Blind Doppler Shift Compensation of Vehicle Noise and its Characterization for Traffic Noise Simulation

Abstract

While a huge amount of literature has been devoted to the study of traffic noise, its aural simulation has received much less attention. Such simulation enables the determination of the sound both outside and inside buildings by means of proper filters. In recent years, a technique based on sound recording of each isolated vehicle in a fixed station, followed by Doppler compensation and noise characterization has been developed. Although this technique takes into account each source more precisely (such as engine and exhaust noise with load effect) as well as aerodynamic noise, it depends on the emission of a pilot tone from the source. This tone allows for the retrieval of the vehicle speed. In this paper, a new Doppler compensation method is proposed, based on the vehicle speed estimation from the tones contained in the recorded signal. This technique may be applied to any vehicle as far as it is reasonably isolated from other vehicles, it does not contaminate the signal of interest, and it needs no pilot tone emission from the source, making the method more practical. For aural simulation, a new model based on a polar scheme is considered and proposed. In this paper, the proposed method is assessed with experimental data, and results are compared with those of other existing techniques.