QoS And Energy Efficiency Analysis For Multiple Data Sources In Wireless Network

The fifth generation mobile communication technology(5g)has many challenges compared with the fourth generation mobile communication technology(4G).In terms of service quality:to achieve high reliability and low delay,the next generation mobile communication network has higher network energy efficiency(100 times higher)and lower end-to-end delay(1/10 lower);in terms of application scenarios,enhanced mobile broadband and large-scale machine communication greatly expand the service objects of the mobile communication network.This makes data traffic grow almost exponentially.In order to meet these challenges,5g adopts super dense networking mechanism.In order to solve this problem,the number of base stations will increase dramatically and the energy consumption will increase.This thesis focuses on how to improve the energy efficiency under the delay limitation in the next generation mobile communication network.Wireless communication network nodes,whether mobile phones or base stations,have corresponding scenarios of multiple data sources,such as multiple applications running on the same phone require service of communication at the same time or the same base station providing services for multiple mobile phones at the same time.For each time slot,different data sources may have data that need to be processed with different lengths.In order to better analyze the QoS and energy efficiency performance of the multiple data sources wireless communication network,we focus on the analysis method of energy efficiency performance under the condition of guaranteeing quality of service(QoS)in the wireless network scenario through the multiple data sources data model.In this thesis,the model of multiple data sources wireless network is established firstly,and the calculation method of buffer length distribution is proposed based on the theory of effective capacity and effective bandwidth.Then,the energy efficiency of multiple data sources wireless network is analyzed by using dual-mode circuit and the theory of effective capacity and effective bandwidth.Finally,a new channel parameter estimation method is proposed,which makes the proposed energy efficiency performance analysis method in use It’s more extensive.First of all,we propose a new multi data sources model,which extends the existing single data source model:a transmitter has several data sources to serve at the same time,they have a certain probability in each time slot without data to be processed.After giving the mathematical definition of this model,we introduce the theory of effective bandwidth and effective capacity and introduce the formula of the length distribution of buffer in transmitter.The matrix calculation method is proposed to calculate the length distribution of buffer more conveniently.After that,we prove the existence of multiple roots of QoS index in the multi data sources model.Secondly,according to the energy efficiency model of multiple data sources wireless network we update system model and with previous achievements,we adopt multi-mode circuit transmission mechanism,which has the high energy consumption transmission state and low energy consumption idle state,and determine the formula of energy efficiency for the new model.The relationship between the concept of energy efficiency and the buffer non empty probability is introduced,which makes it possible to analyze the energy efficiency of multiple data sources wireless communication system to guarantee QoS.Moreover,we prove the existence of multi real roots of QoS exponent in multiple data sources wireless network.Finally,the existing channel parameter estimation methods are summarized.Based on the nonparametric estimation theory,a new channel estimation method is proposed,which widens the scope of application of the new methods for QoS and energy efficiency analysis of multiple data sources wireless network.By introducing the samples into the moment generating function,the channel characteristics can be estimated without any assumptions,and the convergence of the algorithm is faster.