Event-based model predictive control with two-phase predictive detection
Document Type
Article
Publication Date
11-1-2024
Abstract
This paper proposes an event-based model predictive control for continuous-time systems in the presence of bounded disturbances. A two-phase predictive event detection strategy is proposed. In the first phase, a conservative time interval is calculated based on the worst possible system behavior caused by disturbances. The optimal control sequence is applied during this interval without sampling. After the first phase, periodic samplings and detections with adjustable sampling period are then executed in the second phase until finding the triggering instant. This strategy has the advantages in terms of reducing the sampling frequency and triggering rate. The maximum number of samplings can be adjusted to balance computational cost and sensing cost. Sufficient conditions are provided for recursive feasibility and closed-loop stability.
Publication Source (Journal or Book title)
Journal of the Franklin Institute
Recommended Citation
Luo, Z., Zhu, B., Meng, X., & Zuo, Z. (2024). Event-based model predictive control with two-phase predictive detection. Journal of the Franklin Institute, 361 (16) https://doi.org/10.1016/j.jfranklin.2024.107172