Hidden Markov Filtering with Microscopic Traffic Modeling for Vehicle Load Estimation in Cellular Networks

T. Stamoulakatos, A. Yannopoulos, T. Varvarigou, and E. Sykas (Greece)


Location Based Services, Traffic Information, Pattern Recognition, Hidden Markov Model, Microscopic Traffic Model.


Location Based Services is a new category of services for mobile phone (MT) users based on MT location. Various techniques can be found in the literature for MT’s location estimation. Promising appear to be hybrid techniques which overcome existing limitations of cost, accuracy and network coverage. The technique applied in our study is based on pattern recognition together with Time Advance (TA) measurements. The pattern recognition is performed by Hidden Markov Model (HMM) which is trained with downlink prediction data modeling the strength of the received signals for specific areas. Sets of appropriate HMMs are built with respect to the Timing Advance (TA) which indicates the distance between a base station (BS) and the MT. Being able to identify the TA gives us a first estimation of the location of the MT and which set of HMMs we should use to determine MT’s position estimation. This estimation will be based on the comparison between the prediction of the radio level signaling of the area and the RSSI reports of the MT, without the need of MT modifications or cellular network upgrades. After that, Microscopic Traffic modeling can result in vehicle load estimation in main city routes providing in that way Traffic Information Service to MT users.

Important Links:

Go Back