Performance Analysis of Model-Based Functional Identification on Modified Measuring Instruments

Modern measuring instruments are increasingly de-fined by complex software and simple hardware sensors. Currently established mechanisms to identify such systems, however, involve manual labor to verify version numbers and hashes over executable code, which is time-consuming. By ensuring identical functional behavior between a certified prototype and devices in the field, it is possible to realize remote automatic quality control of measuring instruments, even if version numbers or hashes differ. As such, the manufacturers of instruments could introduce minor patches or bugfixes without the need of reidentifying and recertifying the complete updated devices. In this paper, we present an approach that enables quality control of updated measuring instruments based on the
$L_{M}^{*}$
algorithm. With the algorithm, we infer a model that describes the functional behavior of an instrument by means of active automata learning. We analyze the performance of our approach with different test cases. It is envisioned that with our approach, we can pave the way towards realizing fully automatic quality control for software-driven measuring instruments.