Optimizing fault injection in FMI co-simulation through sensitivity partitioning

As society and industry relies extensively on Cyber-Physical Systems (CPS), any malfunctions can have unforeseen catastrophic failures. Fault Injection (FI) techniques perturb a model of a CPS with the intention of causing a failure and measuring the robustness of the CPS. Naturally, the success of a FI simulation depends on three factors: (i) the realism of the faults injected; (ii) how quickly the faults cause catastrophic failure; and (iii) the fidelity of the model used. This paper proposes to improve the success rate of FI studies by addressing each one of these factors. An algorithm is presented that leverages traditional sensitivity analysis in hybrid systems to reduce an uncountable fault search space to a optimal finite set (factors and we use co-simulation as the model integration technique (factor iii). We evaluate our contribution on the power window system developed by MathWorks®.

Jaar van publicatie:2019

Trefwoorden:P1 Proceeding

BOF-keylabel:ja

Authors from:Government

Toegankelijkheid:Closed