Vibration-based Full State In-Hand Manipulation of Thin Objects
Oron Binyamin, Guy Shapira, Noam Nahum, Avishai Sintov
TL;DR
The paper tackles the limitation of low-DOF parallel grippers by enabling full-state in-hand manipulation of thin objects using a Vibratory Finger Manipulator (VFM). It decouples translation and rotation, inducing cyclic center-of-mass motion about the grasp point to rotate the object, and uses duty-cycle modulation to stabilize orientation during translation. The approach is backed by analytical dynamics, finite-element modeling, and experiments, including task demonstrations like credit-card insertion and ruler alignment, showing high accuracy and practical viability. This work has potential to make thin-object manipulation affordable and robust in industrial and medical settings, while outlining future work toward dual-vibration fingers and onboard sensing for 3D capabilities.
Abstract
Robotic hands offer advanced manipulation capabilities, while their complexity and cost often limit their real-world applications. In contrast, simple parallel grippers, though affordable, are restricted to basic tasks like pick-and-place. Recently, a vibration-based mechanism was proposed to augment parallel grippers and enable in-hand manipulation capabilities for thin objects. By utilizing the stick-slip phenomenon, a simple controller was able to drive a grasped object to a desired position. However, due to the underactuated nature of the mechanism, direct control of the object's orientation was not possible. In this letter, we address the challenge of manipulating the entire state of the object. Hence, we present the excitation of a cyclic phenomenon where the object's center-of-mass rotates in a constant radius about the grasping point. With this cyclic motion, we propose an algorithm for manipulating the object to desired states. In addition to a full analytical analysis of the cyclic phenomenon, we propose the use of duty cycle modulation in operating the vibration actuator to provide more accurate manipulation. Finite element analysis, experiments and task demonstrations validate the proposed algorithm.
