A cooperative training course between
Fraunhofer IAIS and SGS-TÜV Saar
The safety of vehicle functions has now become a high priority in the established development processes of the automotive industry. However, the use of AI in safety-critical applications in particular is associated with many challenges. The procedure established in classic systems for building a safety argumentation is not directly transferable. The compact seminar provides an introduction to the problem and gives an overview of the state of standardization and technology.
This training course will give you an overview of the connections between the safety of vehicle functions and artificial intelligence in the automotive sector and provide you with a good introduction to the safety of automated driving.
Dauer: 1 day
Course content
Functionality, challenges and solution strategies for the use of AI in safety-critical applications
- Introduction to the basic functioning of AI
- Challenges posed by the data-driven approach of machine learning with regard to generalizability
- Solution approach of a basic procedure for building an evidence-based safety argumentation
Status of standardization
- Overview of standards for functional safety and AI in the automotive sector
- Introduction to functional safety according to ISO 26262 and ISO 21448 (SOTIF) for the realization of highly automated driving functions
- Activities of the standards situation for the safety verification of AI
Specification of the operational design domain
- Definition & delimitation of the areas of application - Operational Design Domain (ODD)
- Overview of existing approaches and initiatives for the specification of the ODD (OpenDrive, OpenScenario)
Safeguarding AI
- The challenge of robustness, uncertainty assessment and transparency of neural networks
- Selected methods for safeguarding AI functions
Target group
- Technical experts and managers, decision-makers
- Newcomers and career changers in the fields of safety and/or AI
Prerequisites
- Previous technical knowledge of development in the automotive sector is advantageous
- Interest in AI and automotive safety is sufficient