Content area
Full Text
Charles S. Vann: Deputy Program Leader for the Laser Science and Technology Program, Lawrence Livermore National Laboratory, Livermore, CA 94550, USA. Tel: +1 510 423 8201; Fax: +1 510 422 1930; e-mail: [email protected]
ACKNOWLEDGMENT: The views represented are those of the author and do not represent those of Lawrence Livermore National Laboratory, the University of California or the United States Government.
The problem with robots in manufacturing today is that they are flexible but unadaptable machines in comparison to humans. Robots have the physical agility to perform many human-like tasks, but even the most advanced manufacturing robots have almost no human-like ability to interpret their environment and make intelligent decisions on their own. Up to now, we have been forced to treat robots as dumb machines, training them to perform each new task, and requiring them to totally rely on exact positioning of the robot and part. If a new part or tool is introduced, the robot must be retrained - a major cost of retooling a plant. Worse yet, simple process anomalies such as part misplacement, contamination, wear, slack in the robot arm, or temperature changes can lower product quality, damage robots, stop production lines, or even injure workers. The bottom line is that robots need to be smarter to be more productive.
One main obstacle to smarter robots is their inability to sense their environment as humans do. Humans sense displacement and orientation of objects, i.e. all six degrees of freedom (six-DOF), enabling us to perform complex and intelligent tasks so effortlessly that we take for granted how complex those tasks are for a robot to duplicate. For example, drinking from a can, eating with a fork, or getting dressed are everyday six-DOF tasks. These seemingly...