CSAIL Event Calendar: Previous Series

Efficient Motion Planning using Generalized Penetration Depth Computation

Speaker: Liangjun Zhang , Computer Science, UNC Chapel Hill
Date: July 24 2009
Time: 2:00PM to 3:00PM
Location: 32-D507
Host: Russ Tedrake, CSAIL

Contact: Russ Tedrake, russt@mit.edu
Relevant URL: http://www.cs.unc.edu/~zlj

Motion planning algorithms compute valid motion for robots to safely navigate in complex environments by avoiding collisions. Besides robotics, these techniques are also useful for factory automation, CAD/CAM, virtual environment, animation, and computational biology. Even though there is extensive work in motion planning over the last three decades, I address two important unsolved problems in this area: (1) How can a robot determine that no feasible path exists for a given task? (2) How can a robot plan its motion more efficiently in cluttered environments?

In this talk, I will first present novel geometric algorithms for quantifying the extent of intersection between 3D models, namely generalized penetration depth. This leads to a new formulation to query the infeasible space of robot's motion. Based on this formulation, I will present the first practical path non-existence algorithm for low-DOF robots. An extension to feedback motion planning by computing a global vector field over the robot's entire free space is also shown. I will then show a retraction-based sampling planner based on generalized penetration depth computation, which can efficiently solve the "Alpha Puzzle" benchmark and disassembly tasks frequently arisen in virtual prototyping. Finally, I will talk about some future work on feedback motion planning and digital human modeling (DHM).

Bio: Liangjun Zhang is a PhD candidate of the Department of Computer Science at University of North Carolina at Chapel Hill. He completed BS and MS degrees in Computer Science at Zhejiang University, China. His research interests include robot motion planning and geometric computation and robotics. He has published 17 papers in these areas. He received the best paper award at the CAD conference in 2008. He also received the 2008 Chinese government award for outstanding Ph.D. students abroad.

See other events that are part of Robotics Seminar Series 2009/2010

See other events happening in July 2009


About Us Research News Resources Directory