Boston Dynamics and the Robotics & AI Institue are training Atlas in parallel simulation. Source: RAI Institute
Two of the hottest technologies — humanoid robots and artificial intelligence — are converging. Boston Dynamics Inc. and the Robotics & AI Institute today announced a partnership to advance humanoids through reinforcement learning.
The partners said they will establish a shared reinforcement-learning training pipeline for Boston Dynamics’ Atlas robot to build dynamic and generalizable mobile manipulation behavior. Boston Dynamics and the Robotics & AI Institute (RAI Institute) began working this month on new capabilities for Atlas. The robot won a 2024 RBR50 Robotics Innovation Award for its abilities, and an electric version debuted in April 2024.
“We are living in an extremely exciting time for humanoid robot development,” stated Robert Playter, CEO of Boston Dynamics. “But for humanoids to be useful, they must be flexible enough to work in many different kinds of environments and perform tasks in a wide variety of applications. Our collaboration with the RAI Institute brings together two of the world’s leading robotics organizations to help accelerate core capabilities needed to make robots like Atlas a valuable tool in people’s lives.”
Hyundai Motor Group acquired Boston Dynamics in 2020. It launched the Boston Dynamics AI Institute, later the AI Institute and now RAI Institute, with $400 million in 2022.
Aaron Saunders, the chief technology officer of Boston Dynamics, will be speaking at the Robotics Summit & Expowhich will be in Boston on April 30 and May 1, 2025.
Register today to save 40% on conference passes!
Partnership sets goals for advancing humanoids
The partnership expands upon previous collaborations between Boston Dynamics and the Robotics & AI Institute. For instance, they jointly developed the Reinforcement Learning Researcher Kit for the Spot quadruped robot.
Released last year, the kit trains unique behavior and modes of locomotion on Spot, and researchers used it to achieve record running speeds of 11.5 mph (5.2 m/s.) The new partnership adds a focus on humanoid robots. Boston Dynamics and the RAI Institute stated several objectives for the project:
Develop sim-to-real for mobility: Despite advances in fast parallel simulators and sophisticated policy-optimization techniques, transferring simulation results to real robotic hardware remains one of the most challenging aspects of applying reinforcement learning in robotics, the partners said. To bridge the sim-to-reality gap, they will train policies that generate a variety of agile behavior on physical hardware to achieve novel, robust, and practical locomotion behavior.
Improve whole body loco-manipulation: A robot’s capability to manipulate objects and fixtures, such as doors and levers, in conjunction with locomotion would significantly enhance its utility, acknowledged the researchers. The team said it will explore new policies to improve robustness in these scenarios.
Explore full-body contact strategies: This work will explore high-performance, whole-body locomotion and tasks that require full-body contact strategies, such as dynamic running and full-body manipulation of heavy objects, necessitating close coordination between the arms and legs. The aim is to utilize reinforcement learning to generate such behavior during complex contact events without imposing strict requirements.
Robotics & AI Institute looks to upgrade robot skills
“Our aim at RAI is to develop technology that enables future generations of intelligent machines,” said Marc Raibert, founder of Boston Dynamics and executive director of the Robotics & AI Institute.
“Working on Atlas with Boston Dynamics enables us to make advances in reinforcement learning on arguably the most sophisticated humanoid robot available,” he added. “This work will play a crucial role in advancing the capabilities of humanoids not only by expanding its skillset, but also streamlining the process to achieve new skills.”
The Cambridge, Mass.-based RAI Institute said it is focused on solving “the most important and fundamental problems in AI and robotics.” It is developing the perception, situational understanding, reasoning, and other cognitive functions underpinning robot abilities and combining them with advances in their physical capabilities.
The institute is conducting research in four core areas: cognitive AI, athletic AI, organic hardware design, and ethics related to robotics. It asserted that its culture combines the best features of a university research lab with those of a corporate development lab.
“The RAI Institute develops technology that will increase productivity, contribute to human safety, care for people with disabilities, and help people live their lives more fully,” it said. “Partnering with Boston Dynamics for this collaboration will both build upon the RAI Institute’s existing research in AI and learning, and help the organization further advance the technology as it applies to other robot platforms.”
Boston Dynamics commercializing its robots
While the Robotics & AI Institute has focused on pure research, Boston Dynamics has moved to commercialize its systems. In addition to Spot and Atlas, the Waltham, Mass.-based company has developed Stretch, a robot for unloading boxes from trailers and containers.
“Boston Dynamics has a longstanding reputation for creating groundbreaking advances in dynamic robots, from extreme mobility to bimanual manipulation,” said the institute. “The latest generation of Atlas is the result of years of hardware and software co-design aimed at building the most capable humanoid platform, both in terms of physical capability and software interfaces for authoring behavior. This makes it an ideal platform for advancing the science of AI-based manipulation skills.”
Since 2020, Spot has conducted industrial inspections for asset management at commercial customers and become more capable. In October 2024, German retailer Otto Group deployed Spot and Stretch for the first time together in Europe. Also that month, Boston Dynamics said it was working with the Toyota Research Institute (TRI) on developing general-purpose robots.
The electric version Atlas has also advanced in movement and dexterous manipulation with the goal of commercial use.
GIPHY App Key not set. Please check settings