Lecture
This lecture introduces Data-Enabled Predictive Control (DEEPC) as a method to design controllers directly from measured input/output data, reducing the cost of design and commissioning. The instructor discusses the use of data in controller design, the structure of direct data-driven control, Willems' fundamental lemma, data-driven prediction, and the implementation of DEEPC in practice. Extensions to DEEPC, including robust and adaptive control, are explored, along with real-world experiments showcasing the effectiveness of DEEPC in controlling indoor temperature based on electrical power consumption, solar radiation, and outdoor temperature. The lecture concludes by emphasizing the importance of data in simplifying the controller process and offering a master project opportunity related to physics-inspired DEEPC.