A Novel Criterion for Vehicle Classification using Inductive Vehicle Signatures

Inductive Loop Detectors (ILD) are the most commonly used sensors in traffic management systems. Using the acquired inductive signatures, most proposed systems classify the vehicles using a criterion based on the estimation of the vehicle length, which requires to have a good a-priori estimate of its speed. Contrary to such standard proposals, in this paper we present a method for vehicle classification based on the criterion of the Fourier Transform (FT), which shows several interesting properties: firstly, robustness against variations in vehicle speed or constant acceleration, and secondly, only one inductive signature is required. Our method will be evaluated using real inductive signatures captured with a hardware prototype also developed by us. Keywords—vehicle classification, inductive loop detectors, Fourier transform, sensors, signal processing, data acquisition, data classification, decision making, hardware prototype.