Data-driven haptic modeling using polynomial hypersurfaces

In this paper we present a method for generating accurate haptic force models to describe the feel of physical devices. The models are generated by fitting polynomial hypersurfaces to experimentally measured motion and force data obtained by actuating physical devices. The methods are demonstrated by modeling the complex force-deflection profile of a nonlinear mass-spring-damper system. Initial results indicate that polynomial hypersurfaces have the potential to accurately describe the dynamics of nonlinear devices.