Skip to main navigation Skip to search Skip to main content

A Two-Stage Calibration Method for Industrial Robots with Joint Flexibilities

  • Matthias Neubauer (Speaker)

Activity: Talk or presentationContributed talkunknown

Description

One of the main characteristics of industrial robots is their positioning accuracy, strongly depending on geometric parameters. Thus the calculation of this parameters, called geometric robot calibration is crucial. However, the calibration is often treated solely as a kinematic problem [1], neglecting effects due to robot flexibilities. In this paper, the identification of a reliable kinematic model, eliminating this systematic error by including elastic deflections in the calibration, is presented. The procedure is divided into two stages. First, a frequency domain approach to calculate the stiffness and damping parameters is used. Therefore periodic motor torque excitations are applied, in order to obtain nonparametric frequency responses. By an optimization, the match between the parametric transfer function of the linearized system with the measured nonparametric one is maximized leading to proper stiffness and damping parameters. In the second step, geometric error parameters pe like length errors and joint offsets are introduced and the forward kinematics is derived with this parameters. The calibration result strongly depends on the robot poses. Thus a pose optimization algorithm is used. For these poses, elastic deflections according to the identified model are evaluated and included in the kinematics and thus in the calculation of pe. The calibration is carried out for a six-axes articulated robot (St¨aubli TX90L) and the achieved static positioning accuracy is presented.
Period09 Feb 2015
Event titleEuroCAST 2015
Event typeConference
LocationSpainShow on map

Fields of science

  • 202035 Robotics
  • 202 Electrical Engineering, Electronics, Information Engineering
  • 203022 Technical mechanics
  • 203013 Mechanical engineering
  • 203015 Mechatronics

JKU Focus areas

  • Mechatronics and Information Processing