On Input Generators for Cyber-Physical Systems Falsification

Journal article, 2024

This method generates input signals to a simulation of the system under test and uses quantitative semantics that plays the role of objective functions to minimize the distance to falsify a specification. This paper presents and evaluates differently structured parameterizations of input generators: pulse, sinusoidal, and piecewise with different interpolation signals. The input generators are compared based on their results on benchmark examples, as well as coverage measures in the space and time, and frequency domains. Input generators allow covering many different input signals in a single falsification problem, which is especially useful for industrial practitioners wanting to use falsification in their daily development work. Falsification is a testing method that aims to increase confidence in the correctness of cyber-physical systems by searching for counterexamples guided by an optimization algorithm. This method generates input signals for a simulation of the system under test and employs quantitative semantics, which serve as objective functions, to minimize the distance needed to falsify a specification. This paper introduces and evaluates various parameterizations of input generators, including pulse, sinusoidal, and piecewise signals with different interpolation techniques. The input generators are compared based on their performance on benchmark examples, as well as coverage measures in the space-time and frequency domains. Input generators enable the exploration of numerous different input signals within a single falsification problem, making them particularly valuable for industrial practitioners who wish to incorporate falsification into their daily development work.