Some Ideas On Sampling Strategies For Data-Driven Inference Of Passivity Properties For Mimo Systems

We study the problem of determining whether, and if so to which extent, an input-output system is passive. While in many applications the input-output relation is undisclosed, input-output data are available. Therefore, we present an iterative approach that determines the input feedforward passivity index and the shortage of passivity of discrete-time linear time-invariant systems from input-output data. Besides the general multiple-input multiple-output framework, we consider the interesting case of interconnected dynamical systems over graphs and show that certain assumptions on the communication topology lead to a reduction of the number of required (numerical) experiments.