Spatial modeling and analysis of traffic distribution based on real data from current mobile cellular networks

Abstract

A deep understanding of spatial traffic patterns modeling is not only essential for both theory analysis and simulation estimation of cellular network's performance, but also can provide us insight on the energy efficient deployment of future network architecture. Since the user behavior in cellular network is affected by the environment, which leads to that the spatial traffic fluctuations exhibit an obvious non-uniform patterns. Existing spatial traffic models are based on ideal assumptions, and these models fail to depict the real spatial traffic distribution cases in the current mobile cellular network. In this paper, based on real data from a mobile network operator in one large city of China, "Truncated two-dimensional DCT (Discrete Cosine Transform)" model is proposed, which characterize the spatial traffic pattern more precisely. Furthermore, comparison of the modeling results has been illustrated among three typical regions: dense urban, urban and suburban, which shows that the parameters of model formula for these three regions are different.

Modeling Realistic Bit Rates of D2D Communications between Android Devices

“…Certain domains commonly infer the distance between two devices from the received signal strength. This is notably done in wireless sensor networks [13,22], vehicular networks [4,21], and in mobility detection of smartphones [17].…”