Meta Boosts Robotics with New AI Touch and Dexterity Innovations
Meta’s Fundamental AI Research (FAIR) team is making significant advancements in the field of robotics by focusing on advanced machine intelligence (AMI) and tactile sensing. By collaborating with industry leaders like GelSight Inc. and Wonik Robotics, Meta is introducing groundbreaking tools to enhance touch perception, dexterity, and human-robot interaction.
Building a Foundation in Touch Perception and Dexterity
Meta has introduced several tools aimed at improving touch perception and dexterity in AI-driven robots:
- Meta Sparsh: A general-purpose touch representation that enhances AI's ability to perceive details beyond vision.
- Meta Digit 360: A tactile sensor with multimodal sensing capabilities similar to human touch, enabling robots to detect fine spatial details and forces as minimal as 1 millinewton.
- Meta Digit Plexus: A platform that integrates various tactile sensors onto a single robotic hand for coordinated touch perception and motor control.
These innovations are expected to have practical applications in industries such as healthcare and manufacturing, where precise machine dexterity and perception are essential for complex tasks.
Partnerships to Expand Access to Tactile Sensing Technologies
Meta is collaborating with GelSight Inc. and Wonik Robotics to make their technologies accessible to researchers worldwide:
- GelSight Inc. will manufacture and distribute the Meta Digit 360 sensor, with early access available through a call for proposals.
- Wonik Robotics will release a new generation of the Allegro Hand equipped with Meta Digit Plexus and tactile sensing capabilities.
These partnerships aim to encourage the adoption of tactile sensing technology in research and development, making it more widespread and ubiquitous.
PARTNR: A Benchmark for Human-Robot Collaboration
Meta has introduced the PARTNR benchmark to assess planning and reasoning in human-robot interaction. This tool, built on Habitat 3.0, provides a simulation environment for scalable testing of AI-human collaboration in various tasks, facilitating improvements in AI planning, perception, and skill execution in social settings.
Vision-Based Tactile Sensing with Sparsh
Meta has launched Sparsh, a vision-based tactile sensing model capable of processing inputs from different tactile sensors. Trained on a large dataset of tactile images, Sparsh outperforms traditional models, supporting the efficient scaling of touch perception technologies.
Enabling New Frontiers in AI and Robotics Research
Meta FAIR’s commitment to open-source collaboration and partnerships with industry leaders will enhance the accessibility of robotics technologies, fostering innovation in fields like medical robotics, virtual reality, and supply chain management.
For more information about these innovations, please visit Meta AI's blog.
