Teradyne Robotics is bringing new AI capabilities to automation applications by equipping its platforms with the latest microchips from NVIDIA, the chipmaker whose processors have become a critical ingredient to drive generative artificial intelligence (AI) products over the past year.
Massachusetts-based Teradyne will use the technology in its Universal Robots division, which makes picking arms for collaborative robot (cobot) applications and in its MiR division, which makes autonomous mobile robots (AMRs).
Specifically, UR has integrated NVIDIA accelerated computing into its UR5e cobot for faster planning, making path planning 50-80x faster than today’s solutions and increasing application potential and efficiency for automation customers. And MiR will use the technology in its new pallet handling product, the MiR1200 Pallet Jack, which uses 3D vision to identify, pick up, and deliver pallets with tight precision, even in dynamic and complex environments like autonomous material handling workflows at factories and warehouses.
While Teradyne has been working with NVIDIA chips for years, the upgrade to its current generation of advanced graphics chips will serve to “future proof” its robotics roadmap, Kevin Dumas, MiR’s VP of Products, said in an interview.
“We could have custom built solutions ourselves or gone with another provider, but [with NVIDIA] we’re packaging their [central processing unit] and also their [graphics processing unit], since the math and computations for doing graphics turned out to be perfect for AI and machine learning, to handle the high bandwidth of images we use to run our products,” Dumas said.
The robots need that muscular processing power because they navigate with an array of sensors including two-dimensional LIDAR, three-dimensional LIDAR, and five three-dimensional computer vision cameras.
By combining a CPU and a GPU, MiR’s robotic pallet jack can process that data in real time to detect what a pallet is and where it is in relation to the robot and in space, so the AMR can drive to that location and use its forks to lift a loaded pallet off the ground, Dumas said. That enables a different workflow from the current generation of MiR robots, which handle pallets that are placed on an elevated stand, instead of directly on the warehouse floor. The new approach opens up a wider range of use cases while preserving customers’ top needs of reliability, predictability, and repeatability, he said.
