All about Data-driven control.
Since the early 1990’s a variety of control design methods that are constructed directly upon batches of input-output data collected from the process to be controlled have appeared in the literature. These methods contrast with model-based control design mainly in two fundamental aspects: they are not based on the knowledge of a process model and they do not intend to freely determine the controller’s transfer function. Instead, they make direct use of the information carried by the measured data in order to adjust the numerical parameters of a controller whose transfer function has a previously specified and fixed structure. Accordingly, these methods became known as data-based, in opposition to model-based, or alternatively as data-driven methods. Closely related is the area of identification for control, which has also received an important boost during this same time frame. This website’s aim is to provide a forum for the core concepts and the novelties in this field of research and to disseminate our modest contributions to it.