Design of a Safe Navigation Scheme for a Robotic System in a Dynamic and Uncertain Environment
Internship Description:
The design of a robotic system that can move safely in an environment
featuring moving objects whose future behaviour is unknown remains a
challenge. When it comes to collision avoidance, the concept of
Inevitable Collision States (ICS) is useful. An ICS is a state such that
no matter what the future trajectory of a robotic system is, it
eventually collides with an object. This concept is general and permits
to take into account the dynamics of both the robotic system and the
moving objects. Clearly, for its own safety, a robotic system should
never be in one of these ICS. The ICS concept was formally introduced in
[1] and later exploited in a number of applications.
Characterizing the ICS set in a given situation theoretically requires
information about the future evolution of the moving objects. Given the
elusive nature of the future, this knowledge is difficult to obtain in
practice. Prediction of the future is in order here. With prediction
comes uncertainty and the primary purpose of this internship is to
explore how to design a collision avoidance scheme that takes into
account this uncertainty and yet strive to guarantee a given level of
safety. The work will be based upon the concept of Probabilistic ICS
that was recently developed. The navigation scheme proposed will be
evaluated in simulation or, time permitting, on a robotic wheelchair.
References:
[1] Th. Fraichard and H. Asama. Inevitable collision states. a step
towards safer robots? Advanced Robotics, 18(10),
2004. http://emotion.inrialpes.fr/fraichard/publications.
As a general rule, I am looking for highly motivated candidates with a good
educational record, the ability to think independently and work
autonomously. If you are interested in this internship, please send me a
complete Curriculum Vitae (featuring your academic background with the
courses attended, the grades and the rankings obtained - very important),
plus any additional information you might find useful.
