University of Toronto

Supervisor: Prof. Ashish Khisti, Professor, University of Toronto, Toronto, Canada

Partner Organization: Mircom Technologies Ltd.

Duration: Jan 2011 - present

The project requires design and development of a proprietary video-conferencing software over an Ethernet based LAN (wired or wireless), over TX3-Touch Series Telephone Access Systems developed by Mircom Technologies Ltd. The project proposes a new architecture combining adaptive FEC codes, hybrid ARQ techniques and path-diversity over a secure transport protocol in a loopback mechanism for Video LAN streaming solution to ensure efficient and reliable multimedia transmission. A new class of error correction codes; Diversity Embedded Streaming Erasure Codes (DE-SCo) will be put into application for video transmission to ensure minimum delay.

Supervisor: Prof. S. R. M. Prasanna, Associate Professor, Indian Institute of Technology, Guwahati, India

Duration: July 2007 - Apr 2008

The project was motivated by human capability of identifiying a known person even by listening to non-speech sounds like coughing, humming, laughing, crying etc. The course of project involved filtering of speaker-specific features of non-speech vocal emissions which are uncommon to the speech sample space, and inculcating them to conventional Speaker Recognition model to produce a more robust identification system. The new system improved the performance of the conventional system by at most 70%. This system can produce effective results for speakers with vocal disorders and throat diseases.

Supervisor: Prof. T. V. Sreenivas, Professor, Indian Institute of Sciences, Bangalore, India

Duration: May 2007 - July 2007

The project introduced a novel method of echo cancellation exploiting spectral subtraction and peak matching in frequency domain. The main purpose of coming up with a method in frequency domain was to overcome the disadvantages of conventional time-domain adaptive algorithms employed for echo cancellation in real-time systems.

Research Lab: Centre for Applied Research in Electronics, IIT Delhi, India

Supervisor: Prof. Arun Kumar, Professor, Indian Institute of Technology, Delhi, India

Duration: May 2006 - July 2006

The project involved design and development of a simulator for multi-channel signal communication employing various communication schemes and algorithms specific to underwater acoustic signals. It involved various signal encoding and decoding schemes, pulse shaping and filtering techniques and modulation and demodulation of the signals.