A High Performance Disturbance Observer

This paper proposes a novel Disturbance Observer (DOb), termed the High-Performance Disturbance Observer (HPDOb), which achieves more accurate disturbance estimation compared to the conventional DOb, thereby delivering significant improvements in robustness and performance for motion control systems. The conventional and proposed DObs are analysed and synthesised in the discrete-time domain to fully capture their dynamic behaviours in real-world applications. This allows us to adjust the stability and performance of DOb-based digital robust motion control systems more effectively in practice. The conventional DOb is synthesised by assuming that the internal and external disturbances of a servo system does not change within the sampling period. Despite its simplicity, the constant disturbance model is impractical in many motion control applications. To address this issue, a more realistic disturbance model is utilised in the synthesis of the HPDOb, which is carried out in two steps. In the first step, a predictor is employed to estimate the disturbance change within the sampling period. The predictor can be implemented using the conventional DOb. In the second step, the predicted disturbance change is substituted into the high performance observer's dynamic model so that disturbances are estimated more accurately. Simulation results demonstrate that the proposed HPDOb significantly enhances the accuracy of disturbance estimation. This approach benefits a wide range of motion control applications by providing more accurate disturbance estimates to motion controllers, thereby enhancing both the robustness and performance of position control and sensorless force control systems.