A simple magnetic signature vehicles detection and classification system for Smart Cities

Vehicle recognition is one of the main challenges in Intelligent Transportation Systems (ITS). The need to recognize the vehicle type can help insurance companies, public safety organizations, infomobility, and policy-makers in general. In this paper, we propose a vehicle recognition system based on speed estimation, vehicle length estimation and classification of the vehicle type. We developed a real time system for vehicle recognition based on four steps: a storage of the magnetic signature of the vehicle, speed estimation, estimation of the length of the vehicle and vehicle recognition. The latter has been realized through matching between the measured waveform with information in a database containing magnetic signatures of vehicles. Matching was realized using the Dynamic Time Warping (DTW) method. Experimental results involving 10 vehicles and 50 trials show successful identification of approximately 98% of the considered vehicles.