Robot calibration
Robot Calibration refers to the determination of the various parameters of the robot arm in order to obtain a complete kinematic model of the robot. The calibration of robots, tool and workpiece ( cell calibration ) can reduce or minimize the existing inaccuracies. By calibrating the process reliability can be increased.
The positioning accuracy of industrial robots ( IR) is often inadequate for certain tasks. In particular, the robots exchange and during the programming of precision applications problems may occur, their removal can be very time -consuming and costly.
Parameters and error sources
The international standard ISO 9283 defines various performance criteria for IR and propose test methods for their determination before. The most important characteristics are the absolute accuracy ( accuracy of pose or path, AP ) and the repeatability ( repeatability of pose and path, RP) dar. These are the commonly used criteria.
The repeatability is critical not only in programming the robot by Teach ( " teach-in "), but each time demanding process. This is so regardless of whether the robot only " teached " is, or whether the data was by " offline programming " is generated.
If the exercise program is, however, created by 3D simulation ( " offline - programming" ), is also the absolute accuracy of an IR of importance. It is generally adversely affected by various factors. In this case, the Achsnulllagen well as the linear and angular errors between the individual robot links is the most importance. This in conjunction with the variable load on the flange of an IR, are the largest sources of error dar. But also changes in length due to temperature fluctuations contribute to the change in position of a robot who can afford depending on the robot program and the machine heats to contribute significantly to the repeatability.
Measuring systems
There are several possibilities of the position measurement in industrial robots, for example the start of sample workpieces, the use of ultrasonic sensors, laser interferometry, theodolites, probes or laser triangulation. Also, there is camera systems may be mounted in the robot cell and the IR itself and detecting the position of a reference object. Provider of measuring systems are, for example, the company Perceptron, NDI, HGV Vosseler, Dynalog, FARO, Leica, Metris, Wiest and Teconsult.
Mathematical Foundations
The robot error detected by position measurement can be minimized with numerical optimization in the context of a compensation calculation. For this, a complete kinematic model of the geometric structure must be created, whose parameters are then determined by mathematical optimization. So can be calculated from the input and output variables in vector notation, the general behavior with the vectorial model function formulated as follows:
The variables k, l, m, n, and their linkages describe the dimensions of the individual vector spaces. The minimization of the residual error R for the identification of an optimal parameter vector p follows from the difference of the two output vectors using Euclidean norm.
For the solution of the kinematic optimization problems are, among other things, Least Squares descent method, for example a modified quasi -Newton method. This method provides corrected kinematic parameters for the measured engine, which can then be added, for example, in the robot controller to adapt the mathematical model to the real kinematic used therein.
Results
The absolute positioning accuracy of industrial robots varies depending on the manufacturer, the age and wear anywhere from a few tenths to several millimeters. By calibration can usually be achieved a positioning accuracy of approximately 0.5 mm, which can come close in limiting the volume of work also to the usual repeatability of a robot of about 0.1 mm.
Application Examples
In the industry, there is currently a general trend towards the substitution of tools or special equipment by industrial robots for specific manufacturing tasks, their accuracy requirements can be met by calibrated robot. In the figure, a current example is presented: in-line metrology in the body shop, where previously used for 100 % inspection "Measuring Tunnel" with many expensive sensors partly be replaced by IR, each carrying only one sensor. Thus, the total cost of a measuring cell can be significantly reduced. In addition, the system is reusable when you switch models by simple reprogramming without structural changes.
Further examples for precision applications are robotic roller hemming in the body shop, assembly of mobile phones, drilling, riveting and milling in aircraft and increasingly medical applications.
Summary
Through the use of efficient calibration methods, it is possible with currently commercially available industrial robots - especially parallel kinematic - to achieve an absolute positioning accuracy of 0.1 mm, so as to improve the interchangeability to simplify the off- line programming and new, high-precision applications to enable.