Abstract:

This paper deals with the layout of sensors in traffic networks,a key problem related to the economicbenefit of intelligent transportation system ( ITS) . In the investigation,first,a new measure describing the spatialcharacteristics of traffic information is used to construct a maximum integrated value model ( MIVM) based on thegame between the maximum information value and the minimum investment cost of sensors,and the model is solvedby constructing a directed graph and by making use of the shortest path algorithm ( SPA) of the directed graph.Then,the model parameters are analyzed and the corresponding mathematical methods to describe and calibrate theparameters are presented.Moreover,in order to simplify the problem,a simplified model based on the MIVM isgiven and the relationship between the sensor number and the integrated value is determined.Finally,a case studyon Beijing 3rd Ring Road is performed,with the optimal section sensor number and the best sensor spacing beingdetermined as 142 and about 676 m,respectively.The results indicate that,as compared with the widely-appliedgraph theory models,the MIVM helps to obtain not only the optimal sensor number but also the optimal location ofeach sensor. Besides,from the calibration results of model parameters in different conditions,it is found that theMIVM can be universally applied because it can be easily used for different cities,roads and sensor types.

Key words: transportation system engineering, traffic sensor network, sensor layout, maximum integrated value model, optimization