Carnegie Mellon University
Advanced Search   
  Look in
       Name    Email
       Former RI Members 
Peter Rander
Principal NREC Commercialization Specialist, RI
Office: NREC 281
Phone: (412) 681-3466
Fax: 412-681-6961
  Mailing address:
National Robotics Engineering Center
10 40th Street
Pittsburgh, PA 15201
Affiliated Center(s):
 National Robotics Engineering Center (NREC)
Research Interests

My research interests revolve around computer vision and robot perception. The core theme of my work is the exploration of 3D perception and modeling, including stereo vision, range data processing, and sensor fusion. The boundary between vision and graphics has been a driving force behind much of this work, and has also led me into image based modeling and rendering. More recently, and especially on the PerceptOR project, I have been exploring the interaction of perception, planning, and action in offroad mobile robotics.

My projects span these research areas, from mobile robot perception to Virtualized Reality dynamic event modeling and rendering. For the last several years, I have explored many mobile robot perception problems on the PerceptOR program. My early work on this effort included 3D perception for hazard detection from a low-altitude helicopter flying ahead of a ground vehicle, as well as terrain classification based on geometric and visual cues. More recently, my role has expanded to include day-to-day technical direction of the entire program -- active and passive perception, planning, pose estimation, control, remote operator interfaces, and coordination of unrehearsed field experiments.

Previously, I worked on 2D vision problems for a robotic paint stripping system (M2000) that removes paint from ship hulls. I have also been exploring 3D perception for robotic paint stripping of complex 3D surfaces. I have collaborated with the Virtualized Reality project, which explores 4D event capture, playback, and manipulation. This project, part of the inspiration for the CBS Eye Vision system that debuted recently at the 2001 Super Bowl, seeks to use an extremely large number of video cameras to capture the space-time (4D) appearance of events over time. The resulting 4D models can be used not only for immersive playback, but potentially for real-time collaboration, training, and a host of other applications.

Research Interest Keywords
3-D perceptioncomputer graphicscomputer visiondata visualizationimage processingrange finderssensor fusionstereo visionvideo systems