- This event has passed.
Clustering Nature of BS Deployment and Traffic Demand in Cellular Networks
13 September 2017 @ 15 h 30 min - 17 h 00 min
Presenter: Yifan Zhou
Abstract: This work is based on large-scale real data from on-operating cellular networks. Traditionally, the architecture and service paradigm of cellular networks are based on the homogeneous assumptions of base stations (BSs) deployment and user requests pattern, such as hexagonal grid and Poisson arrivals. However, due to the evolutionary progresses, cellular networks become more and more complicated in these decades. In this seminar, we will firstly review the state-of-art real data analysis of cellular traffic consumption on different dimensions (time, space and content) which implies the clustering nature clearly. After that, we conduct the identification process based on our collected real data in different aspects, i.e., the spatial distribution of BSs, the temporal and spatial density of traffic demand. Our results shows that the popular Poisson processes are not accurate for BS locations or traffic modelling and the heavy-tailed distribution (alpha-stable distribution) is good estimator for BS and traffic density on temporal or spatial point of view. Finally, in consideration of the intrinsical relevance, we try to connect the traffic demand and network deployment by revealing the linear relationship between spatial densities of traffic and BSs in cellular networks.