Project Title: Acoustic Modeled Sound Source Localization and Extraction

Faculty Sponsor: Na (Linda) Zhu

Department: Engineering

Telephone: 810.424.5451

E-Mail: nazhu@umflint.edu

Project Description: Sound source localization and extraction have been a very practical yet challenging problem with applications in various areas. For example, noise detection is always the first step in industrial noise control, as well as implemented in structural health monitoring and quality control in a factory floor; aerodynamically generated noise and its localization is under concern in automotive industry; tracking and tracing of a suspect in a noisy environment is interested in homeland security for intelligence applications; extraction of target sound source and de-noising is in high demand of acoustic related devices, such as media players, cell phones, and hearing aids, etc.

The proposed project aims at developing an innovative acoustic modeled sound source localization and extraction methodology. With real-time data collected by multiple sensors, the location of sound sources can be identified with modified triangulation algorithm, and a digital signal of target sound source will be extracted. Numerical simulation will be accompanied, and the influence factors on localization and extraction effectiveness will be evaluated and discussed.

Completion of the project will have a significant impact on the field of industrial noise control, structural health monitoring, and acoustic related devices. Particularly, a numerical model used for hearing aids will be developed. The proposed model will not only have a function to amplify the directly measured sound signal but also be able to extract target sound from a noisy environment. The acoustic assumption of the proposed model is independent of sound frequencies and directions thus break the directional limitation on the hearing aid devices in the current market. In other words, patients or elder people will be able to achieve de-noised target sound from all over directions in three-dimensional spaces, without adjusting their head position or facing direction, therefore, the new sound extraction methodology can lead to a more comfortable experience for the hearing aids users.

Student Tasks & Responsibilities: Students are responsible for part of the literature review and numerical simulation.

Minimum Student Qualifications: Matlab and Labview skills required. Preference give to students took courses in sound and vibration.