RUB 

Source Localization

"Where did that sound come from?" This is the question that acoustic source localization algorithms try to answer. Acoustic source localization finds application in the most diverse of fields, e.g., in the automotive industry, for detecting errors during manufacturing, in the noise reduction components of hands-free systems, teleconferencing etc. The algorithms exploit the spatial distribution of the microphones and the statistical characteristics of the source signals in order to estimate the location of a source. It is obvious that when the environment is noisy and echoic, localization is difficult. Yet, this is the situation encountered in most applications. Thus, robust approaches to acoustic source localization are the need of the hour, especially with the proliferation of speech driven interfaces.

Vergrößerung

8-Channel Microphone Array with Preamplifier



The theses/projects offered in this area focus on the development and implementation of algorithms for robust localization of sources in reverberant and noisy environments. The student is expected to build upon existing frameworks and examine methods to improve the used models for higher precision. Another aspect of the localization problem is the reliable detection of multiple sources: topics in this area would include a statistical modelling of the multi-source problem and development of appropriate localization algorithms.

Vergrößerung

Localization Plot with Source at
80° Azimuth



Prerequisite for these topics is interest in the field of signal processing. Basic programming skills in MATLAB and/or C++ are advantageous, but not compulsory.

Contact: Alexander Schasse