Vehicle classification and accurate speed calculation using multi-element piezoelectric sensor

Abstract

Vehicle monitoring and classification is a necessary Intelligent Transportation System ITS activity, as nationwide departments of transportation (DOT) use the information to effectively design safe and durable roadways. Because over 70% of the weight of goods shipped in the U.S. are trucked, substantial pavement damage is becoming more and more problematic [1]. Thus, an accurate classification system for estimating vehicle parameters is sorely needed. Currently, the most widely used classification solution consists of a combination of inductive loops and piezoelectric sensors. Installing these systems causes pavement damage. Even more challenging is that current systems greatly under-classify class 1 motorcycle vehicles. In this paper we present a novel system for classifying vehicles and determining track width and speed. The system employs a multi-element piezoelectric sensor positioned diagonally across a single traffic lane; a data acquisition unit; and a processing and classification algorithm operating on a computing device. Vehicle front axle tires distinctively impact different element sensors, which aids in calculating track width, speed and axle spacing. Given these factors, a classification decision can be made using vehicle axle spacing. The developed system was tested on highway conditions. Classification accuracy was 86.9% overall and even better for class 1 motorcycles (100%) and passenger vehicles (98.9%).