Significance of observation strategy on the design of robot calibration experiments

This article examines the factors influencing the identification and observability of kinematic parameters during robot calibration. A generalized calibration experiment has been simulated using two different identification techniques. Details of the identification techniques and considerations for implementing them using standard IMSL routines are presented. The factors considered during the simulations include: initial estimates of parameters, measurement accuracy and noise, encoder resolution and uncertainty, selection of measurement configurations, number of measurements, and range of motion of the joints during observations. Results are tabulated for the various cases and suggestions are made for the design of robot calibration experiments.