With Omniverse, a cobot developer from Taiwan speeds up robotics-based inspection by 20% to raise product quality in the electronics manufacturing industry.
How to improve robots? By reenacting considerably more robots.
Jensen Huang, the CEO, and founder of NVIDIA, demonstrated today how Quanta, a major electronics manufacturer, checks the quality of its products with AI-enabled robots.
Huang discussed the digitalization of cutting-edge factories by electronics manufacturers in his keynote address this week at the COMPUTEX trade show in Taipei.
For instance, robots from Quanta auxiliary Techman Robot tapped NVIDIA Isaac Sim — a mechanical technology reproduction application based on NVIDIA Omniverse — to foster a custom computerized twin application to further develop a review on the Taiwan-based hardware supplier’s assembling line.
The demo shows how Techman utilizes Isaac Sim to enhance the examination of robots by robots on the assembly line. Basically, it’s robots building robots.
Manufacturers can more quickly identify flaws and deliver high-quality goods to their global customer’s thanks to automated optical inspection, or AOI. Additionally, the NVIDIA Metropolis vision AI framework, which is now supported by AOI, is utilized to improve inspection workflows for a variety of products, including circuit boards and automobiles.
Techman used Isaac Sim to simulate, test, and improve its cutting-edge collaborative robots, or cobots, and NVIDIA AI and GPUs for cloud training and inference on the robots themselves to develop AOI for its factory-floor robots.
Isaac Sim is based on NVIDIA Omniverse — an open advancement stage for building and working modern metaverse applications.
Novel highlights of Techman’s mechanical AOI arrangements incorporate their situation of the review camera straightforwardly on expressed automated arms and GPUs coordinated in the robot regulator.
The bots are able to use AI at the edge to immediately identify defects and inspect product areas that fixed cameras simply cannot access.
“The unmistakable elements of Techman’s robots — contrasted with other robot brands — lie in their implicit vision framework and man-made intelligence derivation motor,” said Scott Huang, boss tasks official at Techman. ” NVIDIA RTX GPUs power up their computer-based intelligence execution.”
However, programming the development of these robots can time consume.
In order to capture hundreds of images as quickly as possible, a developer must determine the most effective sequence and precise arm positions.
This can include a few days of exertion, investigating a huge number of conceivable outcomes to decide on an ideal arrangement.
The remedy: simulation of a robot
Utilizing Omniverse, Techman fabricated a computerized twin of the examination robot — as well as the item to be reviewed — in Isaac Sim.
Programming the robot in recreation decreased time spent on the assignment by more than 70%, contrasted with programming physically on the genuine robot. The application can be developed in the digital twin before the actual product is made by using a precise 3D model of the product. This saves time on the production line.
Techman then explored a huge number of program options simultaneously on NVIDIA GPUs with the help of powerful optimization tools in Isaac Sim.
According to Huang, the final product was an effective solution that cut the cycle time for each inspection by 20%.
Techman’s manufacturing clients’ bottom lines will benefit from every second of inspection time saved.
Since it takes a lot of time and money to collect and label real-world images of defects, Techman turned to synthetic data to boost inspection quality. It utilized the Omniverse Replicator system to rapidly create great manufactured datasets.
These impeccably named pictures are utilized to prepare the man-made intelligence models in the cloud and emphatically improve their exhibition.
Also, many artificial intelligence models can be run at the edge — productively and with low idleness on account of NVIDIA innovation — while examining especially confounded items, some of which take in excess of 40 models to examine their various viewpoints.
Join NVIDIA at COMPUTEX, where the Techman cobots will be on display, to learn more about how Isaac Sim on Omniverse, Metropolis, and AI are streamlining the optical inspection process across products and industries.
Click on the following link Metrologically Speaking to read more such news about the Metrology Industry.
