Abstract The validation of the development of driver assistance functions over millions of driven test kilometers finds its limits already today. The systems evolve rapidly from pure assistants over automated driving maneuvers towards fully autonomous driving. In the process an explosion of complexity is generated through interconnection and addition of sensors as well as strongly increased model complexity to represent the environmental aspects relevant for driving dynamics better and timelier. Here we would like to present an idea how an effort reduction of the validation can be realized. We propose to start the validation of intelligent systems before the road test: Features that intelligent systems use to classify the environment can selectively be varied in regards of test stimuli. The sensor state space is far easier to handle here than in the road test. Variations can be generated noticeably more efficient and goal oriented. In conclusion relevant parts of intelligent systems can be validated in driving dynamics relevant scenarios with less effort.
Validation of Automated Driving Functions
01.01.2016
5 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Scenario Identification for Validation of Automated Driving Functions
Springer Verlag | 2016
|Test Strategies for Efficient Validation of Automated Driving Functions
Springer Verlag | 2024
|Enhancing Radar Model Validation Methodology for Virtual Validation of Automated Driving Functions
TIBKAT | 2025
|Enhancing Radar Model Validation Methodology for Virtual Validation of Automated Driving Functions
DataCite | 2025
|Challenges for the Validation and Testing of Automated Driving Functions
Springer Verlag | 2017
|