The Quality Assurance Methodologies for autonomous Cyber-Physical Systems (QAMCAS) aims at methods for enabling substantial quality improvements of interacting communicating systems that interact with humans and the physical environment. QAMCAS is intended to carry out research improving quality assurance methods during development together with methods that assure quality criteria like safety as well as reliability and also robustness during operation. The latter deals with fail-operational methodologies based on artifacts obtained during development and also measurements gained from previous similar systems during operation. In QAMCAS we treat quality assurance from a holistic point of view investigating methods to be used at development time as well as methods to be applied during operation of the cyber-physical system. To solve the challenges corresponding to the holistic view, we suggest to integrate testing methodologies like combinatorial testing and model-based testing and to combine them with machine learning approaches for model and test data generation. Furthermore, we carry out research for transferring development artifacts like models to fail-operational systems that follow the model-based reasoning paradigm. For this purpose, we have to work on smart monitoring systems that are capable of identifying failures and triggering fault localization and repair procedures for obtaining truly fail-operational systems. Although, the main ingredients are available, their integration is challenging and requires several research issues to be solved. In this proposal we discuss these issues in detail. In order to show that the proposed methods and techniques contribute to quality assurance of cyberphysical systems, we carry out the development of prototypical implementations that are further on used for providing an experimental evaluation in the context of autonomous driving and other mobile and autonomous systems.
|Effective start/end date||1/10/17 → 30/09/24|
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.