Abstract

Control systems are increasingly complex, often at the point that obtaining a model for them is out of reach. In some situations, (parts of) the systems are proprietary, so that the very equations that rule their behaviour cannot be known. On the other hand, the ever-growing progress in hardware technologies often enables one to retrieve massive data, e.g. from embedded sensors. Because of these evolutions, control theory is moving from a model-based towards a model-free and data-driven paradigm.

For Linear Time-invariant systems, classical results from Identification theory provide a rather straightforward approach. However, these approaches become useless (or at least inefficient) if one relaxes the strong assumptions they rely upon (linearity, gaussian noise, etc.). This is especially the case in safety-critical applications, where one needs guarantees on the performance of the obtained solution.

Despite these difficulties, one may sometimes recover firm guarantees on the behaviour of the system. This may require to change one’s point of view on the nature of the guarantees we ask. I will provide examples of such results for different control tasks and different complex systems, and will raise the question of theoretical fundamental barriers for these problems.

The seminar will be held in the Control Lab BN4 -85, Department of Engineering, and online (zoom): https://us06web.zoom.us/j/87986687566?pwd=MGJScmMwd2lwT0tVMHNmWmxSa05XZz09