NVIDIA and Taiwan’s manufacturing ecosystem, together with Delta Electronics, Foxconn, TSMC and Wistron, are showcasing this week at COMPUTEX in Taipei the essential function digital twins play in accelerating industrial AI.
These electronics, semiconductor and robotics manufacturing leaders are utilizing Common Scene Description (OpenUSD) and NVIDIA Omniverse libraries and blueprints to develop bodily primarily based digital twins. That is remodeling manufacturing unit planning by unlocking new operational efficiencies and accelerating the event, testing and validation of autonomous robots and robotic fleets.
Many of those producers are additionally extending the digitalization of their factories to the true world, utilizing the NVIDIA AI Blueprint for video search and summarization (VSS) — now typically accessible and a part of the NVIDIA Metropolis platform — to deploy video analytics AI brokers into their operations and drive extra automation and optimizations in defect detection and different operations.
Taiwan Producers Optimize Planning and Operations With Simulation and AI Brokers
Taiwan’s main electronics and semiconductor producers are utilizing digital twins, bodily primarily based simulation and AI brokers to optimize present operations and vastly speed up the planning and commissioning of latest factories.
Foxconn is main the best way. At its Taiwan services, Foxconn engineers depend on the Fii Digital Twin platform, developed with OpenUSD, Siemens and Omniverse applied sciences, to design and simulate robotic work cells, meeting strains and whole manufacturing unit layouts.
These digital twins hook up with materials management methods and use Autodesk Flexsim, NVIDIA cuOpt and NVIDIA Isaac Sim to allow engineers to simulate and dynamically optimize the circulate of supplies, tools, autonomous cell robots (AMRs), automated guided autos, and different robots and people. By growing a normal digital twin mannequin for his or her factories, Foxconn can shortly migrate and simply reconfigure its designs and plans for brand spanking new manufacturing unit deployments.
Foxconn is utilizing the NVIDIA Isaac GR00T N1 mannequin, the NVIDIA Isaac GR00T-Mimic blueprint for artificial manipulation movement era and NVIDIA Isaac Lab to coach industrial manipulator arms and humanoid robots for performing complicated duties similar to screw-tightening, choose and place, meeting and cable insertion. Foxconn robotics builders use the Mega NVIDIA Omniverse Blueprint to simulate and check giant robotic fleets comprising AMRs, manipulators and humanoid robots earlier than deploying them in services.
To speed up evaluation and decision-making, Foxconn engineers use their digital twin platform to conduct thermal assessments of POD rooms throughout totally different eventualities. By connecting their digital twins to the Cadence Actuality Digital Twin Platform and integrating NVIDIA PhysicsNeMo frameworks, groups can conduct thermal simulations 150x sooner, cut back thermal dangers and determine energy-saving alternatives.
Utilizing the Omniverse Blueprint for AI manufacturing unit digital twins, Foxconn can simulate and check GB200 Grace Blackwell Superchips in liquid-cooled PODs to copy the situations of an AI manufacturing unit.
The corporate can also be deploying video analytics AI brokers utilizing the VSS blueprint from NVIDIA Metropolis for real-time video evaluation and insights in reside manufacturing eventualities.
TSMC is collaborating with an AI-powered digital twins startup to optimize the planning and building of its new fabs. TSMC faucets into an AI engine and purposes constructed with Omniverse libraries to remodel conventional 2D computer-aided designs into wealthy, interactive 3D layouts of their complicated services, together with specialised areas like clear rooms.
Visualizing these optimized layouts in a digital twin permits planning groups to proactively determine and resolve tools collisions, perceive system interdependencies, and assess impacts on house and operational key efficiency indicators.
This AI-driven strategy is enhanced by NVIDIA cuOpt for optimization and reinforcement studying with NVIDIA Isaac Lab, enabling the era of intricate, multilevel piping methods in seconds — a activity that historically requires substantial effort and time. This permits engineers to just about validate complicated pipe routing and drastically cut back design revisions, finally streamlining all the fab growth course of.
TSMC additionally makes use of imaginative and prescient language fashions and imaginative and prescient basis fashions to enhance automated defect classification workflows — boosting effectivity to categorise wafer product defects for engineers to pinpoint potential root causes for the problems. Past using digital twins and imaginative and prescient AI, TSMC additionally faucets into NVIDIA CUDA-X software program libraries and NVIDIA GPUs to speed up its complete semiconductor chip design workflow — from lithography with NVIDIA cuLitho to semiconductor course of simulation.
Wistron groups drive operational efficiencies, optimize format planning of their vegetation, and practice robots and staff with the Wistron Digital Twin (WiDT) platform. The platform is powered by software program from Autodesk, Cadence and Microsoft and faucets into NVIDIA AI and Omniverse libraries.
By connecting the WiDT platform to generative AI instruments and real-time information from floor mount expertise machines and shopfloor management methods, operations groups can visualize real-time dashboards to shortly diagnose and enhance machine and plant efficiency.
Wistron robotics builders use the platform, and its integration with NVIDIA Isaac Sim, to simulate and check robotic arms. With a simulation-first strategy, groups diminished the time wanted for every arm to assemble components on the manufacturing line by 12 seconds.
