Sensing and Suppressing Abrupt Disturbances for Robot Manipulators via Supervisory Disturbance Observer

Abstract

Abrupt disturbances in robot manipulators can not only significantly degrade tracking control performance, but also reduce system stability. However, it is difficult to directly sense abrupt disturbances and suppress their effects in practice. In this article, a supervisory disturbance observer (SDO) is proposed to estimate abrupt disturbances and suppress their effects by using only joint position sensor data. The proposed SDO consists of a supervisory strategy and two extended disturbance observers (EDOs) with different bandwidths. The high-bandwidth EDO is used for accurately and rapidly estimating abrupt disturbances, while the low-bandwidth EDO is used to suppress the measurement noise in sensor data. The supervisory strategy is designed to combine the outputs of the two EDOs as the final estimated disturbance, based on disturbance change point detection and disturbance effect analysis. Moreover, the virtual measurement of disturbance is introduced in the supervisory algorithm to quickly detect the disturbance mutation time. As a result, the SDO is capable of quickly sensing abrupt disturbances while suppressing the effects of measurement noise. Furthermore, the information from SDO is feedforward to compensate for the effects of disturbance. The experimental results are presented to validate the effectiveness of the proposed method.