This paper was presented at the `Traffic management strategies for the future' session. Traffic data represents a major component of Pavement, Bridge and Safety Management systems (PMS, BMS and SMS). It is essential for investment programming and design purposes. Proper estimation and use of traffic data are highly important. Average Annual Daily Traffic (AADT), traffic composition, speed, truck configuration and axle weights are among the traffic attributes that are essential for PMS, BMS and SMS. These attributes are typically collected using three classes of traffic stations, unclassified traffic counters (UTC), classified traffic counters (CTC) and Weigh-in-Motion (WIM) stations. Selecting the number and locations of these stations to provide adequate coverage for an entire province or state is not an easy task and has a significant impact on the accuracy of the estimated and forecasted traffic attributes, and hence the investment programming. This selection should be performed in a way that maximizes the coverage, allows detection of as many traffic patterns as possible and allows efficient data sharing among different stations. This paper presents a probabilistic based procedure used to select the optimum number of traffic stations and to distribute these stations among traffic regions. Factors considered in this procedure include: homogeneous traffic regions; variation in traffic characteristics among traffic regions and highway classes; budget constraints; performance constraints in terms of accuracy and confidence levels for the estimated traffic characteristics. A Traffic Diversity Index (TDI), which is based on the probability of selecting a traffic pattern, is used to compare the relative abundance of various traffic patterns at different locations. TDI values are computed for all locations. Locations with the highest TDI values are selected until the required number of each type of station is satisfied. (A)
Samenvatting