Hidden Markov modeling and macroscopic traffic filtering supporting location‐based services

Abstract

Abstract In this study, we present a technique that combines pattern recognition techniques with cellular signaling measurements and more precisely information extracted from Abis air interface in GSM network. The pattern recognition is applied to measurement reports that mobile terminal (MT) sends to its serving base station (BS). Modeling of these reports is performed by hidden Markov model (HMM) while employing clustering large applications (CLARA) as clustering method. The accurate results during MT velocity estimation located inside a probe vehicle show the potential of the method when applied to large scale of MTs in order to estimate basic parameters for road traffic. Copyright © 2006 John Wiley & Sons, Ltd.