Automated Vehicle Classifier for Traffic Management and Control

Abstract

Vehicle classification is the process of vehicle type recognition based on given vehicle characteristics. Accurate vehicle classification has many important applications in transportation. One example is road maintenance, which is highly related to the monitoring of heavy vehicle traffic. Because trucks and oversized vehicles exhibit distinctly different performance characteristics from passenger cars, the continuous updating of those vehicles with respect to their share in daily traffic will help estimate the life of current road surface and assist in the scheduling of road maintenance. Design of a toll system can also use the same information. Moreover, by obtaining the heterogeneity of traffic flow, vehicle classification information can lead to more reliable modeling of vehicle flow. Incorporating the information of vehicle classification in the analysis of environmental impact is also highly desirable since different vehicle types have different degree of airborne and noise emission. The class of vehicle is one of most important parameters in the process of road traffic measurement. Improvement of highway safety can also benefit from vehicle classification information, knowing thatthe severity of traffic accidents is highly correlated with vehicle types. To summarize, an area-wide assessment of the component of vehicle classes in traffic is essential for more reliable and accurate traffic analysis and modeling. This paper focuses on automatic vehicle classification algorithm development based on the advanced loop detector data and expands' the study scope by testing proposed algorithm transferability.