A neural blueprint for human-like intelligence in soft robots
AI Summary: Researchers from the Mens, Manus and Machina (M3S) group, in collaboration with the National University of Singapore and MIT, have developed a novel AI control system for soft robotic arms that enhances their adaptability and functionality in real-world environments. The system employs two types of synapses: structural synapses, which are pre-trained on foundational movements, and plastic synapses, which adapt in real-time to changing conditions. This dual approach allows soft robots to learn a variety of tasks and adjust their movements dynamically without the need for retraining, addressing key challenges in the deployment of soft robotics. The findings, published in *Science Advances*, suggest significant advancements toward the safe and intelligent operation of soft robots in assistive and medical applications.