From handling sheet metal and working in the body shop to managing warehouse operations, the robot is extremely dexterous and versatile at a height of 1.6 meters and a weight of 60 kilograms. What’s more is that Figure also announced that all robots will undergo a training period of 12 to 24 months to fine-tune the robot’s skills to meet the precise requirements of each task in the facility. Crucially, its neural networks enable ultra-precise manipulation of objects by converting the visual data from its cameras into specific robot actions. Plus, the humanoid’s intelligence and logic is also powered by OpenAI’s cutting-edge vision language model, which processes data directly from the robot’s cameras.

Impressively, these neural networks analyze images at 10 hertz and then produce actions at 200 hertz with 24 degrees of freedom. This setup allows for sheet metal to be placed with intolerances tighter than one centimeter solely through simulation. And while the system currently operates at a moderate pace, the precise and autonomous nature of these actions are leading to a future where humanoid robots will become commonplace in industrial settings. To do this, Figure aims to create a global model capable of managing billions of humanoid robots at once, highlighting approximately 10 million unsafe or undesirable jobs in the U.S.

alone that the robot could address. Meanwhile, Chinese robotics firm UBTEC has partnered with Four Volkswagen to be deployed in car factories in East China. Its Walker S humanoid is 1.7 meters tall with an enhanced body structure, perception and more. As for the body, the Walker S is newly equipped with force-compliant drive joints and a rigid flexible coupling hybrid structure, which together ensure that the robot can walk stably even in dynamic environments. This improvement marks a significant leap towards integrating humanoid robots into fast-paced industrial settings. Equipped with 41 servo joints that provide force feedback, along with multiple visual, audio and distance sensors, the Walker S is designed for comprehensive perception.

Innovative multimodal sensor fusion and dynamic environment exploration enable it to reliably perceive its surroundings. Additionally, the Walker S features high-resolution RGBD visual sensors that allow it to observe its environment in great detail. Using these sensors, the robot can create 3D semantic maps of its surroundings, design walking routes, and avoid obstacles efficiently. This capability ensures not only the safe operation of the robot, but also its ability to execute tasks with high precision. Another standout feature of Walker S is its integration with a highly intelligent large language model. This integration enhances the robot’s intent understanding and fine-grained planning capabilities, building on its existing multimodal perception and motion control abilities.

As a result, Walker S can perform a wide variety of tasks across different environments, making it an incredibly versatile tool for industrial applications. And thanks to its ability to process 3D point cloud information, Walker S also excels in object detection and grasping. This allows the robot to recognize the 6D pose of complex objects and complete grasping tasks with remarkable hand-eye coordination. Such precision in object handling is crucial for intricate manufacturing processes. And finally, in order for Walker S to integrate with manufacturing management systems, it can also automatically connect to these systems to exchange real-time information about production status.

This capability guarantees data synchronization throughout the factory for smooth and efficient operations. Meanwhile, Elephant Robotics has unveiled its latest Mercury X1 humanoid robot designed to handle precision tasks and enhance efficiency in various environments. In a simulated environment, Mercury X1 showcases its dynamic visual recognition and grasping abilities. Utilizing simultaneous localization and mapping for navigation, Mercury X1 precisely avoids obstacles and moves to the front of a table. Equipped with a camera mounted at the end of its arm, the robot scans the tabletop using tag recognition to identify objects for picking up.

Once the scan is complete, it determines the location of the objects and executes the grasping action with precision. In another scenario, Mercury X1 demonstrates its adeptness in handling a box. The robot starts by scanning the tabletop to locate its target. It then skillfully grips the box and lifts it steadily, ensuring the objects balance throughout the transportation process. This demonstration highlights Mercury X1’s capability to handle delicate and precise tasks, making it an invaluable asset in various settings. Mercury X1 also excels as an office assistant. In the document delivery demo, the robot retrieves documents from the printer and precisely packs them for delivery.

It then navigates to the reception desk on the fourth floor, showcasing its robust navigation capabilities. As Mercury X1 moves through corridors and avoids obstacles, it uses high-precision sensors and advanced algorithms to ensure safe and accurate movement. And upon reaching the elevator, Mercury X1 autonomously presses the button and boards it. The robot intelligently recognizes the elevator’s position and status, ensuring a smooth ride to the fourth floor. Finally, Mercury X1 delivers the documents accurately to the staff member, successfully completing the task. Because of this, Mercury X1 transcends the traditional roles of robots by embodying the intelligent assistant for future offices, enhancing efficiency and convenience in our daily tasks.

With its advanced capabilities in precision handling, navigation and interaction, Mercury X1 is poised to revolutionize the way we approach office automation and efficiency. And finally, Home AI has launched an AI-as-a-service spatial planning solution to revolutionize the $650 billion home decor market. This innovation aids furniture makers and home designers in exploring new business opportunities in the digital age. Leveraging NVIDIA Omniverse, Home AI is a service to deliver accurate 3D furniture assets directly to users’ homes. The Universal Scene Description framework behind Omniverse allows Home AI to connect diverse customer data sources, enhancing their spatial AI solution.

Furthermore, their home optimal planning engine uses USD metadata to recommend ideal furniture and placement, powered by NVIDIA RTX technology for real-time rendering and AI performance. With this new tech, consumers can now visualize furniture in their homes through photorealistic OpenUSD-based 3D models. This technology streamlines digitalization efforts for home decor companies and empowers consumers to try products virtually before purchasing. And Home AI’s platform is quickly gaining traction among furniture manufacturers and suppliers who can now connect their asset libraries to consumer homes to accelerate the purchasing process. [tr:trw].