ABB And PSYONIC Want Robots To Learn Dexterity From Human Hands

ABB Robotics is teaming up with PSYONIC on one of the hardest problems in automation: getting robots to handle objects more like people do.

The collaboration combines ABB’s GoFa cobot with PSYONIC’s Ability Hand, a bionic hand originally developed for prosthetic use. The goal is to use touch, motion and grip data generated by real human prosthetic users to train robots for delicate, variable tasks that are still difficult to automate.

That is a big shift from the usual robot gripper story. Traditional grippers can be great at repeating the same motion with the same part all day. They struggle when objects are fragile, irregular, soft, oddly shaped or constantly changing. Human hands handle those situations without much thought. Robots need a lot more help.

PSYONIC’s Ability Hand brings some of that missing information into the loop. The hand uses myoelectric control, touch sensing and compliant mechanics in a lightweight multi-articulating design. Its pressure sensors and vibration feedback system allow users to detect contact, grip force and release, while the flexible fingers can conform to irregular and deformable objects.

That data is what makes the partnership interesting. PSYONIC founder and CEO Dr. Aadeel Akhtar said dexterous manipulation is “a data challenge as much as a hardware challenge.” By using the same Ability Hand on people and on robots, PSYONIC and ABB can capture real-world data on movement, contact and grip force, then use it to improve robotic systems.

ABB’s GoFa cobot adds the industrial side of the equation. The robot provides the accuracy and repeatability needed to test subtle variations in grip force, finger placement and motion. That is how human-derived manipulation data can start to become repeatable robot behavior rather than just an impressive lab demo.

ABB says grasping and dexterity are central to its vision for Autonomous Versatile Robotics, or robots that can sense, reason, move and handle objects precisely in dynamic environments. The company says the work could apply across automotive, aerospace, packaging, logistics and life sciences, especially in tasks that are repetitive, ergonomically difficult or hard to perform consistently at scale.

The stakes are practical. ABB says advanced gripping and digital integration can reduce engineering time by up to 30 percent, citing the International Federation of Robotics. That matters because complicated end-of-arm tooling can slow automation projects and make them more expensive to deploy.

For now, the project is about integration, testing and development. But it points to a bigger robotics trend. The next wave of useful robots may not just need better arms or better AI models. They may need better hands, better touch data and better examples from humans who already know how to grab things without crushing, dropping or fumbling them.