Classification of Mobility of Cellular Phone Using Linear Classification and k-Clustering

Abstract

Road traffic data is a fundamental element of intelligent traffic system. However, due to the high investment of the road sensor, the availability of the traffic data is so limited that it can’t satisfy the requirement of current situation. Using cellular phone information as road traffic data becomes an attractive alternative because of its low cost, widespread and high cover rate. Until now, there are several algorithms to process the cellular phone information and most of them present promising conclusion. However, in order to get the satisfying conclusion, nearly all of these methods depend on a high amount of sample, in which way will detract the real-time performance of the traffic regulation and control. This is we don’t want to see. In this paper, we proposed a process to collect the information of cellular phone based on the simulation of working mode of the real base station, i.e., putting an appropriate instrument on the side of the road to detect the cellular phone passing by. Using the data we got, then we proposed a method to classify the mobility of the cellular phone, which is the critical problem of the analysis of the cellular phone information. Two key attributes are the average vehicle velocity and the variance of the vehicle velocity.