Soft Robotic Finger

Soft robotic fingers aim to replicate the dexterity and adaptability of human fingers for robotic manipulation, focusing on safe interaction with delicate or irregularly shaped objects. Current research emphasizes developing accurate and computationally efficient models of finger deformation, often employing physics-informed neural networks or NARX neural networks integrated with embedded sensors like optical waveguides, to improve state estimation and control. These advancements, coupled with innovations in tactile sensing (e.g., GelSight integration) and proprioceptive feedback mechanisms, are enhancing the capabilities of soft robotic grippers for applications ranging from underwater manipulation to in-hand object manipulation.