Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/83802
Title: Localization and characterization of noise sources in urban environment by microphone array signal processing
Authors: Tam, Kai Chung
Degree: Ph.D.
Issue Date: 2013
Abstract: This thesis is concerned the application of microphone array technology in characterizing sound sources propagating outdoor. Microphone array processing has proved its flexibility in estimating the location or the directional of arrival of sound sources. The performance of the estimation is highly depending on the knowledge and assumption of the sound source and sound propagation model. While considering the sound source propagating outdoor, the system of sound propagating is distorted from basic sound propagation model. The distortion would induce performance degrade of the estimation. To enhance the flexibility of microphone array processing technology, a further development of the processing method is required. The highway traffic noise is decomposed in spatial domain with delay and sum beamforming. A half hour traffic noise monitoring with microphone array is conducted at sideway of a highway in Hong Kong. The contribution of the traffic noise from particular carriageway is quantified in A-weighted sound pressure level. The result from microphone array is validating with an empirical traffic noise calculation model (CRTN). Point moving source parameters are extracted with time-frequency approach. Estimation of source velocity, frequency and amplitude based on change of instantaneous frequency and time-frequency difference is derived. Four different time-frequency transformation method is studied for the estimation of instantaneous frequency and their performance is compared. The method of localization of source and estimation of the ground reflection parameter for a point source above a locally reactive ground surface is proposed. A non-linear least square approach is adopted to solve the estimation problem. The Levenberg-Marquardt (LM) method is selected based on the flexibility of estimating parameters with different scale. MUSIC localization method is used to give the initial estimation of the source location. Forward backward spatial smoothing is adopted as a pre-processer of the localization to minimize the effect of source coherence. Performance of the proposed method is analyzed statistically to illustrate its repeatability. The estimation of the ground parameter is further extended into a point moving source in uniform speed. The source is considered travelling in direction parallel to a planar microphone array. With the prior knowledge of the source velocity, and a close initial estimation of source location, the ground parameter with a point moving source is estimated with the Levenberg-Marquardt method. This thesis extend the microphone array processing technique from conventional simple source propagation assumption into models with consideration of practical situation arises outdoors. Further research is suggested for a robust estimation of source parameters and more complex sound propagation models.
Subjects: Microphone arrays.
Signal processing.
Hong Kong Polytechnic University -- Dissertations
Pages: xxix, 172 p. : ill. (some col.) ; 30 cm.
Appears in Collections:Thesis

Show full item record

Page views

42
Last Week
0
Last month
Citations as of Mar 24, 2024

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.