Data-Driven Control and Model Reduction

The field of control systems is moving towards more data-driven approaches, with a focus on developing innovative methods for estimating and controlling complex systems. Recent developments have seen the integration of machine learning techniques, such as deep learning and Koopman operator-based methods, to improve the accuracy and efficiency of control systems. Additionally, there has been a push towards model reduction techniques, which aim to simplify complex systems while preserving their essential dynamics. These advancements have the potential to improve the performance and robustness of control systems in a variety of applications, including quadrotor systems and precision motion control. Noteworthy papers include:

  • Integrating Uncertainties for Koopman-Based Stabilization, which establishes a unified framework for robust stabilization via the Koopman operator.
  • Trajectory Tracking and Stabilization of Quadrotors Using Deep Koopman Model Predictive Control, which presents a data-driven control framework for quadrotor systems using a deep Koopman operator with model predictive control.

Sources

Direct data-driven interpolation and approximation of linear parameter-varying system trajectories

Integrating Uncertainties for Koopman-Based Stabilization

Data driven feedback linearization of nonlinear control systems via Lie derivatives and stacked regression approach

Power-Series Approach to Moment-Matching-Based Model Reduction of MIMO Polynomial Nonlinear Systems

Trajectory Tracking and Stabilization of Quadrotors Using Deep Koopman Model Predictive Control

Iterative Youla-Kucera Loop Shaping For Precision Motion Control

Built with on top of