Combined feedforward and model-assisted active disturbance rejection control for non-minimum phase system.

Area of Science:

• Control Systems Engineering
• Non-linear Dynamics
• Robotics

Background:

• Controlling non-minimum phase (NMP) systems presents significant challenges due to inherent instabilities and sensitivity to uncertainties.
• Existing control methods often struggle with modeling uncertainties and external disturbances in NMP systems.
• Active Disturbance Rejection Control (ADRC) offers a robust framework, but conventional Extended State Observers (ESO) are ill-suited for NMP dynamics.

Purpose of the Study:

• To develop an advanced control strategy for non-minimum phase (NMP) systems.
• To enhance tracking performance and disturbance rejection capabilities.
• To propose a novel observer suitable for NMP system dynamics.

Main Methods:

• A combined feedforward and model-assisted Active Disturbance Rejection Control (MADRC) strategy is presented.
• A model-assisted Extended State Observer (MESO) is designed based on the nominal observable canonical form, overcoming limitations of conventional ESO for NMP systems.
• The feedforward controller ensures minimum settling time with an undershoot constraint, while MADRC compensates for unknown disturbances and model uncertainties.

Main Results:

• The convergence of the proposed MESO is rigorously proved in the time domain.
• Frequency domain analysis confirms the stability, steady-state characteristics, and robustness of the closed-loop system.
• The proposed strategy features a single tuning parameter (MESO bandwidth), simplifying implementation and allowing for prescribed robustness levels.

Conclusions:

• The model-assisted ADRC strategy demonstrates significant efficacy in controlling complex NMP systems.
• The developed MESO provides a viable solution for state estimation in NMP systems.
• This approach offers a promising direction for advanced control of systems with challenging dynamics.