Session 11: Toward Online Data-Driven Control: The Theory, Methodology, and System Stability
Presentation Type
Oral
Track
Methodology
Abstract
Because of the rapid development in data computation science and computer technology, online (real-time) date-driven control is becoming possible. The control strategy in the data-driven system comes from abstracting or learning from data. This so-called data-driven control with AI technology possesses distinctive advantages for nonlinear dynamical system control without a system-identified model. The data-driven method has advantages in dealing with nonlinear dynamics characteristics. High-state dimension, system nonlinearity, time-variant parameters, signal noise, uncertainties, and stochasticity are often studied in real-world application presentations. This work briefly presents the discussions on methods from model-based control (MBC) toward data-driven control (DDC) and the strategies adopted in data-driven methodology. The data-driven model-predictive control (MPC), data-driven machine-learning control (MLC), and some commonly adopted mathematical algorithms for optimization. Other discussions on specific concerns in the practice of DDC. The stability of a system with a data-driven dynamic approach is also studied as a necessary condition for an automatic (closed-loop) control system and the control function.
Keywords: Data-driven control, dynamic systems, AI, MBC, DDC, MPC, MLC
Start Date
2-7-2023 3:00 PM
End Date
2-7-2023 4:00 PM
Session 11: Toward Online Data-Driven Control: The Theory, Methodology, and System Stability
Herold Crest 253 C
Because of the rapid development in data computation science and computer technology, online (real-time) date-driven control is becoming possible. The control strategy in the data-driven system comes from abstracting or learning from data. This so-called data-driven control with AI technology possesses distinctive advantages for nonlinear dynamical system control without a system-identified model. The data-driven method has advantages in dealing with nonlinear dynamics characteristics. High-state dimension, system nonlinearity, time-variant parameters, signal noise, uncertainties, and stochasticity are often studied in real-world application presentations. This work briefly presents the discussions on methods from model-based control (MBC) toward data-driven control (DDC) and the strategies adopted in data-driven methodology. The data-driven model-predictive control (MPC), data-driven machine-learning control (MLC), and some commonly adopted mathematical algorithms for optimization. Other discussions on specific concerns in the practice of DDC. The stability of a system with a data-driven dynamic approach is also studied as a necessary condition for an automatic (closed-loop) control system and the control function.
Keywords: Data-driven control, dynamic systems, AI, MBC, DDC, MPC, MLC