Inertial Properties in Haptic Devices: Non-Linear Inertia Shaping vs. Force Feedforward

Abstract

The inertia of haptic devices limits the user's manipulation and dynamically couples the Cartesian motion, which influences the transparency and fidelity in haptic feedback. By employing force-torque sensing, we investigate two approaches to reduce the apparent inertial effect of haptic devices and to overcome dynamic coupling. First, in order to shape the apparent inertia felt by the user, non-linear inertia shaping (NIS) is presented and introduced to the field of haptics. NIS is based on non-linear dynamic decoupling (NLD). Second, as a standard approach, force feedforward control (FF) is presented that uniformly scales down the apparent inertia. We demonstrate that FF is a special case of NIS, under the assumption that gravitational, centripetal and Coriolis terms are neglected. Simulations and experiments were conducted on DLR's bi-manual haptic device HUG. It is shown that NIS is suited to compensate for the coupling effects, while FF can reduce the apparent inertia more effectively.