Facebook AI unveils open source tactile tech, including ReSkin, a robot “skin”, and DIGIT, a high-resolution fingertip sensors to give robots a sense of touch
to help teach the robot - Cheap: < $6 per unit @ 100 units https://ai.facebook.com/... https://twitter.com/... Matt Beane / @mattbeane : This is the kind of thing that shows that Facebook's rebranding as “Meta” isn't just cheap talk. They're building new sensors. A real, distinctive commitment to the vision outlined in the Meta launch. But not a huge one. And it's going to take “huge” to make the shift happen. https://twitter.com/... Soumith Chintala / @soumithchintala : For example, ReSkin here is put as dog “socks”, to collect fine-grained contact data as the dog runs around: https://twitter.com/... @jfclifford : @schrep Also this would be a great feedback system for haptics, have an array of microchannels that you could inflate against a finger, use the particles to provide distance feedback and create a closed loop system. Maybe even use an array of micro coils to control inflation points? Meta AI / @facebookai : We're also open-sourcing ReSkin, an affordable tactile sensor that autocalibrates through self-supervision. ReSkin will help to advance AI's ability to obtain & use tactile data, getting us closer to deploying reliable touch sensing in research.
Robotics researchers gain open-source designs for ReSkin and DIGIT, along with a low-cost ReSkin option for gathering fine-grained contact data.
Facebook AI broadens its robotics tooling from PyRobot's hardware-agnostic programming layer into the sensor hardware that supplies touch data.
Second-order effects
Robot developers can evaluate touch-enabled models without designing a tactile sensor from scratch, shifting more experimentation toward data collection and integration.
Specialist tactile-sensor suppliers face a more accessible reference design, while companies able to commercialize higher-performance hardware retain a path to differentiation.
Third-order effects
If open designs become common evaluation tools, tactile robotics may organize around shared sensor baselines and proprietary data, calibration methods, and production hardware.
Meta's later work with Wonik and GelSight suggests a wider progression from open research instruments toward commercial tactile-sensing components.
The trend: Robotics AI is moving beyond vision and control software toward accessible, sensor-native touch data as a core input for physical intelligence.
In addition to hardware such as DIGIT, developing touch sensing as a modality requires an ecosystem of touch processors, simulators, benchmarks & data sets. Read more about our ongoing work in developing & open-sourcing this ecosystem: https://ai.facebook.com/... https://twitter.…
Giving robots the sense of touch! @facebookai is open sourcing a new class of inexpensive touch sensors - DIGIT https://ai.facebook.com/... and ReSkin https://ai.facebook.com/... - and aiming to develop an ecosystem of data sets, benchmarks, and processing software around touch s…
Facebook has developed a suite of tactile technologies that will impart a sense of touch into robots. It's a critical development in robotics, that Meta is driving this shows the scope of their ambitions. https://www.engadget.com/...
New touch sensing tech from Meta AI: DIGIT: optical touch sensor now available commercially. Can sense texture, force, shape, etc, at high resolution. ReSkin: magnetic touch sensing technology, now open source. PyTouch: PyTorch library for touch sensing. #robotics #metaAI https:/…
ReSkin: a deformable elastomer with embedded magnetic particles. When it deforms, local magnetic signal changes - Robots wear it for touch sensing - Humans wear it for data-collection — to help teach the robot - Cheap: < $6 per unit @ 100 units https://ai.facebook.com/... https:/…
This is the kind of thing that shows that Facebook's rebranding as “Meta” isn't just cheap talk. They're building new sensors. A real, distinctive commitment to the vision outlined in the Meta launch. But not a huge one. And it's going to take “huge” to make the shift happen. htt…
@schrep Also this would be a great feedback system for haptics, have an array of microchannels that you could inflate against a finger, use the particles to provide distance feedback and create a closed loop system. Maybe even use an array of micro coils to control inflation poin…
We're also open-sourcing ReSkin, an affordable tactile sensor that autocalibrates through self-supervision. ReSkin will help to advance AI's ability to obtain & use tactile data, getting us closer to deploying reliable touch sensing in research. More: https://ai.facebook.com/... …
Today, our Meta AI team is outlining the progress in developing an open source ecosystem (including hardware, simulators, libraries, benchmarks, and data sets) necessary for building AI systems that can understand and interact through touch: https://ai.facebook.com/... https://tw…