Real-Time Tracking of Multiple Acoustic Sources in Reverberant Environments

In 2014 I graduated from my study Electrical Engineering at Eindhoven University of Technology. The result of this graduation project is a paper with the title 'Real-Time Tracking of Multiple Acoustic Sources in Reverberant Environments'. About locating acoustic sources in indoor environments using multiple microphones.

Abstract - Traditional acoustic source localisation methods attempt to find a single source with audio data of the current time only. For many applications, localising multiple sources is necessary. Furthermore the effects of reverberation regularly disturb the accuracy of such localisation algorithms. This work proposes a Steered Response Power derivative which allows to efficiently localise multiple acoustic sources. A framework is presented that is composed of the proposed localisation algorithm and a tracking system which is based on a particle filter. This particle filter takes the temporal data of sources into account to enhance the accuracy and to track the path of a source over time. A real-time implementation of the framework is developed which is able to process 128 microphone channels at a sample frequency of 46875Hz and achieve an accuracy of 5cm in a real reverberant and reasonably noisy environment. Continue reading →