At Danfoss, a Danish manufacturer of mobile hydraulics as well as electronic and electrical components, robotic systems equipped with 3D sensors now reliably and precisely recognize and move a large number of different components.
In the new production line implemented by Danish integrator Quality Robot Systems (QRS), a total of six KUKA robots take over work steps that were previously performed manually. For this, the robotic cells must recognize and move up to 100 different components without manual intervention.
The Challenge: Handling a Large Variety of Parts
“The requirement for the robotic cells at Danfoss was fully automatic feeding of the components – a task that can only be solved with a camera system, given the large number of variations,” explains Olav K. Sjørslev, Sales Manager at KUKA integration partner QRS.
The alternative, namely feeding the blanks meticulously ordered and aligned, was simply not economically feasible.
Instead, the objects arrive at the first robot station in standard roll boxes, sorted by box, but with the boxes stacked on top of each other more or less indiscriminately, and with the layers of objects separated by paper sheets.
The Solution: A Vision System That Works That Really Works
“Our experience with camera systems in the past was not particularly good,” recalls Sjørslev. “We looked into various systems over the past decade, but did not find a single 3D vision solution that met our needs. What works at a trade fair or in a manufacturer’s demonstration room is often not viable under real-world conditions.”
To meet the challenges of object recognition, QRS was able to benefit from its close partnership with KUKA. For when it comes to camera systems and object recognition, KUKA turns to Munich-based partner Roboception and their rc_visard 3D stereo sensor.
3D Vision is Key To Success
At Danfoss, the rc_visard, using the SilhouetteMatch software, reliably detects the position and orientation of the flat, unordered objects on the flat paper surface and also provides the robot with the gripping points at the same time.
A 2D solution was quickly ruled out for this project, as the paper slipsheets are not always completely flat, or may be oily due to the components. In tests with other systems, the camera had sometimes misinterpreted impressions as workpieces. SilhouetteMatch, on the other hand, recognizes exactly which components are supplied to the robotic cell and reliably transmits the appropriate gripping points for a KR CYBERTECH from KUKA.
“Despite the major challenge of automatically recognizing more than 100 components, we were able to solve the task in a highly satisfactory Roboception manner – and provide our customer with the perfect solution“ comments Olav K. Sjørslev, Sales Manager at QRS.
A Synergetic Coupling of Experience
For those involved, the project was very much a case of jumping in at the deep end. “Implementation of the object recognition was a new experience for us. When we venture to take on something like this, we only select partners of whom we are absolutely convinced and who we are sure will support us in the long term,” emphasizes Olav Sjørslev. He has only positive things to report about the cooperation with Roboception and KUKA.
“For this project, everyone involved willingly went the extra mile in terms of transfer of know-how and time invested.“
The effort was worth it in many respects. In this way, not only was the client satisfied, but the successful integration of the camera system from Roboception also ensured follow-up orders at QRS.
For more information: www.roboception.com