Project Title:

Robotic Guidance Systems Using Specialized and Generalized Targets

05.09-2567
901781
Robotic Guidance Systems Using Specialized and Generalized Targets

Abstract:

Robotic guidance requires sensing and control in three dimensions. This can be accomplished
through the use of external sensors which measure robot location relative to target
object locations. A robot guidance system will be developed based on hybrid laser
illumination and video sensors which can use both simplified targets for object location
and identification, and more complex three-dimensional geometrical feature-based
recognition schemes as well (at the cost of greater computational complexity). Many
designed objects like ORUs, tools tailored for robotic use, and satellite docking
fixtures can easily incorporate docking targets and object coding tags (like bar
codes). These recognition "fixtures" offer the advantage of rapid reliable recognition
(even in a vision system that can handle more complex recognition tasks) and are
analogous to specially designed grapple points which offer easy sure robotic grip
points. While designed targets are good for manipulation and recognition of designed
objects, untagged geometrical feature-based recognition and measurement must be supported
as well. This requirement is necessary for grasping untagged objects, manipulating
damaged objects, or aligning objects so that particular features (such as holes,
sides, corners, etc.) mate properly.
Space, industrial, and military applications of autonomous robotics would be cost
effective through use of more robust vision guidance systems and object tagging schemes
that support sensorized robot pick and place.
CAD-based computer vision, object identification/location sensorized robotics, bar
codes, holograms, computer vision