New vehicles are increasingly equipped with a variety of Advanced Driver Assistance Systems (ADASs). As these systems have the potential to prevent accidents, accidents of the future will differ from those of today. Predicting the type and characteristics of these future accidents is therefore essential to current research and development in the occupant restraint and new ADAS fields.

In this study, accident avoidance of 15 ADASs was modelled using simple deterministic rules for each, creating both a conservative and an optimistic ruleset to account for current limitations and future possibilities. The rulesets were applied to the US National Automotive Sampling System Crashworthiness Data System data from 1995-2015 and verified through the literature. The residual passenger vehicle to passenger vehicle accidents were analysed, treating all accidents and accidents with at least moderate injuries in modern passenger vehicles (model year 2007 and later) separately.

Many accidents were found to be avoided through such systems, and their combined effectiveness ranged from 51% to 97% depending on ruleset. Electronic Stability Control (ESC), Lane Keep Assist (LKA), and Crossing and Rear End Automated Emergency Braking (AEB) were highly effective, individually preventing over 25% of accidents in the optimistic calculation. Importantly, remaining accidents will have a different distribution across accident types compared to today:​ rear end collisions will reduce, leaving turning and crossing scenarios to dominate future accidents.

For passenger vehicle to passenger vehicle accidents with at least moderate injuries in modern vehicles, four accident types alone were found to account for 93% of all remaining accidents in the optimistic estimate:​ Head On, Turn Across Path, Turn Into Path Opposite Direction and Straight Crossing Paths;​ the latter three are intersection-related and together represent three quarters of all remaining accidents.

The intersection accidents are analysed further for deformation pattern, impact direction, 90% cumulative delta velocity and injured occupant position in order to identify possible new impact conditions to be used when evaluating occupant restraints. The well-established frontal and side impacts will still generate many AIS2+ injuries, while new more oblique impact conditions will also be needed to represent the variety of intersection accidents remaining.

The description of future accidents and impact conditions presented here can serve as a basis for the research and design of future ADASs and occupant restraints. We propose virtual assessment methods with Human Body Models (HBM) based on these impact conditions.