Real Data Based Validation of Highly Automated Driving Functions Using Simulation Methods
In order to implement highly automated driving, scenarios have to be derived from recorded data and tested with different simulation methods. By combining real data with virtual tests, scenarios can be efficiently generated and tested. At the same time, this approach improves test coverage and creates a statistically sound assessment basis for fulfilling a function.
A key challenge for the approval of highly automated driving functions and autonomous driving is to ensure the functional safety of these vehicle systems. At the latest with the achievement of level 3 automation according to BASt (German Federal Institute of Roads) or level 4 of the SAE classification , conventional validation strategies for driver assistance systems reach their limits, since the driver no longer has to be at disposal as a fallback level, at least temporarily. Consequently, all conceivable situations must be controllable by the system, which means that the requirements to be fulfilled by the...