Dynamic Modeling and Analysis of Vibration-Induced Friction Modulation for Controlled Sliding in Robotic In-Hand Manipulation
Abstract: In-hand manipulation, the ability to reposition an object within a robotic hand without regrasping is a fundamental yet challenging capability in robotic dexterous manipulation. A critical bottleneck is the difficulty of achieving controlled sliding at the finger-object contact interface, primarily because friction at contact points is difficult to regulate without precise knowledge of object surface properties. This thesis investigates vibration-induced friction modulation as a strategy to enable controlled sliding for in-hand manipulation. Specifically, we develop a dynamic model of the Eccentric Rotating Mass (ERM) vibration system embedded in robotic fingers, capturing the position-dependent force amplification behavior. A parametric analysis is conducted to study the effects of vibration frequency, amplitude, and object mass on friction modulation efficacy. Simulation results from Ansys Mechanical modal analysis are validated against accelerometer measurements to confirm the velocity direction predictions. This work provides a principled modeling foundation that extends prior empirical findings and lays the groundwork for design optimization of vibration-induced variable-friction (VIVF) grippers without requiring per-design velocity field characterization from scratch.
Advisor: Professor Berk Calli
Committee: Professor Vincent Aloi, Professor Connor McCann