We have developed blade control algorithms for maximizing the material removal rate of automated dozers. For preliminary testing, we have created an integrated experimental environment with an instrumented scaled-down robot bulldozer, a vision-based robot localization and steering system, and a laser scanning system for measuring the height profile of the material pile. This video shows one automated dozing pass.
Demonstrates our prototype robotic surgical instrument. Unlike cable-based designs, this surgical robot is internally
actuated. This enables increased dexterity for robot-assisted surgery, and the potential for "plug and play" modularity.
Summarizes three of our computer vision (AKA machine vision, or robot vision) projects, specifically: rapid automated surface inspection,
vision-guided grasping of objects with unknown 2-D position and orientation, and automated online 3-D modelling and grasping of objects with unknown location and shape.
Demonstrates a machine we developed for a local company which quickly
inspects the bearing surfaces on automotive rocker arms using machine
vision; and then sorts them into reject/accept bins. (1.3 MB AVI
file using DIVX compression, the CODEC is available here
or from www.divx.com):
Demonstrates the machine vision-guided assembly of four automotive
components. (2 MB AVI file using DIVX compression, the CODEC is available
here or from www.divx.com):
Shorter clip showing a 2 degrees-of-freedom servo gripper accurately
locating and holding a part which has been roughly placed on a table.
1.1 MB MPEG file:
Longer clip showing a 2 degrees-of-freedom servo gripper reconfiguring
and picking up two distinctly shaped automotive body parts. 2.4 MB
MPEG file: