The Role of Machine Sensing in CIM

Ayres, R.U. & Funk, J.L. (1989). The Role of Machine Sensing in CIM. IIASA Research Report (Reprint). IIASA, Laxenburg, Austria: RR-89-013. Reprinted from Robotics and Computer-Integrated Manufacturing, 5(1):53-71 [1989]

Preview

Text RR-89-13.pdf - Published Version Available under License Creative Commons Attribution. Download (9MB) | Preview

Abstract

Computer-integrated manufacturing (CIM) is still an objective for the future rather than a present reality. Mechanical integration was introduced over 60 years ago by Ford to increase output and cut unit costs, but at the cost of making high volume production extremely inflexible. The answer is apparently "computer integration" with multi-purpose machines linked together by digital communications networks and adaptively controlled by computers.

In practice this means replacing flexible human workers with high quality sensory interpretative abilities as decision-makers by "smart sensors" with artificial intelligence. However, adaptive controls are severely constrained by the capabilities of existing sensors and interpretative computer software, especially the latter. Most existing sensors are narrow-band, producing at most a few bits of data per second for control purposes. This provides enough information for a small class of machine control decisions, but is insufficient for part recognition, part orientation, or quality inspection.

CIM also means that workpieces (components and work-in-progress) will have to communicate with machines, as machines will have to communicate with each other. Thus, the true economic significance of recent breakthroughs in machine vision/taction is that they will finally unlock the door to CIM, or "5th generation" automation. It is argued that the economics of machine vision/taction should not be assessed in the narrow context of specific tasks in direct competition with human workers but as the hitherto missing link that will permit all the elements of the factory of the future to communicate efficiently with each other so as to function as an organism rather than as a set of independent cells.

Actions (login required)

View Item