Humanoid robot able to do complex tasks with little code added
Boston Dynamics and Toyota Research Institute announced this breakthrough Wednesday in a press release.
In a video jointly released by the two organizations, Atlas performs a long, continuous sequence of complex tasks that require combining object manipulation with locomotion. By adopting LBMs, new capabilities that would have been laboriously hand-programmed in the past can now be added quickly and without writing a single new line of code.
The video shows Atlas using whole-body movements, such as walking, crouching, and lifting, to accomplish a series of packing, sorting, and organizing tasks. Throughout the sequences, researchers interject unexpected physical challenges mid-task, such as closing the lid of a box and sliding it across the floor, requiring Atlas to self-adjust in response.
Humanoid robots that have demonstrated this capability before typically separate the low-level walking and balancing control from the control of the arms for manipulation. But a single large behavior model has direct control of Atlas, treating the hands and feet almost identically.
This breakthrough is the result of the October 2024 joint research partnership between Boston Dynamics and TRI, which was designed to use their combined strengths and expertise to speed the development of smart robots.
The project, co-led by Scott Kuindersma and Russ Tedrake, is conducting research to answer fundamental questions about humanoid robots and large behavior models, to advance the field's understanding of large models for whole-body control, including advanced manipulation and dynamic behaviors.
"This work provides a glimpse into how we're thinking about building general-purpose robots that will transform how we live and work," Kuindersma, vice president of Robotics Research at Boston Dynamics, said in a statement. "Training a single neural network to perform many long-horizon manipulation tasks will lead to better generalization, and highly capable robots like Atlas present the fewest barriers to data collection for tasks requiring whole-body precision, dexterity, and strength."
Humanoids can help in many ways.
"One of the main value propositions of humanoids is that they can achieve a huge variety of tasks directly in existing environments, but the previous approaches to programming these tasks simply could not scale to meet this challenge," Tedrake, senior vice president of Large Behavior Models at Toyota Research Institute, said in a statement. "Large behavior models address this opportunity in a fundamentally new way -- skills are added quickly via demonstrations from humans, and as the LBMs get stronger, they require less and less demonstrations to achieve more and more robust behaviors."
Humanoid robots have been an ongoing challenge. The World Humanoid Robot Games highlighted this, showing that tumbling robots is the norm.
The games featured more than 500 robots from 16 countries, including Japan, the United States and Germany, competing in 26 events.
One robot was disqualified from a 1,500-meter race when its head fell off in mid-stride.
"Keeping [the head] balanced while in movement is the biggest challenge for us," said Wang Ziyi, 19, a member of the Beijing Union University team.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
11 hours ago
- Yahoo
Hyundai built its newest factory around AI-powered technology like robotics and digital twins
Hyundai's Georgia EV and hybrid plant has AI-powered robots, drones, and quality-control systems. The technology helps to manage inventory, final inspections, and the parking of finished vehicles. This article is part of "How AI Is Changing Everything: Supply Chain," a series on innovations in logistics. At America's largest car manufacturing plant, nearly everything is powered by AI. Drones buzz and scan stacked pallets to inventory parts. Robotic arms swivel through their programmed arcs. Cameras direct sanding tools to paint blemishes. Uncrewed taxis move vehicles and parts to their next manufacturing destination. A four-legged Boston Dynamics AI dog named Spot walks through the facility and scans components to ensure they're snapped into place. The facility, called Hyundai Motor Group Metaplant America, is roughly equivalent to 278 football fields and comes with a nearly $7.6 billion price tag, according to the company. The HMGMA facility, which celebrated its grand opening in March, puts Hyundai in a unique position compared to other automakers. While most US automakers are retrofitting decade-old plants with AI and robotics, Hyundai built HMGMA from the ground up with AI, Nvidia chips, and robotics at its core. Each vehicle built in the plant — including cars in all three of Hyundai Motor Group's brands — runs through at least 23 AI-based or robotic systems before it's finalized, according to Hyundai. Industry analysts told Business Insider that Hyundai's AI-first approach could be a boon for navigating policy changes and customer trends that impact the carmaker's bottom line. The technology can help improve quality control measures and reduce manufacturing costs in ways that traditional manufacturing can't match. "Hyundai's integration of humanoid robots and such sets a new benchmark for smart manufacturing," Erin Keating, an executive analyst at Cox Automotive, told Business Insider. "Potentially, there are cost savings, and it likely offers the ability for Hyundai and others to respond faster to any industrial disruption." AI isn't replacing Hyundai's workforce At the center of Hyundai's Georgia operation is an integrated digital command hub that acts as a digital twin, or a computer simulation that mirrors the physical plant in real time, said Miles Johnson, Hyundai's communications representative for its electric-powered Ioniq-brand vehicles and tech. Sensors collect data from across the production floor, then an AI algorithm creates a virtual replica of each step in the real-life manufacturing process. "AI can play a significant role in predicting optimized outcomes and identifying root causes of production issues," Johnson told Business Insider. "If a quality issue arises in production, the AI system can search past data it has been trained on to identify the likely cause and suggest a corrective action plan." Keating said that Hyundai's AI can help "detect defects in real time, analyze root causes, and suggest corrective actions that can significantly reduce rework, scrap, and warranty claims." The brand's two retro-modern electric vehicles, the Ioniq 5 and Ioniq 9 SUVs, are built with these highly technical processes. The plant will eventually serve Hyundai's other brands as well; Kia production is set to begin in 2026, with Genesis vehicles planned for an unspecified future date. In a press release, Hyundai said that the AI and robotics systems will assist humans, who will remain in charge of on-the-ground robotics maintenance, data