Motivation

Industrial robots are gaining interest due to their versatility in many areas of industry as well as in measurement technology. A large working area, high payloads and low investment costs are essential factors that contribute to increasing the productivity and quality of goods. However, the main disadvantage of using industrial robots is their accuracy. Many industrial applications require accuracies in the sub-millimetre range. So that these applications can also be implemented, measures to increase accuracy are absolutely necessary. The accuracy can be decisively influenced by determining the robot’s position in relation to its workpiece and the dimensions of a mounted tool. Various strategies are used for this, which are more or less suitable depending on the task area and thus the accuracy requirements for the calibrated industrial robot. Therefore, as part of a student work, an investigation of economic concepts for the simultaneous determination of the robot and tool position was carried out.

Activities

Mainz University of Applied Sciences has owned several robots since 2016, with the opening of the ROBOLAB laboratory for applied robotics, funded by the Carl Zeiss Foundation “Invest”. The robot on which the investigations are based is assigned to the “digital production” area and is used to produce various and innovative spatial structures (KUKA KR60 HA with electro-spindle from HSD Mechatronics). The core of the investigations is the further development of measurement concepts with which the robot base (transformation from the robot position to the workpiece) and robot tool (transformation from the robot end point to the processing point of an external tool) can be determined most economically and the results obtained can be reliably checked. On the one hand, the influence of different measurement configurations was examined. Furthermore, new determination methods were tested, strategies for checking the results obtained were developed and the influence of different tool weights on the base determination was simulated.

Results

The 10-point method was optimized. For this purpose, various configurations for the control points for solving the base and the tool were tested using the OpenIndy plug-in “RoboCalc”. Based on the results, the existing measurement concept was supplemented and also validated. This resulted in interesting insights into the number of points and the distribution of points. The simultaneously obtained results from base and tool were compared with reference data accurate to a few tenths of a millimeter. A good base control workflow was also found.

Furthermore, a new strategy for determining the base was developed in the course of the measurements. Despite good feasibility and accuracy, their main disadvantage is the need for a specific tool. The specification of a simultaneous determination could therefore not be met. However, the methodology represents a good alternative to other methods discussed, which also only determine the base.

The investigation of the influence of different tool weights on the base showed that serious deviations result if the correct tool data is not observed and that determining and entering this is therefore to be recommended.