Intelligent built-in torque sensor for harmonic drive systems

Abstract

The unique performance features of harmonic drives, such as high gear ratios and high torque capacities in a compact geometry, justify their widespread industrial application especially in many electrically actuated robot manipulators. In many robotic control strategies it is assumed that the actuator is acting as a torque source, and in order to implement such algorithms it is necessary to accurately measure the transmitted torque by the harmonic drive. In this paper a built-in torque sensor for harmonic drive systems is developed and examined in detail, in which strain-gauges are directly mounted on the harmonic drive flexspline. To minimize sensing inaccuracy, four Rosette strain gauges are used employing an accurate positioning method. To cancel the torque ripples, the oscillation observed on the measured torque and caused mainly by gear teeth meshing, Kalman filter estimation is used. A simple fourth order harmonic oscillator proved to accurately model the torque ripples. Moreover, the error model is extended to incorporate any misalignment torque. By on line implementation of the Kalman filter, it has been shown that this method is a fast and accurate way to filter torque ripples and misalignment torque. Hence, the intelligent built-in torque sensor is a viable and economical way to measure the harmonic drive transmitted torque and to employ that for torque feedback strategies.