Adaptive Inverse Dynamics Control Of APHAN To M Haptic Device

Abstract—The PHANToM haptic device is being commonly used in haptics-based master-slave teleoperation research. The dynamics of the PHANToM robot vary significantly depending on how it is mounted and whether a sensor or tool is attached at its endpoint. Moreover, it has been shown that fixed impedance-reflecting controllers cannot preserve master-slave position tracking when the environment impedance changes. In this paper, a neural network is used to implement adaptive inverse dynamics control of a PHANToM haptic device. Experimental results show that the neural network controller successfully represents the inverse dynamics of the PHANToM and can adapt to changes in the dynamics to maintain master-slave tracking.