Nongeometric error identification and compensation for robotic system by inverse calibration

In order to achieve the stringent accuracy requirement of some robotic applications such as robotic measurement systems, it is critical to compensate for nongeometric errors such as compliance errors and thermal errors in addition to geometric errors. This paper investigates the effect of geometric errors, link compliance and temperature variation on robot positioning accuracy. A comprehensive error model is derived for combining geometric errors, position-dependent compliance errors and time-variant thermal errors. A general methodology is developed to identify these errors simultaneously. A laser tracker is applied to calibrate these errors by an inverse calibration method. Robot geometric errors and compliance errors are calibrated at room temperature while robot parameter thermal errors are calibrated at different temperatures when the robot warms up and cools down. Empirical thermal error models are established using orthogonal regression methods to correlate robot parameter thermal errors with the corresponding temperature field. These models can be built into the controller and used to compensate for quasi-static thermal errors due to internal and external heat sources. Experimental results show that the robot accuracy is improved by an order of magnitude after calibration. (C) 2000 Elsevier Science Ltd. All rights reserved.