Talk Abstract:
This talk discusses an ongoing research effort regarding
the development
of autonomous underwater vehicles with particular emp

hasis on vision-based
sensing. We have been developing an underwater veh

icle for several
applications notably the environmental assessment of c

oral reefs habitats.
Semi- autonomous behavior underwater is especially
challenging since it
combined 6 degree of freedom mobility restricted

communications hard
real-time constraints and unstructured environments

. I will describe the
system design of a small underwater and amphibiou

s robot that uses
computer vision as its principal sensing modality and
some of the ongoing
challenges we have encountered. This includes an o

utline and discussion
of how to accomplish operator control of the vehi

cle using a vision-based
human-robot interface. The exploits a combinat

ion of a symbol-recognition
system with a gestural inference and a speci

al- purpose visual language.
We also make use of Markov Random Fields f

or color correction (and we
hope for scene reconstruction) I will comm

ent on the use of physical
feedback for behavior control and the develo

pment of a vision-based user
interface. This is join work with doctoral

candidates Philippe Giguere and
Junaed Sattar as well as Anqi Xu and o

ut colleagues at York University
led by Michael Jenkin.

Speaker

Bio:
Gregory Dudek is a Professor with the School of Computer Science a

nd an
Associate member of the Department of Electrical Engineering at Mc

Gill
University. He is the Director of McGill''s Research Center for Int

elligent
Machines a 20 year old inter-faculty research facility. In 200

2 he was
named a William Dawson Scholar (an honorary chair). He directs
the McGill
Mobile Robotics Laboratory.

He has recently been on t

he organizing and/or program committees of
Robotics: Systems and Science
the IEEE International Conference on
Robotics and Automation (ICRA) t

he IEEE/RSJ International Conference
on Intelligent Robotics and System

s (IROS) the International Joint
Conference on Artificial Intelligence
(IJCAI) Computer and Robot Vision
IEEE International Conference on M

echatronics (ICM2005) and International
Conference on Hands-on Intellige

nt Mechatronics and Automation
(HIMA2005) among other bodies. He is pre

sident of CIPPRS the Canadian
Information Processing and Pattern Recog

nition Society an ICPR national
affiliate.

He was on leave in 20

00-2001 as Visiting Associate Professor at the
Department of Computer Sc

ience at Stanford University and at Xerox Palo
Alto Research Center (PA

RC). He obtained his PhD in computer science
(computational vision) from
the University of Toronto his MSc in computer
science (systems) at the
University of Toronto and his BSc in computer
science and physics at Qu

een''s University.

He has published over 150 research papers on subj

ects including visual
object description and recognition robotic navig

ation and map construction
distributed system design and biological per

ception. This includes a book
entitled Computational Principles of Mobil

e Robotics co-authored with
Michael Jenkin and published by Cambridge Un

iversity Press. He has
chaired and been otherwise involved in numerous

national and international
conferences and professional activities conce

rned with Robotics Machine
Sensing and Computer Vision. He research int

erests include perception for
mobile robotics navigation and position e

stimation environment and shape
modeling computational vision and col