A High Performance Tracking Control Method Based on a Disturbance Observer With Parameter Adaptation

Abstract

A disturbance observer (DOB) is a useful control algorithm for systems with uncertain dynamics, such as nonlinearity and time-varying dynamics. The DOB, however, is designed based on a nominal model, and its stability is sensitive to the magnitude of discrepancy between a controlled system and its nominal model. Therefore, to increase the stability margin of the DOB, it requires an accurate model identification, which is often difficult for nonlinear or uncertain systems. In this paper, the parameters of the nominal model are continuously updated by a parameter adaptation algorithm (PAA) to keep the model discrepancy small, such that the DOB is able to show its desired performance without losing stability robustness even in the presence of nonlinearity and/or time-varying dynamics. In the integration of the DOB and the PAA, however, there exists a complicated signal interaction. In this paper, such interaction problem is solved from a practical point of view; signal filtering. The proposed method shows improved performance for an electric motor system, and is verified by experimental results in this paper.

Parameters self‐turning controller fused with an adaptive nominal model for disturbance observer: Application to direct drive manipulator with significant payload changes

“…Hyun et al provided a parameter adaptation algorithm (PAA) to continuously update the parameters of the nominal model, aiming to maintain a small model discrepancy. It was verified that the proposed method achieved the desired performance without compromising stability robustness, even in the presence of nonlinearity or time‐varying dynamics 28 . Chen et al introduced a nonlinear disturbance observer (NDOB), which is dependent on the maximum velocity and physical characteristics of the robot manipulator.…”

“…It was verified that the proposed method achieved the desired performance without compromising stability robustness, even in the presence of nonlinearity or time-varying dynamics. 28 Chen et al introduced a nonlinear disturbance observer (NDOB), which is dependent on the maximum velocity and physical characteristics of the robot manipulator. This approach overcomes the limitations of traditional DOB using linear system techniques to design the nominal model.…”

“…These disturbances are estimated by comparing the output deviations of the plant with a nominal model, and they are subsequently compensated at the input to mitigate their effects. At present, DOB has been widely used in the controller design of the traditional manipulator, and the results are satisfactory 24–29 . In reference, 24 the impact of other joints and external forces on the manipulator is regarded as the overall system interference.…”

“…At present, DOB has been widely used in the controller design of the traditional manipulator, and the results are satisfactory. [24][25][26][27][28][29] In reference, 24 the impact of other joints and external forces on the manipulator is regarded as the overall system interference. A DOB with a fixed nominal model is employed at each joint to eliminate the effects caused by external forces and joint coupling.…”