Robustness and performance trade-offs in torque control of robots with harmonic drive transmission

This paper focuses on designing torque control laws for robots equipped with harmonic drive transmissions. A nominal linear model of the joint is first identified from input-output experimental tests. Subsequently, by varying the input signal amplitude level, a set of models, incorporating the effect of nonlinearities in the system, can be extracted. The differences between the nominal model and this set are formulated as uncertainty bounds for control design purposes. Utilizing the uncertainty bounds, an H_∞-based optimal controller is designed. Experiments are performed for different uncertainty levels on the IRIS facility (a versatile, modular and reconfigurable prototype robot developed at the Robotics and Automation Laboratory of the University of Toronto)