Kane Robotics’ GRIT Vision System is customizable for robotic applications from weld grinding to polishing of aerospace parts.
AUSTIN, Texas—Kane Robotics is harnessing the transformative power of artificial intelligence (AI) with visual sensors that replicate the human eye, assessing uneven surfaces and enabling its GRIT™ collaborative robot (cobot) to make real-time adjustments in material removal operations, the company said in a release.
Kane’s new GRIT Vision System is reported to expertly handle manufacturing tasks such as sanding, grinding, and finishing, thanks to a high-speed camera and AI-driven computer vision integrated with Kane’s GRIT cobot.
“Thanks to AI, we can now teach a cobot to flexibly respond to changes in surfaces based on what it ‘sees’ through a camera attachment,” said Arlo Caine, Kane’s robotics engineer who spearheaded the GRIT Vision System, in the release. “The cobot world hasn’t seen this before.”
A human operator sets up the cobot and attaches the necessary tools for the weld-grinding process, then monitors and adjusts the process through a live custom interface. Next, artificial intelligence takes over. Using the camera, the vision system “sees” uneven surfaces on a weld or seam. Thanks to rapidly-learning AI, the system alerts the cobot’s robotic arm and steers it accordingly.
The GRIT Vision System was first introduced at the Paul Mueller Company, a Missouri-based global manufacturer of stainless-steel tanks and equipment. The company needed a solution for seamlessly grinding welds on large steel tanks used in the dairy, food, brewery, beverage, pure water, and pharmaceutical industries. GRIT’s AI software performs live object detection, ensuring accurate tracking even as the weld seam disappears during rotation. When faced with imperfect welds, the system proved it could be taught to detect variations and refine its grinding accuracy.
At the Paul Mueller Company, the GRIT cobot is said to work in a “perfectly collaborative” manner. It does the tedious and strenuous work of holding the grinder, and the vision system manages the tiring task of tracking large seams for long periods.
Finally, a human operator sets and adjusts the force level most suitable to the chosen abrasive, grinding speed, and number of grind passes to achieve the required finish.
Although originally developed for weld grinding, the GRIT Vision System is compatible with various robotic components and customizable for diverse applications, from polishing aerospace parts to sanding furniture, Kane said in the release.
“The GRIT Vision System is a breakout for collaborative robotics,” said Kane Robotics CEO John Spruce, in a statement. “Kane was among the first to successfully integrate robots, tools, cameras, and AI software into a cohesive product. Today, that means more efficient, precise, and safe material removal in manufacturing. The future possibilities are limitless.”