Kinematic accuracy and self-calibration of an object integrative handling system

• Kinematische Genauigkeit und Selbstkalibrierungeines Objekt-integrativen Handhabungssystems

Detert, Tim; Corves, Burkhard (Thesis advisor); Schmitt, Robert Heinrich (Thesis advisor)

1. Auflage. - Aachen : Apprimus Verlag (2018)
Book, Dissertation / PhD Thesis

In: Apprimus Edition Wissenschaft
Page(s)/Article-Nr.: 1 Online-Ressource (118, xxiii Seiten) : Illustrationen

Dissertation, RWTH Aachen University, 2018

Abstract

The cooperative manipulation of rigid objects with industrial robots leads to redundant actuation, which is a challenging configuration for the control of cooperating robots. In such a configuration, the object is connected to multiple robots and integrated into the kinematic structure, resulting in actuation redundancy. Inaccuracies of the robots and tolerances of the object lead to deviations of the grasp points at the object that cannot be determined precisely. These deviations of the kinematics model from the real system are amplified due to coupling effects. This consequently results in inadequate object positioning and causes internal tensioning of the overall structure. A recent example of this object integration is the PARAGRIP handling system that was investigated in this thesis. Today’s control and calibration approaches, however, do not target the identification of the uncertain grasp points. Accordingly, in this thesis a kinematic calibration procedure for the PARAGRIP robotic arms and a self-calibration procedure for the object integrative handling system were developed to identify the actual grasp points at the object. The kinematic calibration was investigated for a mathematically efficient serial and a hybrid kinematics mod-el, both including the compensation of gravitational effects. The choice of identification points and the optimization residual were investigated and the limited absolute accuracy of the PARAGRIP arms could be improved significantly based on the findings. In the context of the kinematic calibration, a new stiffness modeling approach was implemented by extending the concept of Matrix Structure Analysis. The implemented modeling approach allows for the automatic calculation of arbitrary kinematic structures and the compensation of the gravitational deformations. Furthermore, a new self-calibration method for the object integrative handling system was developed based on the combination of direct and inverse kinematic calculations. The redundant sensor-information of the cooperating robots is evaluated to identify the actual grasp points at the integrated object, to increase the system accuracy and to limit the impact of the coupling effects. The sensitivity and influencing factors of the method were investigated and verified in simulations. The results show that the available redundant sensor information for object integrative robots or handling systems can be used to identify the grasp points at the object and compensate the internal inaccuracies automatically. The application of self-calibration to object integrative handling systems allows for an efficient and easy way of reconfiguration and calibration without additional external metrology. This offers the opportunity to extend the capabilities of cooperating robots and reduces the demand for complex robot control. The re-search and results described in this thesis yielded new findings for the PARAGRIP handling system, which can be generalized for every object integrative handling system with redundant actuation, in particular cooperating industrial robots.

Institutions

• Chair and Institute of Mechanism Theory, Machine Dynamics and Robotics [411910]