STGEM is a tool for black-box testing of cyber-physical systems. It supports falsification of requirements described in Signal Temporal Logical (STL) by robustness optimization. This is achieved by training a generative machine learning model online to produce system inputs that yield a low robustness.
STGEM is under development and we are adding new algorithms and features. So far, the tool implements the algorithms presented in these articles:
-
Jarkko Peltomäki and Ivan Porres, Falsification of Multiple Requirements for Cyber-Physical Systems Using Online Generative Adversarial Networks and Multi-Armed Bandits. The 6th. Intl. Workshop on Testing Extra-Functional Properties and Quality Characteristics of Software Systems, ITEQS 2022. To appear. Preprint available here.
-
Jarkko Peltomäki, Frankie Spencer and Ivan Porres, Wasserstein Generative Adversarial Networks for Online Test Generation for Cyber Physical Systems, The 15th Intl. Workshop on Search-Based Software Testing, SBST 2022. To appear. Preprint available here.
- We participated in the SBST 2022 CPS tool competition. A preprint of the competition note is available here.
- Related presentations videos here ( slides ) and here ( slides ).
Read the CONTRIBUTING.md file to learn how to set up stgem.
STGEM is developed at Åbo Akademi University. Contact Ivan Porres ivan.porres@abo.fi for more information.