Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

A vision system for robotic inspection and manipulation

Conference ·
OSTI ID:5294094
Advanced robotic systems should be able to perform a variety of tasks in complex, unstructured environments with increased level of autonomy. Robots provide the physical link between intelligence and action. Structurally, a robot can be considered as having three modules. These include, (1) Mechanical assemblies, such as robot arm, end effectors, and mobility platforms, (2) Sensors, for sensing the work environment of a robot, and (3) the Perception, Planning and Control unit, which is utilized for interpreting the sensory inputs and for planning and controlling the actions of the robot. Most of the robotic systems currently employed in the industry require a highly structured environment for the robot to operate. This requirement can be relaxed if the robot is endowed with an array of external sensors to sense its environment. The sensor-driven operation is critical for making robots more versatile and flexible to use. Advanced robotic systems which are capable of utilizing sensor modalities such as vision, range, force, and touch can be employed in a variety of application domains. Of these, vision is recognized to be a very important sensory modality. It offers rich sensory data for accurate and detailed interpretation of the composition of a robot's work environment. Object recognition, determination of their locations in the 3-dimensional workspace, inspection of their status are all important tasks where vision derived information can be effectively utilized. 10 refs., 6 figs., 3 tabs.
Research Organization:
Tennessee Univ., Knoxville, TN (USA). Dept. of Electrical and Computer Engineering
Sponsoring Organization:
DOE/NE
DOE Contract Number:
FG02-86NE37968
OSTI ID:
5294094
Report Number(s):
DOE/NE/37968-11; CONF-890143--3; ON: DE90005412
Country of Publication:
United States
Language:
English