Soft robotic arms, composed of flexible materials and actuated by artificial muscles, have struggled with stable and versatile control due to unpredictable shape changes under real-world disturbances. The interdisciplinary Mens, Manus and Machina group at MIT, together with Singaporean universities, designed a neural control system that enables soft robots to learn multiple motions once and dynamically adapt to new tasks without retraining or compromising stability.
This advance surpasses prior methods that achieved only one or two key capabilities among task transfer, rapid adaptation, and stability assurance. By addressing these three simultaneously, the researchers bring soft robots closer to human-like flexibility, potentially benefiting fields like rehabilitation and assistive robotics by improving intelligence, versatility, and safety.
Source & further reading
A neural blueprint for human-like intelligence in soft robots ↗Source published 2026-02-19. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.