Funded by ProFIT
FuMaMa 2019-2021: Fuzzy Measures Management
Project Partner
Model Engineering Solutions GmbH (MES) Go to About Us Page of MES
TU Berlin Field of Expertize Quality Science Go to Partner Website
Lead partner: Prof. Dr. Roland Jochem Get more Infos
Starting Point: Why Quality Alone Is Not Enough?
Using one method or procedure alone does not ensure lasting quality. High-quality work results from constant improvement and adaptation. This mindset is reflected in the PDCA model:
- Plan
- Do
- Check
- Act
Software development in the automotive industry follows ISO 26262-6, as the developed software handles safety-critical tasks. For example, software designed for steering control or a braking system is safety relevant. A software failure in these areas could lead to injury or serious accidents. That is why ISO 26262-6 defines strict quality assurance steps for development of such software products.
However, there is a challenge:
- Many tools and manual reports exist
- Data is spread out and hard to evaluate
- Developers lack a clear, unified view
Therefore, automotive software manufacturers seek a central view of the software development process to identify countermeasures efficiently.
Objectives and Benefits
The motivation of the cooperation project aimed at holistic quality management within the framework of the PDCA model. For continuous improvement, it was essential not only to systematically measure and check quality but also to improve it effectively and efficiently.
The main focus of the project was on the automation of measure determination. Based on the identified compliance status in a project, the comprehensive set of options was used to define those actions that would produce compliance efficiently and effectively. For this purpose, a to-do-list generator was developed that proposed suitable countermeasures based on the fuzzy set theory for given distributions of the quality parameters as well as their aggregation by the quality model. Beyond mere assessments, countermeasures were also identified and the PDCA cycle was fully implemented. This corresponded to the high practical demand for efficient, high quality software development.
The recommendation system was designed for specific countermeasures in such a way that potential users who lacked prior knowledge or in-depth expertise could systematically and objectively initiate and implement improvements.
Automotive software manufacturers are looking for a holistic view of ISO compliance and automated generation of countermeasures. Therefore, the recommendation system is integrated into the tool MES Quality Commander. In this way, necessary steps for ISO 26262-6 compliance, such as identifying countermeasures, can be ascertained automatically.