🚀 go-pugleaf

           DEPARTMENT OF COMPUTER SCIENCE
             COLORADO STATE UNIVERSITY

                    BARNEY'S
           MONDAY AFTERNOON CLUB SEMINAR

     Speaker:	Rejeev Sharma, The Beckman Institute
                Dept of Electrical & Computer Engineering
	        University of Illinois at Urbana-Champaign

	        Dr. Sharma is a candidate for a tenure track
	        position with us.

	Date:	Wednesday, April 17, 1996 [Note:  Not Monday]

        Time:	3:00 - 4:00 p.m.

       Place:	100 USC

 	    Refreshments:  following the talk

   GEOMETRIC MOTION PLANNING WITH VISUAL CONSTRAINTS

The configuration space (C-space) has been used as a unifying 
computational framework for defining geometric motion planning 
problems in robotics.  However, for a robot operating in an 
uncertain and dynamic environment when sensing plays a critical
role, the traditional C-space approach may be inadequate.  For 
example, to utilize the sensor feedback effectively, the motion 
plans should also account for properties of the sensed data.  
We introduce a general motion planning framework using the
concept of perceptual control manifold (PCM) that augments the 
C-space with a parameter space defined on sensor data.  For 
example, the PCM could involve a set of image features that 
are used for visually controlling a robot.  The framework 
allows the motion plans to incorporate various constraints 
and optimality criteria derived from the robot kinematics, 
the control system, and the sensing mechanism.  In computer
vision, one such constraint is the ability of the camera 
system to perceive the instantaneous motion of an object in 
its field of view.  We formulate this constraint in terms
of the PCM and show how it leads to better motion plans.  
Finally, we discuss related issues including representation 
of the PCM, computational complexity, modeling of various 
uncertainties and active vision.