Early Stage Researchers
Research training opportunities
Two research training positions are available at the Institute of Sound and Vibration Research (ISVR) at the University of Southampton as part of the AUDIS project. They will give first-hand experience of conducting cutting-edge research and be an ideal preparation for scientists or engineers contemplating registering for a PhD afterwards. The start date is 3 January 2012 and the maximum duration is until the end of September 2012. They are open to candidates with a bachelor’s or master’s degree in engineering or similar discipline having appropriate skills in signal processing and its application to human problems. Applicants must have less than 4 years’ experience of research as a graduate and must satisfy the requirements of EU Marie Curie schemes, including transnational mobility (normally must move from outside the UK to commence the position).
Salary
Trainees will receive a generous package of salary and allowances up to a maximum of approximately 35,000 Euros for 9 months.
How to apply
Send in the following documents via email to the AUDIS lead at ISVR (Prof Mark E Lutman, mel@isvr.soton.ac.uk):
• CV (including personal details, academic history, work experience, experimental skills, publications)
• Copies of relevant diplomas and transcripts
• Two letters of recommendation
• Proof of proficiency in English (IELTS 6.5 or equivalent)
• Letter of motivation (research interest, reasons for applying to programme and host)
Closing date for applications: 05 November 2011. Applicants may be contacted by telephone or email for further information. Successful applicants will be notified by 20 November at the latest.
Summary of research projects
The following are examples of research projects to be undertaken during the training period. There will be limited flexibility to accommodate the skills and interests of trainees. This is not an exhaustive list.
Sparse speech processing for cochlear implants
Cochlear implants (CI) are devices for profoundly deaf people that stimulate the auditory nerve by means of electrical pulse trains to provide artificial hearing. A novel speech processing method based on sparse coding theory has been proposed for the CI speech processor and has shown promising performance for enhancing speech in noisy environment. To further improve the performance of the novel speech processor, this project will investigate key elements of the algorithm, including efficient sparse representation, optimal weighting function and robustness by using dictionary learning and Bayesian methods. Other aspects of sparse coding may also be explored.
Development and validation of an experimental real-time processor for cochlear implants
Algorithms for cochlear implant (CI) speech processors are typically developed using generic software tools such as Matlab. They are evaluated by processing signals off-line and presenting prepared stimulus sequences to patients via their CI. This approach is laborious and prevents immediate adjustment of parameters to meet patient requirements; it is also unnatural for the patient. Work is underway to develop new algorithms on a real-time platform using Simulink and the Matlab Real-time Workshop. The project will implement sparse coding algorithms on the real-time platform and evaluate their performance on patients with CIs.
