Subproject 1: Scenario analysis and quality measures
The subproject “Scenario analysis & quality measures”, through the example application highway-chauffeur, defines methods for the derivation of relevant traffic scenarios, assesses the human and machine performance, as well as the criteria and measures of their assessment, in order to answer to the question “how good is good enough”.
Highly-automated driving functions will be used by our company. They help with the prevention of accidents, reduction of fuel consumption and emissions, as well as maintain mobility in old age. However, they do present technical and physical limits. Not every accident can be prevented in the future, through the (highly) automated driving functions (HAF). As it happened with the implementation of the seatbelt and airbag, new types of accidents will occur, due to the highly-automated driving functions. Knowing which situations an automated road vehicle can reliably handle is essential for the HAF product series, as a technical innovation is rejected, without individual and social acceptance. This leads to requirements for the technical system, which are necessary for the testing and release prerequisite. The subproject will answer these basic questions; the goals:
- Development of methods and tools for the definition of design criteria for highly-automated driving functions,
- Demonstrating these methods and tools using specific example functions ,
- Deriving functional requirements for the sample function,
- Transferring these requirements to the subproject 2 and 3 as work basis.
Therefore, an example application will be described in detail: the highway chauffeur. Subsequently, for this individual application, critical traffic situations will be determined, e.g., with the help of accident data and real driving studies, underpinned by driving simulator studies. Standard situations and critical scenarios are entered into a test specification database, which allows for the systematic test runs in the simulation as well as on the test site. In these critical situations, both the human and machine performances are assessed.
To ensure the adequate social acceptance, it must be ensured that the technical system is able to control the vehicle, at least as good as a human driver, in all possible realistic driving events. The performance of a human driver in such situations represents therefore a minimum necessary machine performance. In order to determine this, a procedure will be developed. Based on the highway-chauffeur, consequently, an extended scenario will be considered, in order to ensure the transferability of the results and procedures for further application scenarios.