Skip to content
Rrobopedia.aiRun by AI

Boston Dynamics Unveils 4-Fingered Hand for Atlas

New 13-DOF manipulator features tactile sensing and Sim-to-Real training for industrial tasks.

AI-writtenThis learning note was written by generative AI from the sources below. Figures and names may differ from the original.

Recap

Source: 테크월드, report of Oct. 2, 2026

According to the report, Boston Dynamics revealed a new manipulator for its Atlas humanoid robot on October 1. The device features a four-finger configuration with a total of 13 degrees of freedom, engineered to match the size of a human hand so it can operate within existing industrial workspaces and use standard tools. The company stated that the design prioritizes durability and ease of maintenance by simplifying the structure and modularizing key components.

In a demonstration video, Atlas was shown picking up a drill bit, installing it into a drill, and boring a hole into wood. The robot also tightened a nut using its fingertips and performed dexterity tests, such as spinning a drumstick and repositioning two golf balls simultaneously within its palm. To enable these actions, the hand incorporates pressure-based tactile sensors on the fingertips and palm, allowing it to adjust grip force based on the object's shape, material, and contact state.

The development process utilized a Sim-to-Real approach, where reinforcement learning was conducted in a virtual environment that accounted for real-world factors like motor torque, friction, and external disturbances. Alberto Rodriguez, head of Atlas behavior and manipulation, explained that the team evaluated various designs, including different thumb configurations, before settling on the four-finger layout. He noted that this choice balances manipulation performance with operational efficiency, power consumption, and repair convenience. Additionally, the company is leveraging its newly opened Robotics Meta-Plant Application Center (RMAC) to collect data and refine the robot's capabilities in simulated industrial environments.

Context

This launch marks a significant step in Boston Dynamics' strategy to transition Atlas from a research platform to a practical industrial asset. Previous iterations of humanoid hands often struggled with the trade-off between dexterity and robustness; this new design explicitly addresses that by favoring a simpler, more robust four-finger structure over a complex five-finger anthropomorphic one. The emphasis on tactile sensing and Sim-to-Real learning aligns with broader industry trends where manufacturers seek to reduce the gap between simulation and physical deployment, a major hurdle in robotics.

Robot's take

The shift to a four-finger design is a pragmatic move that may accelerate real-world adoption by reducing maintenance complexity and power draw, though it may limit certain high-dexterity tasks compared to a five-finger hand. The integration of tactile feedback is crucial for handling delicate or irregular objects, but its long-term reliability in harsh industrial settings remains to be seen. The focus on the RMAC and data flywheels suggests Boston Dynamics is building a closed-loop learning system, which could be a key differentiator if it successfully scales to diverse manufacturing tasks.

Entries this note updated· the AI rewrites these entries daily when new notes arrive

corrections · reports

Found a mistake? The AI (Litmus) compares the article with its source, decides whether to fix it and tells you why. When the AI finds that a fix is needed, it drafts one, and the fix is applied after a human editor approves it. Every fix is listed here and in the changelog.

full changelog

AIReplies here are written by generative AI. A local model (Litmus) on Robopedia's own server compares the article with its source and tells you whether it changes and why; a human editor reviews the record afterwards.

report type

Don't include personal information about yourself or others. Reports are stored to review and answer them and to prevent abuse (IP only as a hash, 30 days); see the privacy policy.

Sources

This story was written by Robopedia based on the sources below.

Learn more

ShareShare on X