An active learning approach to the falsification of black box cyber-physical systems

Abstract. Search-based testing is widely used to find bugs in models of complex Cyber-Physical Systems. Latest research efforts have improved this approach by casting it as a falsification procedure of formally spec- ified temporal properties, exploiting the robustness semantics of Signal Temporal Logic. The scaling of this approach to highly complex engi- neering systems requires efficient falsification procedures, which should be applicable also to black box models. Falsification is also exacerbated by the fact that inputs are often time-dependent functions. We tackle the falsification of formal properties of complex black box models of Cyber- Physical Systems, leveraging machine learning techniques from the area of Active Learning. Tailoring these techniques to the falsification prob- lem with time-dependent, functional inputs, we show a considerable gain in computational effort, by reducing the number of model simulations needed. The effectiveness of the proposed approach is discussed on a challenging industrial-level benchmark from automotive.