Acoustic noise pattern detection and identification method in doppler system

Abstract

This paper discusses detection and identification of the moving vehicles based on prerecorded acoustic noise patterns. Object movement and object type differences would lead to the significant vehicle pattern changes. Doppler shift estimations of the measured signal spectral components will improve the recognition and classification performance. The proposed pattern identification methods could be used for the robust single sensor system or as a part of complex multi-sensor and learning solutions. The experimental results illustrate measured signal processing stages. © Kauno technologijos universitetas, 2012.