Risk-Based Continuous Quality Control for Software in Legal Metrology

Measuring instruments are increasingly defined by complex software while using simple hardware sensors. For such systems, software conformity between certified prototypes and devices in the field is usually demonstrated using version numbers and hashes over executable code. Legal requirements for regulated instruments could equally be satisfied if prototype and device in the field display identical functional behavior even if hashes differ. Such functional identification can give instrument manufacturers room for software patches and bugfixes without the need for recertification. Based on the L∗ algorithm, which is used to learn the language which deterministic finite automata accept, a risk-based method is proposed that realizes automatic functional identification of software to a certain extent, thereby enabling quality control of regularly updated measuring instruments without the need for frequent manual inspections. Risk assessment may be used to identify critical state transitions in monitored devices, which can be used to trigger recertifications if needed.

---

1. Thematic Tracks Regular Papers

---

2. Opracowanie rekordu ze środków MEiN, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2024).