Putting the Turing into Manufacturing: Recent Developments in the Science of Automation

Manufacturing, which includes everything from electronics to industrial equipment, represents a major sector of the American economy. In 2006, durable-goods manufacturing accounted for 6.9 percent of the GDP whereas all information-communications-technology industries comprised 3.9 percent.

Yet manufacturing today is where computer technology was in the early 1960’s, a patchwork of ad-hoc solutions lacking a rigorous scientific methodology. Computer Aided Design has been widely adopted for modeling of mechanical parts and behavior. What is missing is an associated science for Manufacturing: the required handling, assembly, inspection, and storage of these parts, the robots and tooling for assembling of these parts into products through networks to reach customers in a timely and quality manner. What are the models of manufacturing analogous to the models of computing that Alan Turing invented to establish the mathematical and scientific foundations for computer science?

In short: it’s time to put the Turing into Manufacturing. A science of manufacturing will require mathematical and algorithmic abstractions for basic elements of robotics and automation such as part feeding and fixturing and production systems. Abstractions allow functionality to be specified independent of hardware and software implementations,which in turn provides the foundation for formal specification, algorithmic design, consistency checking and optimization. Abstraction facilitate the integrity, reliability, interoperability, and maintainability of automation, and streamlines upgrading as new technology and theory becomes available. Such research would have broader impacts into laboratory automation, agriculture, and health care.

Over the past decade, researchers have made progress, developing a variety of algorithmic models and results for part feeding and fixturing. I’ll review selected results from my lab and others, including a new framework for fixturing deformable parts and new geometric primitives for vibratory bowl feeders, and propose several open problems for future research.

I’ll also describe the Berkeley Center for New Media and its current cross-disciplinary research projects.

Presenter(s)

Ken Goldberg is Professor of IEOR, EECS, and the iSchool at UC Berkeley and Director of the Berkeley Center for New Media. He is Vice-President of Technical Activities for the IEEE Robotics and Automation Society. His research addresses robot manipulation, geometric algorithms for automation, and networked robots. More information on his work, art, and projects are available at: http://goldberg.berkeley.edu/.