Calibration, modeling and sensor guided approaches are proposed to improve robot workcell accuracy to achieve effective offline programming and precise contact operations. Challenges in the proposed solution include:
- Implementation of calibration results in closed architecture industrial robots,
- Non-geometric error sources that are difficult to model, identify and compensate, and
- Sensor-based localization of free form work objects with geometric uncertainties.
Some of the research results include:
- A model-based calibration with online compensation was implemented in a ABB robot and achieved an average error of less than 0.3mm.
- A calibration and compensation framework uses non-model based approach to improve local accuracy and to provide measurement results for the outer loop model-based calibration to improve model and subsequently global accuracy.