The precise and reliable prediction of vehicle movements based on information from environmental sensors such as radar, camera or LiDAR is an essential constituent of future pre-crash safety functions triggering irreversible actuators. These predictions require novel motion models that go beyond the state-of-the-art and the comparatively low requirements regarding accuracy in today's AEB systems. Because of this, methods for existing driver assistance systems are not suitable for the prediction of crash parameters in inevitable crash situations and the subsequent activation of irreversible systems with ASIL D classification like prospective smart airbags.

The proposed highly accurate inevitability model allows the integration of a specially adapted motion model for worst-case assessment within the physical limits. The focus is on integration into pre-crash functionality with near-field environment models in order to provide reliable information of the expected crash constellation to subsequent crash severity estimation. An adapted single-track model describes the nonholonomic vehicle behavior. Physically possible trajectories for crash avoidance are calculated using Kamm circle as the limitation for the vehicle accelerations. This model is modified to minimize the dependency on the unavailable and hence only roughly estimated vehicle parameters of the collision partner. Innovative concepts for crash severity estimation require detailed information on the expected collision parameters (location, angle, velocities) as well as the estimated times to the inevitable collision.

Depending on the application and the traffic scenario, this information can then be categorized into collision classes. These classes can be defined according to standard crash test scenarios (ODB40, AZT, etc.).
Particularly interesting examples for the inevitability model are critical traffic situations with crossing vehicles. This paper focuses on the effects of different scenario constellations in the urban to rural speed range, on the inevitability model, and the subsequent crash severity prediction. Therefore, heat maps are presented as a suitable tool for the evaluation of the scenarios. They provide a simple illustration of how critical the unavoidable collision is as well as the functional requirements of pre-crash systems in different situations.

The proposed model is the groundwork for further developments and investigations towards a holistic methodology for pre-crash safety systems. The prototype application in a CARISSMA research vehicle is currently in preparation. In addition, the presented methodology will also be extended to investigations of sensor requirements and system tolerances. This enables the comprehensive analysis of sensor systems and algorithms, which are essential components of future irreversible safety systems.

Mr. Robert Lugner, CARISSMA Institute of Safety in Future Mobility, TH Ingolstadt, GERMANY; Mr. Maximilian Inderst, CARISSMA Institute of Safety in Future Mobility, TH Ingolstadt, GERMANY; Mr. Gerald Sequeira, CARISSMA Institute of Safety in Future Mobility, TH Ingolstadt, GERMANY; Mr. Kilian Schneider, CARISSMA Institute of Safety in Future Mobility, TH Ingolstadt, GERMANY; Prof. Dr.-Ing. Thomas Brandmeier, CARISSMA Institute of Safety in Future Mobility, TH Ingolstadt, GERMANY