Minimum Variance Estimation of Load-side External Torque Using Load-side Encoder and Torque Sensor

Sensorless external force/torque estimation is required for force/torque control in cooperative robots and human-support robots. Our group has proposed a novel load-side external torque observer using load-side encoder information, which can reduce the measurement noise effects and modeling error effects of the motor side and the transmission part in our previous study. Based on the fact that most of the cooperative robots in industry have torque sensors, this paper proposes a load-side external torque observer using encoders' information and torque sensor's information for robust estimation against modeling errors and measurement noises. The advantages of the proposed observer are validated by the simulations and experiments.