Soft microtubular sensors as artificial fingerprints for incipient slip detection

Abstract

Incipient slip detection constitutes a crucial aspect of adaptive grasping and dexterous manipulation in robotics. The primary challenge lies in the subtle nature of incipient slip across temporal, spatial, and force dimensions. This work reports a soft robotic finger capable of accurately detecting incipient slip using artificial fingerprints composed of two piezoresistive microtubular sensors. Experimental results reveal distinctive peak patterns in the sensing signals during incipient slip on smooth and rough surfaces. For smooth surfaces, the direction of slip can be determined by the opposite changing trends in the sensing signals. Finite element analysis elucidates that the underlying mechanisms are driven by the asymmetric local geometry around the sensors when sliding on a smooth surface, and by the relative position of the sensors to the surface micro-structure when sliding on a rough surface. A customized program is then developed for real-time incipient slip detection based on peak recognition in de-noised rolling windows. The feasibility of this method is demonstrated through the adaptive grasping of deformable, moving, and weight-unknown objects using a robotic hand integrated with the soft tactile fingers.