Seminar: "On-line acoustic source localisation with VoxNet"
Title: On-line acoustic source localisation with VoxNet
Speaker: Mike Allen, Cogent Computing ARC, Coventry University
Date/Time: 11am Thursday 4 December
Location: 59/4240 (ESD Demo Room)
This talk will describe the development, deployment, evaluation cycle for an interesting distributed acoustic source localisation application: in-situ monitoring of animals. The end users of this application are bioacoustics researchers, who aim to learn about the behaviour of animals based on the calls they make and the locations/situations they are made in. The application is built using VoxNet, a hardware and software platform for distributed, high data rate acoustic sensing. The platform supports the provision of on-line position estimates, in order that user can act on them in-situ (for example to take a picture of the animal of interest).
Learning from a succession of VoxNet-based network deployments, several approaches to improve the timeliness of this system are proposed (Adaptation of processing and Lazy Grouping), intelligently distributing the burden of processing between the sink and nodes in the network.
Whilst the proposed improvements have been evaluated for this particular animal monitoring application, their suitability is more general to a wider class of acoustic localisation applications.
Michael Allen received a BSc in Computer Science from Coventry University in 2005. He is currently a PhD student with the Cogent Computing Applied Research Centre at Coventry University, where he is advised by Dr. Elena Gaura.
Michael's research is based around high data rate wireless embedded sensing systems, with particular emphasis on acoustic sensing, and localisation-motivated applications. In 2007, Michael was a visiting graduate student at the Centre for Embedded Networked Sensing (CENS), UCLA, under the guidance of Prof. Deborah Estrin (UCLA) and Dr. Lewis Girod (MIT). Michael has gathered expertise in the development and in-situ deployment of wireless sensing systems, as well as ad-hoc wireless networking and signal processing for acoustic localisation.
Michael expects to graduate in 2009, and has taken a position as a post-graduate research associate with the Singapore MiT Alliance for Research and Technology in Singapore (SMART), as part of the Centre for Environment Sensing and Modelling (CENSAM). He will continue to work with high data rate sensing systems, for water quality monitoring.
Posted by Dr Geoff Merrett on 02 Dec 2008.