Industrial robots as a reference: Static calibration of inertia measurement systems

This paper deals with a calibration method for inertial measurement systems by means of a conventional industrial robot. By dividing the calibration procedure up into three steps, whereby the first one only deals with the vector norm of the three-dimensional measurements, it is possible to identify offsets and scale factors of the sensors as well as nonorthogonalities independently of the robot's positioning errors. Focus is set on a third step in which the robot's end effector drives a circular path to generate reference rates. Taking advantage of the special structure of the equations for circular paths, it is possible to formulate the problem as a mixed-linear parameter estimation in which the dimension of the optimization problem can be reduced by the number of the linear parameters. Thereby, the convergence speed of the nonlinear optimization algorithm is improved.