Randomized Robot Motion Planning
Speaker: Prof. David Hsu ,
Date: September 21 2004
Time: 4:00PM to 5:00PM
Location: 32-D463 (Star Conf. Room)
Host: Prof. Daniela Rus, CSAIL
Contact: , theresa@csail.mit.edu
Relevant URL: Abstract:
Motion planning is a fundamental and yet computationally hard problem in
robotics. Over the past decade, random sampling brought tremendous
progress to motion planning of robots with many degrees of freedom,
by providing an effective tool to capture the connectivity of complex,
high-dimensional geometric spaces. It can plan paths for robots with as many as
36 degrees of freedom with relative ease. Despite the success, it is not well
understood why random sampling works well for motion planning. In this talk,
I will describe the notion of expansive spaces, in an attempt to characterize
robot configuration spaces for which random sampling works
well (or not so well). I will also show how to construct simple, effective
samplers to deal with narrow passages, the main computational bottle-neck
faced by randomized motion planning algorithms.
See other events that are part of AI Colloquium Series 2004/2005
See other events happening in September 2004
