Selective data collection in vehicular networks for traffic control applications

摘要

Vehicular sensor network (VSN) is an emerging technology, which combineswireless communication offered by vehicular ad hoc networks (VANET) withsensing devices installed in vehicles. VSN creates a huge opportunity to extendthe road-side sensor infrastructure of existing traffic control systems. Theefficient use of the wireless communication medium is one of the basic issuesin VSN applications development. This paper introduces a novel method ofselective data collection for traffic control applications, which provides asignificant reduction in data amounts transmitted through VSN. The underlyingidea is to detect the necessity of data transfers on the basis of uncertaintydetermination of the traffic control decisions. According to the proposedapproach, sensor data are transmitted from vehicles to the control node only atselected time moments. Data collected in VSN are processed using on-linetraffic simulation technique, which enables traffic flow prediction,performance evaluation of control strategies and uncertainty estimation. Ifprecision of the resulting information is insufficient, the optimal controlstrategy cannot be derived without ambiguity. As a result the control decisionbecomes uncertain and it is a signal informing that new traffic data from VSNare necessary to provide more precise prediction and to reduce the uncertaintyof decision. The proposed method can be applied in traffic control systems ofdifferent types e.g. traffic signals, variable speed limits, and dynamic routeguidance. The effectiveness of this method is illustrated in an experimentalstudy on traffic control at signalised intersection.