Memory-Based Passivation Approach for Stable Haptic Interaction

Abstract

Inspired from the graphical interpretation of passivity in a position versus force xy graph, this paper proposes a new concept of passivation method to increase the dynamic range of impedance in which a haptic interface can passively interact. The position versus force graph represents the energy behavior in a one-port haptic interface while the haptic interface is traveling in (pressed) or out of (released) a virtual environment (VE). It interprets that to make the one-port system passive, it is sufficient to bound the releasing path below the pressing path in the position versus force graph. To realize the bound, the computed force output from the VE is saved into a designated memory, addressed by current position, while the haptic interface is pressed. Thereafter, the saved force can be reused to upper-bound the releasing path below the saved pressing path while the haptic interface is released. The proposed method is generalized from preliminary work [9] by including interaction with a moving virtual object and multi-DOF extension, together with more systematic and detailed explanations about the methods. In particular, we reconfigure the one-port moving virtual object as a two-port haptic controller and a connected one-port moving inertia. This reconfiguration allows implementation of the preliminary approach including the interaction with moving virtual objects by using an error versus force passivity instead of a position versus force passivity. This two-port extension also allows generalization of the proposed method to multi-DOF interactions by introducing virtual proxy as a virtual moving object. The proposed method is tested with a single and a multi-DOF haptic interface and shows better performance than the recently proposed field-programmable gate array-based time-domain passivity approach [12].