Recently, we spoke Simulation in robotics, and associated challenges. Within days, at Computex 2024, NVIDIA head Jensen Huang spoke enthusiastically about how simulation and robotics are going to change everything.

“The next wave of AI is here. Robotics, powered by physical AI, will revolutionize industries,” Huang said.

Screenshot 2024 06 05 at 11.04.57 AM

Jensen Huang’s Keynote at Computex 2024 Source: NVIDIA

The rising tide of physical AI

“Physical AIs are models that can understand instructions and perform complex tasks autonomously in the real world,” said Huang, who is optimistic about how robots will become a part of every industry.

“Everything is going to be robotic,” he said. Huang believes that there will be an entire ecosystem of robots, where all factories will orchestrate robots and those robots will produce robotic products. NVIDIA is banking on Omniverse to do just that.

NVIDIA’s leading simulation

Underpinning it is NVIDIA’s Omniverse, a platform designed for real-time 3D design collaboration and simulation. Digital twins which is important for simulation. Digital twins are virtual replicas of physical objects or systems where robots can be tested for real-world fit.

Demonstrating several scenarios on NVIDIA’s Omniverse where robots have been trained, Huang talked about how companies are building robotic warehouses around it.

In digital twins, factory planners optimize floor layouts and line configurations and find optimal camera placement to monitor future operations. Also called ‘Robot Gym’, Foxconn developers train and test NVIDIA ISAC AI applications for robotic perception and manipulation in Omniverse digital twins.

Multimodal LLM has only accelerated the robotic training process. “Multimodal LLMs are breakthroughs that enable robots to learn, perceive and understand the world around them and plan how they will act,” Huang said.

By combining this technique with human demonstrations, robots can acquire the skills needed to interact with the world using gross and fine motor skills.

Although simulation may sound like everything, it is Training robots isn’t the only way.. Gokul NA, co-founder of Bangalore-based CynLr Robotics, believes that no single method is perfect. He believes that this does not work when we move from artificial hypothesis to reality. It fails completely because it never learned that; He has learned something else independently.

Also suitable for a certain number of simulation tasks. Tasks such as walking, backflips, and other movements that require robotic balance work best in an artificial environment. However, tasks that can be learned through imitation do not require a simulated environment.

Independent everywhere

“One day, everything that moves will be autonomous,” Huang said. And it is most likely that it will be powered by NVIDIA.

Teaching robots to grasp and handle objects is one thing, but autonomously navigating environments and avoiding obstacles or hazards is another capability inherent in physical AI.

Interestingly, Huang recently mentioned that “Tesla is the leader in self-driving cars, but every car will someday have autonomous capability”.

Earlier last year, NVIDIA and Foxconn partnered to build AI factories that would Help expand production of EVs and autonomous vehicles.. They also partnered with major electronics manufacturers Delta Electronics, Pegatron and Westron.

The robotics race is on.

The robotics race is only gaining steam with all the recent developments. Big tech companies have invested aggressively in robotics companies over the past few years. Figure 01, a humanoid created by deep-tech robotics company Figure AI, is backed by some of the biggest players like NVIDIA, Microsoft, Jeff Bezos, and others.

Similarly, 1X Technologies, a robotics company backed by OpenAI, recently released its humanoids that display multiple autonomous tasks from behind.

Given all these developments, the wave of ‘physical AI’ is already underway.

Source link