The Wistron digital twin platform additionally makes use of the VSS blueprint to create and curate coaching movies for educating staff the right way to carry out and handle complicated duties and eventualities. The platform makes use of NVIDIA Cosmos Tokenizer to assist groups analyze and break down employee actions on the manufacturing line and enhance customary working procedures. This strategy is enabling Wistron to speed up onboarding, enhance employee productiveness and guarantee security.
Wiwynn makes use of AI-enabled digital twins constructed with Omniverse applied sciences to optimize manufacturing unit layouts, simulate manufacturing, combine cobots and improve high quality management via improved inspection and evaluation. These options have pushed important manufacturing and logistics innovation and efficiencies.
Pegatron’s PEGAVERSE and PEGAAi platforms equip engineers and manufacturing unit managers with digital twins that assist many use instances, together with manufacturing unit planning, predictive upkeep, course of optimization, useful resource planning, distant monitoring and high quality management.
Groups additionally use the platforms to construct visible AI brokers to assist staff good complicated meeting duties. These AI brokers, developed with the NVIDIA AI Blueprint for VSS and NVIDIA Metropolis, have enabled Pegatron to reinforce meeting processes, cut back labor prices by 7% and reduce meeting line defect charges by 67%.
Kenmec and MetAI are utilizing Omniverse applied sciences and the Mega NVIDIA Omniverse Blueprint to construct bodily correct digital twins for simulating, testing and deploying warehouse automation options. Collectively, the groups virtualized all the Chief Good Logistics Middle, making a full-fidelity simulation surroundings that brings collectively bodily dynamics, real-time controller logic, AI-driven testing and optimization — all inside a simulated surroundings.
GIGABYTE operations groups are utilizing digital twins developed with Omniverse libraries and linked to reside IoT information from the manufacturing flooring to enhance operational monitoring of manufacturing methods. By visually flagging anomalies, together with tools points and delays, the digital twins assist groups shortly determine points, conduct root trigger evaluation and take corrective actions.
Quanta Cloud Know-how engineering, operations and logistics groups collaborate utilizing digital twin options constructed with Omniverse to speed up manufacturing unit planning. Digital twins present these cross-functional groups with entry to the most recent design information, enabling them to offer rapid suggestions on proposed layouts, which ends up in optimized workflows and improved house utilization. Groups can additional lengthen collaboration periods to exterior clients and suppliers to allow them to remotely contribute to design critiques and validation.
Producers Embrace Digital Twins to Speed up Robotics Growth
Along with creating the long run in manufacturing, Taiwan producers are utilizing digital twins, powered by Omniverse libraries and blueprints, to develop the following wave of AI-enabled robots.
Delta Electronics is utilizing Isaac Sim to optimize digital part manufacturing and to simulate, practice and validate its complete vary of commercial robots — from AMRs to industrial manipulators.
The corporate is remodeling its experience right into a service by designing a cyber-physical built-in classroom to be launched quickly in Taiwan, the place clients be taught to make use of the DIATwin platform to simulate and combine Delta’s industrial tools and robots to make sure a simpler implementation into their very own manufacturing strains.
Techman Robotic is advancing clever automation at Volkswagen’s Clear Manufacturing facility. Utilizing Isaac Sim, Techman’s AI Cobots be taught to function on GESSbot AMRs in bodily correct simulations to carry out real-time meeting, inspection and adaptive manipulation duties with precision. By simulating robotic habits and workflows just about, Techman Robotic has diminished the time to program robots by 70% and improved robotic productiveness by 20%.
Foxlink is utilizing the Isaac GR00T N1 mannequin so as to add generalized intelligence and autonomy to its industrial robots utilized in manufacturing services.
Solomon’s AI imaginative and prescient answer, powered by NVIDIA Isaac Manipulator CUDA-X acceleration libraries, helps Inventec considerably speed up its robotic server inspection course of by boosting complicated movement planning velocity by as much as 8x and lowering errors by 50%.
Kudan is integrating its Visible SLAM expertise with Isaac Perceptor CUDA-X acceleration libraries into NexAIoT’s AMR, NexMOV-2. This integration makes use of superior 3D notion and navigation, enabling them to navigate complicated, unstructured environments similar to manufacturing, logistics and healthcare services with higher precision and reliability.
MSI is powering its industrial robots with the NVIDIA Jetson AGX Orin module to carry out a wide range of duties, from pick-and-place and materials dealing with to delivering payloads inside giant warehouses and services.
In healthcare, Adata and Advantech are collectively utilizing Isaac Sim, Isaac Perceptor and Jetson Orin to develop AMRs for disinfecting hospitals. This collaboration has diminished deployment time by 70% and made the disinfection course of 3x sooner. Ubitus can also be utilizing the Isaac platform to coach G1 humanoid robots to ship medical checkup supplies and specimens, serving to alleviate labor shortages in hospitals.
Be taught extra by watching the COMPUTEX keynote from NVIDIA founder and CEO Jensen Huang and attending periods at NVIDIA GTC Taipei, operating via Could 22.
See discover relating to software program product info.
Featured picture courtesy of Quanta (high left), Wistron (high proper), Foxconn (backside left), Pegatron (backside proper).
