Models for Diagnosing Robot Error Sources

Accuracy of a machine is determined by its ability to position and orient a tool at a prescribed point. Usually this positioning is performed by positioning joints (actuators The actuator positions are comsured. The tool position is computed from knowledge of actuator position through some equation. It is impossible for machines to position tools with zero error. The error is introduced from a number of sources. Some of these error sources are warped, or bent components, inaccurate actuator position sensors, or even external (sometimes internal) loads. In the case of industrial robots and on some machine tools, calibration methods have significantly improved position accuracy. At the heart of most of these calibration methods is an error model. This paper discusses the idea of using error models to predict the sources of the error in a machine. It is shown that the somewhat ad-hoc modelling methods used for robot calibration, although satisfactory for improving accuracy, are not useful for detecting the true causes of the error. monly measured 13. ut the actual tool position is unmea-