Kane Robotics Inc. has integrated artificial intelligence with visual sensors to make its collaborative robot automatically track and grind weld seams with great precision and speed. The company said its GRIT vision system applies computer vision to the GRIT cobot to perform material removal tasks such as sanding, grinding and polishing in manufacturing.
Launched in 2023, the GRIT cobot works in collaboration with humans to achieve labor-intensive finishes, adapting to any size and type of manufacturer, Kane Robotics said in a statement. Although it was originally designed for material removal in the aerospace industry, the company explained that the robot can be configured to work with metals, woodworking and other types of manufacturing.
It is important to mention that Kane Robotics highlighted the case of Paul Mueller Co., which was looking for a more efficient way to grind welds on large steel tanks. This stainless steel equipment manufacturer also aimed to reduce fatigue-related injuries and improve working conditions.
The GRIT vision system demonstrated its capabilities in moving object detection and adaptive recalibration, allowing Paul Mueller to polish tank shells of different sizes and various types of weld beads , Kane Robotics said.
How does the GRIT Vision System work?
Dr. Arlo Caine, a consulting engineer at Kane Robotics and a member of the company’s advisory board, played a crucial role in creating the GRIT vision system.
This system is made up of a camera integrated into Kane’s cobot arm and patented artificial intelligence software. The AI uses the images captured by the camera to “think” while guiding the cobot to follow the weld seam. In situations where weld seams exhibit imperfections, AI auto-steering identifies the uneven pattern and adjusts the robotic arm accordingly .
Kane engineers train artificial intelligence to recognize various welds before deployment. Through software upgrades, the vision system is refined to identify variations in welds, thereby improving grinding accuracy.
This methodology differs from other welding cobot systems, as the Kane AI vision system “visualizes” the path that the grinding tool must follow , even when the weld seam is no longer visible. Unlike most vision systems that detect and react to objects, in this case objects, such as weld seams, can disappear.
Likewise, the Kane vision system addresses this challenge by learning the different phases of each specific weld grinding process. In this way, the GRIT cobot can recognize when it should continue or stop the grinding process. This approach represents an unprecedented innovation in the field of collaborative robots.
The collaborative robot and humans
Vision systems use machine learning to adjust cobot movements in near real time. However, since manufactured parts are rarely exact copies of their ideal CAD models and the cobot’s movements must be precise, human judgment is still required. Human operators are responsible for determining the best way to perform weld polishing.
These human operators make decisions about tool placement, abrasive choice, and when finishing specifications must be met. Once the GRIT vision system takes over control of the cobot, a custom live interface allows the human operator to evaluate and make adjustments as necessary.
It is essential to highlight that the system operates in a truly collaborative way, since the cobot takes care of the tedious and monotonous tasks, while the expert operator shapes the final finishing result.
Don’t miss any of our posts and follow us on social media!