Analysis and design of aperiodic sampling in control systems

This thesis addresses data-driven analysis as well as prediction-based design of aperiodic sampling in control systems. In many modern control applications, the traditional view of sampling and controlling in a periodic fashion must be abandoned due to varying transmission delays, data loss or uneven computational latencies. The effect of these undesired aperiodicities on the closed loop must thus be analyzed. In addition, a targeted aperiodic design of the sampling pattern, e.g. via event-triggered control (ETC), often allows for a better usage of bandwidth than periodic concepts.
We consider both perspectives in this thesis. In particular, first we provide methods to analyze stability of control systems under arbitrary aperiodic sampling directly from recorded data. Previous approaches to this end required model knowledge, which might be challenging to obtain. Second, we propose a prediction-based approach for design of the aperiodic sampling pattern, called rollout ETC. This approach allows for predictability of the required communication resources and a guarantee of improved sample efficiency, which were both difficult to address in existing ETC schemes.