High Res TIFF (14.6 MB)

The Tactical Mobile Robot called "Urbie"

One of the important technologies that must be developed for any sort of free-roaming robot is stereoscopic vision. Otherwise the robot will not be able to accurately and quickly navigate toward its target, seeing and avoiding obstacles along the way. Real-time stereoscopic "machine vision" is one of the advanced technologies being developed and demonstrated for the "Urbie" Tactical Mobile Robot.

In the past, robots used a single camera as its "eye" and human operators used the picture from that one camera to "see" obstacles and move the robot around to avoid them. This approach doesn't work too well, though, when the robot is moving quickly or in a hazardous situation. There just isn't time for a human operator to analyze the image and react in time with commands to the robot. Also, for exploration of other planets, moons, asteroids and comets, the robot must be autonomous. It takes minutes to hours for an image signal from a robot on a distant body to reach a human operator on Earth and for a command to be sent back to the robot. By that time, the robot could have fallen into a deep hole or gotten itself stuck between a rock and a hard place!

Urbie's initial purpose is mobile military reconnaissance in city terrain. However, many of its features will also make it useful to police, emergency and rescue personnel. The robot is rugged and well-suited for hostile environments and its autonomy will make Urbie ideal for working in dangerous situations. Such robots could investigate urban environments contaminated with radiation, biological warfare, or chemical spills. And, of course, such a robot will make an ideal space explorer.

The Defense Advanced Research Projects Association (DARPA) has enlisted JPL's Machine Vision Group in leading the design and implementation of its urban robot project, a Tactical Mobile Robot. This urban robot (Urbie) is a joint effort of JPL, IS Robotics, the Robotics Institute of Carnegie Mellon University, Oak Ridge National Laboratory, and the University of Southern California Robotics Research Laboratory.

Read more about it: http://robotics.jpl.nasa.gov/tasks/tmr