Calibration of Serial Manipulators: Theory and Applications

Abstract

All the steps necessary to perform a kinematic calibration of serial manipulators have been discussed and applied to an actual manipulator. This first part of the chapter has compared and extended two procedures for the identification of the geometrical parameters necessary to describe the inaccuracies in the kinematic structure of a generic serial robot . The discussion led to a new methodology called `Modified Incremental' which holds the positive characteristics of the others. Each procedure generates a Minimum, Complete and Parametrically Continuous set of parameters. First of all the formula for the determination of the total number of parameters has been discussed pointing out the distinction between internal and external parameters. The D&H approach for the parameters identification has been extended to deal with all the special cases (adjacent parallel joint axes, prismatic joints and gripper frame). This approach has the good quality to include the joint offsets in the parameter set and shows that some parameters ( l and ) are intrinsic of the link while the others represent the joint motion and the assembly condition. The Incremental approach is the most simple to be applied (only the type of the joints must be considered) and can be more easily automatized. The main drawback is that the joint offsets are not explicitly included in the parameters set. The Modified Incremental approach solves this problem and the proposed procedure for the elimination of the redundant parameters ensures the Minimality of the model. The last part of this chapter fully describes a practical calibration of a measuring robot discussing different estimations algorithms and their performances. All these concepts will be reviewed in the next chapter in order to be applied to a generic parallel manipulator .