Site Verification of Weigh-in-Motion Traffic and TIRTL Classification Data

Quality weigh-in-motion (WIM) traffic data is essential not only in general transportation application, but also in pavement design. The new AASHTO Mechanistic-Empirical Pavement Design Guide (MEPDG) for New and Rehabilitated Pavement Structures requires information on the detailed truck traffic, such as truck traffic volume, truck traffic monthly and hourly variations, vehicle class distribution, axle load, and axle load distributions, instead of the traditional ESALs. In addition, the Indiana Department of Transportation (INDOT) needs to collect traffic data frequently so as to timely provide accurate traffic information for planning, program development, operations, and pavement management. Currently, INDOT is using the pneumatic road traffic counters in traffic data collection, such as particular short-term or temporary traffic data collections. However, the pneumatic road traffic counter requires installation of rubber tubes on the pavement surface. As a result, the installation of rubber tubes usually creates safety issues to our workers and is timely consuming and labor intensive. Therefore, it is an urgent need for INDOT to utilize new devices to enhance the safety of field traffic data collection without compromising data quality. This study consists of two parts. The first part is to verify the accuracy of WIM vehicle classification and develop models for vehicle classification corrections using image processing technologies. The second part is to install and then evaluate a traffic surveillance system, i.e., the Transportable Infra-Red Traffic Logger (TIRTL). In the first part, the investigators collected video and WIM traffic data at WIM sites statewide. A digital image based vehicle monitoring and classification system was developed for verifying weigh-in-station data, in particular the vehicle classification counts. Based on the real world WIM and video traffic classification data, allocation factors were determined for correcting the unclassified vehicle counts associated with the WIM traffic data. In the second part of this study, a TIRTL system was installed to collect traffic data near a WIM site. Hourly traffic data was first gathered manually and by video cameras to verify the potential errors associated with the TIRTL vehicle counts. Large amount of daily WIM traffic data was also utilized as baseline data to evaluate the field performance of TIRTL and assess the impact of various weather conditions, such as fog, rain and snow, and thunderstorm on TIRTL’s performance. The evaluation was based on the FHWA Scheme F Vehicle Classification and solely a data-driven process