Toward automated haptic modeling using commercial haptic interfaces: surface normal estimation and static model identification

The use of commercial haptic interfaces to create models of the feel of physical objects has the potential to transform the field of haptic modeling by broadening the availability of haptic modeling technology. In this paper we develop methods that enable general haptic interfaces to actuate and measure the dynamic properties of physical objects and generate models of their behavior. Specifically, methods are presented that allow commercial haptic interfaces to 1) estimate interaction forces without the use of force sensors, 2) estimate surface normal directions, and 3) actuate objects and identify haptic models from experimental data. Results are presented for two 1-D push buttons, one foam ball, and one inflatable vinyl ball. Initial results suggest that the methods have promise, but challenges must be overcome related estimating forces without the use of force sensors.