analysis, and parts installation. Hundreds of roles now involve programming the factory's 850 robots, troubleshooting AI, or interpreting production data rather than performing repetitive physical tasks. The company aims to hire 8,500 employees at HMGMA by 2031 and has dubbed the project the largest single business investment in Georgia's history. In March, Hyundai said it had already hired 1,000 full-time staff members at the plant. The company also built a nearby Hyundai Mobility Training Center of Georgia, where it plans to train a pipeline of college-level talent for the plant. The program, which is in coordination with three local technical colleges, is expected to open this year. AI could help carmakers with cost management and supply-chain disruptions Hyundai's massive AI investment faces significant challenges in the current automotive market. Both the Ioniq 5 and Ioniq 9 were positioned to benefit from the then-President Joe Biden's Inflation Reduction Act. The cars are both eligible for the $7,500 tax credits for EVs assembled in American plants. But Americans' uptake of EVs has underperformed analysts' predictions since Hyundai broke ground on HMGMA in October 2022. Now, EV adoption is facing more federal pushback as subsidies for EVs are set to end in September. Hyundai is now betting that its EVs, including the three-row Ioniq 9, which launched in May 2025, will attract buyers without the need for tax incentives. However, some industry analysts say that Hyundai's AI-powered manufacturing can still work, even if it currently relies on EV sales. AI's advantages — like real-time quality control and faster problem solving — could prove crucial as automakers face pressure from increasingly cheap Chinese vehicles and ongoing supply chain volatility. Morningstar analyst David Whiston said that companies adopting AI for manufacturing and logistics are better positioned to manage costs and respond to disruptions. "Automakers leveraging AI for smart factories, autonomous logistics, and predictive analytics will be better positioned to scale production efficiently and meet regulatory and consumer demands faster," Keating said. "AI and robotics are becoming non-negotiable for competitiveness." Read the original article on Business Insider
Digital Trends
12 hours ago
- Digital Trends
Watch this humanoid robot nail complex tasks and think on the fly
Boston Dynamics' humanoid robot may have skipped the inaugural 'robot Olympics' in China last week, but that doesn't mean the engineers behind the machine have been sitting around watching the world go by. Indeed, a video released by the Massachussetts-based company on Wednesday reveals that the team has been hard at work on Atlas, its advanced and highly talented bipedal bot. Working with experts at the AI- and robotics-focused Toyota Research Institute (TRI), Boston Dynamics has equipped Atlas with a Large Behavior Model (LBM), essentially a sophisticated AI system trained on vast datasets of human actions, aimed at enabling the robots to understand, generate, and adapt complex, human-like behaviors for activities in real-world environments. The video shows Atlas performing a lengthy sequence of complex tasks that force it to combine object manipulation with locomotion. They include walking, crouching, and lifting objects, while at the same time packing, sorting, and organizing. 'By adopting LBMs, new capabilities that previously would have been laboriously hand-programmed can now be added quickly and without writing a single new line of code,' the Massachussetts-based company said in a release. To test its ability to adjust itself, an engineer interrupts Atlas in the way that an annoying co-worker might do, by repeatedly closing the lid of the box from which it's taking things, and by sliding the box across the floor. If Atlas had a voice — and no doubt one day it will — it would probably have said: 'Can you quit messing around — I'm trying to get a job done here.' Atlas passes with flying colors, refraining from decking the troublemaker and instead readjusting its position to continue with the task in hand. 'This work provides a glimpse into how we're thinking about building general-purpose robots that will transform how we live and work,' said Scott Kuindersma, Boston Dynamics' vice president of robotics research. 'Training a single neural network to perform many long-horizon manipulation tasks will lead to better generalization, and highly capable robots like Atlas present the fewest barriers to data collection for tasks requiring whole-body precision, dexterity, and strength.' Boston Dynamics is one of a growing number of tech firms working on humanoid robots, with rapidly advancing technology paving the way for increasingly agile, dexterous, and intelligent bipedal robots that could one day perform a huge variety of activities. Even the laundry …
UPI
17 hours ago
- UPI
NASA's TOMEX+ rockets to study turbulence in Earth's mesosphere
One of the three rockets for the TOMEX+ mission sits on a launcher at Wallops Flight Facility in Virginia. Photo by Danielle Johnson/NASA Aug. 20 (UPI) -- NASA is planning to launch TOMEX+ rocket mission to study the turbulence where Earth's atmosphere ends and outer space begins sometime over the next two weeks. The earliest the agency expects to launch the three sounding rockets is Saturday, with the launch window closing Sept. 3, NASA announced Wednesday. The launch window has been repeatedly pushed back, this time due to high sea states in the rocket recovery area from Hurricane Erin. Sounding rockets are those that can be aimed to reach the Earth's mesopause, an area of the atmosphere that's too high for weather balloons and too low for traditional satellites to reach. The mesosphere is the coldest layer of Earth's atmosphere and is the place where weather patterns on the planet can transfer energy into space, creating turbulence that can affect satellites. NASA will use three sounding rockets to carry out the study, the first two launching within about a minute of each other. They will release vapor that will help scientists on the ground map wind patterns in the mesosphere. The third rocket will then use a lidar -- light detection and ranging -- device to send out pulses of light to detect sodium atoms in the mesosphere. The sodium comes from tiny grains of dust that enter the Earth's atmosphere from space. The lidar will allow scientists to further map the mesosphere's density and motion over time. "Together, the TOMEX+ payloads will provide the clearest 3D view yet of turbulence at the edge of space, improving our understanding of high-altitude cloud formation, satellite drag and even atmospheric processes on other planets," NASA said in a blog post on the project. NASA will launch the rockets from Wallops Flight Facility on Wallops Island, Va. The project is led by Jim Clemmons, a professor of physics and astronomy at the University of New Hampshire.
