The integration of active and passive safety systems is considered as a significant contribution towards further improvement of traffic safety. The present article describes an approach to integrate these systems. This is done by development of a novel control algorithm where force levels and activation times of an assumed adaptive restraint system are predefined based on the oncoming collision. Reference values for these force levels are generated in order to minimise the acceleration of the occupants.

The method takes into account the actual crash severity by a forecast of the acceleration behaviour of the passenger cell, based on prediction of collision speed, mass and stiffness of opponent and own vehicle. The prediction of mass and collision speed is not part of the present paper and currently under investigation. A forecast of the acceleration pulse is calculated by a simplified multi body model of the impact. The vehicle deformations are considered by non-linear springs with hysteresis. Their characteristics are derived from 53 crash tests published by NHTSA. The occupant of the egovehicle is considered also by a simplified multi body model, taking into account its mass and seating position. Optimisation algorithms determine suitable force levels and trigger times of the adaptive restraint components by minimising the acceleration of the occupant while avoiding bottoming-out of the restraint system. Currently, only straight frontal collisions with full overlap are considered. The algorithm is developed in order to provide a real-time application and is verified by detailed off-line crash simulations.

With numerical simulations several configurations with different collision severities and occupant masses were investigated. In almost every configuration significant reductions up to 90 % of the occupant acceleration were observed. The present study forms the basis of future work which includes a real-time application in a vehicle.