Novel vehicle and motorcycle classification using single element piezoelectric sensor

Abstract

Accurate data reporting ensures suitable roadway design for safety and capacity. Currently, vehicle classifier devices employ inductive loops, piezoelectric sensors, or some combination of both to identify 13 different FHWA vehicle classifications. Systems using inductive loops have failed to accurately classify motorcycles and record relative pertinent data. Previous investigations have focused on classification techniques utilizing inductive loop signal output, magnetic sensor output with neural networks, or the fusion of several sensor outputs. This paper presents a novel vehicle classification setup that uses a single piezoelectric sensor placed diagonally across the traffic lane to accurately identify motorcycles from among other vehicles by detecting the number of vehicle tires. A vehicle classification algorithm based on number of tires detected and axle/tire spacing was formulated and deployed in an embedded system for field testing.