Teradyne Robotics, parent company of Universal Robots and Mobile Industrial Robots (MiR), is entering into a partnership with NVIDIA. The company is integrating AI functionalities based on NVIDIA technology into its automation solutions. A concrete example is the new MiR1200 autonomous pallet truck, which uses AI-based 3D vision technology to identify pallets. In addition, consider Universal Robots' cuMotion planner, which can create routes up to 80 times faster thanks to AI.
NVIDIA is a US manufacturer of computer hardware. The company specialises in video cards and chipsets for parent companies. NVIDIA is benefiting from the popularity and rapid embrace of AI. GPUs play an important role in training many AI models. NVIDIA is responding to this, among other things, with graphics cards specifically designed for AI applications.
Plan routes up to 80 times faster
Through the partnership, Teradyne Robotics aims to integrate AI technology from NVIDIA into its autonomous mobile robots. In doing so, the company is betting on "physical AI solutions". Universal Robots demonstrated a concrete example at the NVIDIA GTC AI conference in San Jose, California. Here, the robot company showed an AI-supported autonomous inspection solution: cuMotion path planner.
cuMotion path planner features an NVIDIA Jetson AGX Orin module, hardware for running autonomous systems. This is a system-on-module designed specifically for AI applications at the edge of the network.
Increase productivity of robots
Universal Robots combines the Jetson AGC Orin module with its UR5e platform for collaborative robots and its PolyScope X software. PolyScope X is a software platform that helps manufacturers increase robot productivity. Among other things, the software speeds up the programming of specific tasks and simplifies the integration of collaborative robots into a variety of tasks. With the combination, Universal Robots aims, among other things, to support new applications whose full automation was not feasible until now.
Incidentally, seven types of Jetson modules are available. The entry-level model is the Jetson Orin Nano, whilst the Jetson AGX Orin modules used by Teradyne are the top-of-the-range models. By way of comparison: whilst the 4GB Jetson Orin Nano offers 20 TOPS (trillion operations per second) of performance, the 64GB AGX Orin delivers around 275 TOPS. The AGX Orin 64GB is therefore almost 14 times as powerful as the Orin Nano 4GB. More information about the modules can be found here.
MiR1200 Pallet Jack
Another product using NVIDIA's AI technology is the MiR1200 Pallet Jack. Like Universal Robots' cuMotion path planner, this robot uses the NVIDIA Jetson AGX Orin. This module supports integrated 3D vision technology, allowing the autonomous pallet jack to identify, pick and deliver pallets with greater accuracy.
The MiR1200 Pallet Jack can interoperate with existing MiR AMRs. With this, MiR aims to provide a one-stop shop for material handling for large companies. They typically work with more complex workflows, which can involve larger fleets of AMRs and multiple locations.
Customise routes dynamically
Robot's 3D vision technology to enable automation in complex environments. MiR is thus responding to the labour shortage in many industrial sectors, for which robots can provide a solution. Among other things, the MiR1200 Pallet Jack is capable of dynamically adjusting its routes. This allows the robot to avoid obstacles in its path, such as loose objects or people.
To make this possible, the robot is equipped with multiple robots and LIDAR sensors. The data they generate is processed by the robot in real time using the GPU and CPUs of the NVIDIA Jetson AGX Orin module.
Author: Wouter Hoeffnagel
Image: Teradyne Robotics
