Non-Invasive Axle-Based Vehicle Classification Utilising Tracking Radar Technology

This paper demonstrates the feasibility of using tracking radar technology to provide non-invasive axle-based vehicle classification. A new algorithm that extracts the physical characteristics of the vehicle from its radar signal return is presented. The vehicle-induced Doppler is decomposed into two parts: translational Doppler shift from which we estimate the length of the vehicle; and wheel rotation-induced micro-Doppler from which we estimate the vehicle's axle configuration (i.e. axle groups, axle numbers and wheelbases). A decision tree provides the final vehicle classification according to an axle-based vehicle classification scheme based on the estimated characteristics.