Saturday, June 27, 2009

PhD studentship at Institute of Intelligent Systems and Robot

PhD studentship

PhD studentship available at Institute of Intelligent Systems and Robotics
(http://www.robot. jussieu.fr/ ?op=view_ page&id=2& page=Object〈=en)
University of Paris 6 (www.upmc.fr) .

The position is available within the framework an interdisciplinary
project HuPer and Lunettes Intelligentes
(http://www.robot. jussieu.fr/ fra/projets/ lunintel/ index.html)
on "Cognitive navigation in space ; application to design of mobility
assistance for visually impaired". Cooperation with London University
(professor Steve Maybank).

Project summary.
ISIR (Institute of Intelligent Systems & Robotics) (former Laboratoire de
Robotique de Paris,) of Paris 6 University studies neuro-cognitive
human-space interaction in order to design and realize prototypes of
intelligent systems, robotics and mechatronics systems, highly autonomous
and/or able to assist healthy and handicapped (especially visually
impaired), in performing everyday life tasks. The provision of autonomous
mobility for the visually impaired is a difficult but important societal
problem. Solutions based on GPS and a digital map of the environment are
not suitable because of the low, 10m, precision of GPS, poor satellite
signal reception in urban areas and indoors, and the difficulty of
responding to changes in the environment. Artificial vision in the form of
SLAM (Simultaneous Localization and Mapping) provides a satisfactory
solution for autonomous mobile robots. However, SLAM is not suitable for
assisting the visually impaired, because human navigation is far more
complex than that of a mobile robot. Human navigation leads to the
definition of natural visual SLAM (SLAM-VN). SLAM-VN is a simulation of
the visual perception-motion loop, including feedback from inertial
sensors to the imaging system, in analogy with the human vestibulo-ocular
feedback. This should reinforce the robustness of classical stereovision
systems and solve the problem of the spatio-temporal instabilities caused
by uneven ground and by sudden and unpredictable human movements. This
thesis aims to define precisely the concept of SLAM-VN and model it in
hardware and software. We will use the ISIR convergent stereovision system
to which we will add an inertial system, which will play the role
analogous to the human vestibular system. The combined system is named
intelligent glasses (IG). One of challenges of the IG system is the
correct synchronization and fusion of the visual information and the
inertial information. The information about the environment obtained by
SLAM-VN will be used for object recognition and tracking, and obstacle
detection. The Fisher-Rao metric will be used to quantify the accuracy
with which the objects and obstacles can be described. This will lead to
the design of new vision processing algorithms, never investigated before.
3D object localization (precision of ~ 0.5 m) will be established via
RANSAC type algorithms. Object tracking will use a classical approach such
as a particle filter. The SLAM-VN model will be validated by experiments,
including navigation in real environments by both the visually impaired
and by the robot iCUB (European humanoid robotic platform of ISIR
laboratory). The interface for visually impaired will include a touch
stimulating interface upon which obstacles will be displayed
Key words: visual perception, stereo, inertial data, data fusion,
localization, object recognition, Fisher-Rao metric, RANSAC, particle
filter,visually impaired mobility assistance

- Starting date: October 2009
- Duration: 3 years
- Required education: Master in Computer Science, or Robotics, or
Intelligent Systems, or Cognitive Science, or Human/animal
Physiology/Biology obtained in 2009 in a European Union University;
- Be 26 year old or less

- Required skills: Programming in C and C++; good level of written/spoken
English
(French not required)
- Experience/studies in computer vision, image processing, robotics, VLSI is
an advantage.
- The successful candidate will be hired as PhD student and receive
Ministry of Education scholarship with social benefits.

Deadline for application : June 28, 2009 (first preslection) , July 2st
(final application)

If interested, please post, e-mail (pdf format only), or fax:
for preselection
1- C.V.
2- Copy of original study certificates with list of exams, grades, and scale
(indicate min, max, and pass threshold) for Master
3- Copy of international publications, if any (max 3)

For final application
1- C.V.
2- Motivation letter (career plan, motivation for PhD, statement of research
interest)
3- Copy of original study certificates with list of exams, grades, and scale
(indicate min, max, and pass threshold) for Master
4- Contact addresses of two reference persons
5- Copy of international publications, if any (max 3)

to:

Prof. Edwige Pissaloux
ISIR
Université Paris 6
Contact : Edwige.Pissaloux@ upmc.fr
Fax : + 33 2 32 10 69 90

Given the high number of applications and enquiries, further information can
be provided only to applicants who have submitted all items listed above.

============ ========= ========= ========= ========= =
Edwige Pissaloux ph. 33(0)144272874
Professeur fax : 33(0)144275145
Vision VLSI/System Design email:Edwige.Pissaloux@ upmc.fr
ISIR/UPMC & CNRS UMC 7220
Pyramide tour 55
4 place Jussieu
75 252 Paris Cedex 05, France
IEEE France Section EMBS Chapter Founder and Past Chair
Présidente de la Section Signal & Image du Club
EEA Section
============ ========= ========= ========= ========

No comments: