Motorcycles Exhaust Recognition based on Acoustic Signal using Time-Domain Audio Features
Efraheim Ghera Mahanaheim and Bambang Heru Iswanto

Department of Physics, Universitas Negeri Jakarta, Jl. Rawamangun Muka, Jakarta 13220, Indonesia

Abstract

Motorcycle noise levels are not only determined by the sound pressure level but also by the audio characteristics produced by their exhausts. In this study, motorcycle are classified into racing and non-racing type based on their the exhaust sound produced by them. The simple time-domain feature is used for classifying both types of vehicle sounds. The characteristic features of the two types of vehicles are analyzed based on the main components and reported. The results of vehicle type classification using neural networks show an accuracy above 90%.

Keywords: signal classification, acoustic signal processing, time-domain audio features, motorcycles exhaust

Topic: Computer Science