Have you heard the exciting news about Boston Dynamics teaming up with the Robotics & AI Institute to develop a cutting-edge humanoid robot named Atlas? This groundbreaking partnership, led by Boston Dynamics’ visionary Marc Raibert, aims to enhance Atlas’s abilities through reinforcement learning, a form of machine learning that teaches machines through trial and error.

The project, which just kicked off, will focus on refining Atlas’s skills in tasks like opening doors and running. By leveraging the expertise of the Robotics & AI Institute, Boston Dynamics is pushing the boundaries of humanoid robotics and making it easier for Atlas to acquire new skills.

Raibert, the mastermind behind Boston Dynamics, founded the company in 1992 after leaving MIT, where he was a professor for nearly a decade. In addition to their famous robot dog “Spot,” Boston Dynamics has a diverse lineup of commercial robots, including Stretch, a warehouse box-moving robot. Their original hydraulic humanoid robot, introduced in 2013, has been replaced by an electric version for real-world applications.

In a strategic move, Hyundai acquired a majority stake in Boston Dynamics in 2021, valuing the company at $1.1 billion. This led to the launch of the Robotics & AI Institute with $400 million in funding, led by Raibert. The institute focuses on addressing AI challenges in robotics, including cognitive AI, athletic AI, and robotics ethics.

The research team at the Robotics & AI Institute is dedicated to enhancing Atlas’s functionality by focusing on object manipulation, full-body coordination, and dynamic running strategies. Collaborating with Toyota’s research unit, the project aims to enhance Atlas’s capabilities and versatility across various environments and applications.

Boston Dynamics CEO Robert Playter is optimistic about the partnership with the Robotics & AI Institute, highlighting the accelerated development of core capabilities crucial for integrating humanoid robots like Atlas into everyday life. As progress continues, the research team will provide updates through publications and demonstrations, showcasing Atlas’s evolution as a valuable tool for diverse tasks and environments.