A model-independent force observer for teleoperation systems

Transparent implementation of bilateral teleoperation or haptic controllers necessitates the measurement of operator hand force. This requires the use of expensive commercially available 6 degree-of-freedom (DOF) force/torque sensors. Since the bandwidth of human operator force output is reasonably low, observers can be used for hand force estimation. The drawback of conventional force observers is their dynamic lag and the need for exact knowledge of the dynamic model of the haptic device. As an alternative, a novel model-independent force observer (MIFO) is proposed in this paper, in which a multilayer perceptron neural network (MLPANN) is utilized and trained for force estimation. The performance of the proposed observer is verified on a 1-DOF experimental setup.