A hexapod has fewer moving parts, greater rigidity, and higher accuracy than comparable serial manipulators. Typically CNC machine tools rely on serial mechanisms, where the joints are displaced from the end point of the tool and therefore angular errors multiply by the displacement to cause large errors in the tool position. This reduces accuracy and stiffness. In parallel manipulators, stiffness is increased over conventional manipulators because the loads are divided among all of the links, and accuracy is increased because, over much of the workspace, leg length errors tend to average. Another significant characteristic of the hexapod is that the forces applied to actuators as the result of machining loads are all axial (tension and compression) without any bending moments. In conventional machine tools or coordinate measuring machines, the axes are generally calibrated independently. The results leave many errors unaddressed, but at least the calibration procedure can be done one axis at a time. This is not true in a parallel link manipulator. Recently, IAI discovered a method which they believe will allow the major errors in a hexapod structure to be calibrated using only ball bar data. The scheme is so effective that it can be implemented in real time. If instead of a `fixed length` ball bar, a single tracking interferometer is used which produces only single axis data, then the calibration procedure could be executed continually as the machine operates. The ultimate result of this extension is that researchers could build a machine whose controller is continually measuring the machine's errors to interferometer level accuracies, and dynamically compensating for all the machine's primary errors. With a hexapod-based coordinate measuring or machine tool, all the errors could be corrected by adjusting only the six leg lengths, so a hexapod-based machine tool could be as accurate as the controller's knowledge of the machine's errors. IAI's approach to continuous calibration should also be applicable to conventional machine tools, and they will investigate this possibility as well.

Agency
National Science Foundation (NSF)
Institute
Division of Industrial Innovation and Partnerships (IIP)
Type
Standard Grant (Standard)
Application #
9460521
Program Officer
Kesh S. Narayanan
Project Start
Project End
Budget Start
1995-01-01
Budget End
1996-04-30
Support Year
Fiscal Year
1994
Total Cost
$74,761
Indirect Cost
Name
Intelligent Automation, Inc
Department
Type
DUNS #
City
Rockville
State
MD
Country
United States
Zip Code
20855