architecture of the C-RAN.Finally,the simulation results show that the solution converges well,and the computational complexity can be greatly reduced.