Abstract
Accurately modeling of many practical contact scenarios such as robotic grasping and assembly is a challenging problem because of complex contact geometry and surface uncertainties. This paper presents a new hybrid contact model (HCM), which combines a traditional physical model and a data-driven model to make a more accurate description of a contact dynamics phenomenon. When a physical model is employed to describe a complex contact case, it usually has error from an experimental data measured from the contact case to be modeled because of inevitable unmodeled and/or unknown factors. The data-driven model is used to represent this error, which is an artificial neural network model trained from experimental data using a machine learning technique. A bouncing ball example is presented to demonstrate the feasibility of the presented approach.
