Performance Of Wireless Network With Spatio-temporal Traffic

With the continuous evolution of wireless networks,the era of fifth generation wireless systems is coming.Various new wireless applications are constantly emerging,and traffic in wireless networks tends to be complex and diversified.New methods are needed to describe the spatio-temporal random characteristics of traffic in order to better study network performance.The transmission performance in wireless network is deeply related with the traffic.The working status of the base station in wireless networks will be affected by the traffic distribution.The denser the traffic distribution,the busier the base station will be.At the same time,the busy situation of the base station and the interference of the base station are related to each other.The interference pattern,in turn,affects the performance of the transceiver links in the wireless network,which directly determines the service process of the traffic.Therefore,modeling and analysis of the traffic are essential to design and configure the wireless networks so as to match the network service with the traffic.Studying the traffic in the wireless network can better analyze the network performance and is helpful to the network design and performance improvement.In order to characterize the status of various new wireless applications in the network and meet the users’ service requirements,it is necessary to conduct some research and analysis on the traffic.The content of this thesis can be classified into two parts.The first part is about the research of wireless network performance in static networks with spatio-temporal traffic.Inspired by the custom of human social activities and the real measured data from cellular network operators,we find that the spatial distribution of traffic tends to be clustered in the actual network.The Poisson point process model commonly used in the academic research cannot reflect the clustering characteristic of traffic,so we introduce a non-uniform point process to describe the clustering and make a comparative study with the uniform point process.We first analyze the statistical characteristics of traffic and explore the fluctuations of the uniformly or non-uniformly distributed traffic in the network.We find that the fluctuation of non-uniformly distributed traffic is significantly larger than that of uniformly distributed traffic.Then,we derive the delay and the stability of users’ queues.According to the simulation results,we find that when the density of users is large,the stability of users’ queues is higher in the non-uniformly distributed case.The second part is about the research of wireless network performance in mobile networks with spatio-temporal traffic.Due to the exponential growth of mobile traffic,more and more communication scenarios need to consider the movement of nodes.Therefore,it is of practical significance to explore the spatio-temporal traffic in the networks with mobility.In order to study the network performance in the networks with mobility,we first propose a constrained independent and identically distributed mobility model,which is described as the homogeneous mobility model and the Gaussian mobility model.We explore the effects of spatio-temporal fluctuation traffic on the network performance under two scenarios: fast mobility and partial mobility.Then,we explore the relationships between the quality of experience(QoE)and the system cost in the networks with mobility.We find that QoE does not always increase as the system cost increases,but begins to stabilize when the system cost increase to a certain extent.We also find that the mobility in wireless networks may be useful to reduce the delay and improve the QoE.Conclusively,in this thesis,we uncover the non-uniform nature of the traffic in the cellular networks.From the two dimensions of space and time,combining the two theoretical tools of stochastic geometry and queueing theory,we explore the effects of traffic with different spatial distribution and time arrival distribution on network performance.The spatial distribution of traffic is modeled by using stochastic geometry and the temporal arrival process of traffic is modeled by using queueing theory.We extend the research on traffic from the static networks to the networks with mobility,and explore the relationship between QoE and system cost in the network with mobility considering the spatio-temporal traffic.