Tactile Sensor Analysis during Early Stages of Manipulation for Single Grasp Identification of Daily Objects

Dexterous robotic manipulation in unstructured environments is still challenging, despite the increasing number of robots entering human settings each day. Even though provides complete solutions factories and industries, it lacks essential techniques, displaying clumsy or limited operation environments. Daily objects typically aim at hand, somatosensory system responsible for solving all complex calculations required dexterous manipulations settings. Borrowing concepts visuotactile can improve increase robotics usage In humans, finger wrist joint adjustments occur after fast identification object initial stages manipulation. Fast during those phases may performance. The present paper explores human-inspired such as haptic glance to develop single-grasp identification. This concept assist early manipulation, helping automated decision-making, type grasp position, tasks. main developed here are detecting sensor activation sample collection using signal-to-noise z-score filtering on tactile data. procedure automates touch detection reduces space classification. Experiments a daily dataset presented compelling results that will later grasping.