A quarter-car vehicle model based feature for wheeled and tracked vehicles classification

Seismic signals are widely used in ground vehicle classification due to their inherent characteristics. In this paper, a kind of feature extracted by the Short-time Power Spectrum Density (STPSD) from ground-vehicle-induced seismic signals was introduced for wheeled and tracked vehicle distinction. According to theoretical analysis, based on a simplified quarter-car vehicle model, the cepstrums of the STPSD of seismic signals can depict the structure differences of wheel and track. In addition, the extracted feature is also less affected by the underlying geologies than traditional Power Spectrum Density based features. It is verified with mixed datasets from our field experiments and SensIT project.