Technical University of Darmstadt – FZD
FZD brings, to PEGASUS, its experiences and previous works in the field of research, regarding functions and test procedures for the vehicle automation as well as active sensors for the automotive field. The following projects provide an overview: PRORETA I to III: integral safety and maneuver-based driving; Conduct-by-Wire: serial assistance in automobile with a maneuver interface for the vehicle guidance; the Darmstadt procedure (EVITA) for the testing and assessment of frontal collision prevention measures; Ko-FAS – Ko-KOMP; realization and cost-benefit analysis of a protection concept based on the cooperative sensor technology; objective assessment of the controllability of non-situation, appropriate reactions of environment-assisted driver assistance systems.
FZD is working thereby on the three research areas of criticality metrics and assessment of the human performance capability, development of alternative testing methods – decomposition approach, development and validation of simulation models for active sensors.
CRITICALITY METRICS AND ASSESSMENT OF THE HUMAN PERFORMANCE CAPABILITY (PP1)
are automatically evaluated from the field studies. The metrics that describe the criticality of an individual scenario are developed with the assistance of FZD. The performance capability regarding automation will be compared in the next step, with the human one. The determination of the human performance capability is thereby also part of the FZD content.
DEVELOPMENT OF ALTERNATIVE TEST METHODS – DECOMPOSITION APPROACH (PP2)
As amendment to this scenario-based approach, the potential of alternative test methods for the reduction of the release expenditures will be demonstrated. For this, the approach of the functional decomposition of the vehicle guidance will be applied to partial functions of the highway chauffeur. This approach will be assessed, while the partial functions of the highway chauffeur, as well as the scenario-based and the decomposition approach for the test, will be used and compared.
Figure 1 shows possible levels for the decomposition of the driving skills. The challenge regarding the implementation of the decomposition approach lies in the definition of existent criteria for the particular tests on each respective level and the proof of their independence.
DEVELOPMENT AND VALIDATION OF SIMULATION MODELS FOR ACTIVE SENSORS (PP3)
The real traffic offers a larger diversity and a high complexity of possible traffic scenarios. The release of highly-automated driving functions, through traditional test drives, is therefore, neither efficient, nor economical. Because of this, in addition to the real testing, there is the need of additional virtual test methods, in order to reduce the complexity of safety verifications. This is why high-performance models of vehicle, environment and driving function are necessary. FZD is especially involved in the development and validation of models of active environmental sensors. Hereby, the conflict of objectives between the model accuracy and computing time is resolved. Since the application is within the scope of the PEGASUS, in the virtual driving test and X-in-the-loop simulation, real-time capability is a basic prerequisite.