In a rollover, the lap part of the 3-point belt certainly restrains the occupant from being fully ejected out of the vehicle, however the upper torso of the far side occupant may slip out of the shoulder belt. In this study the combination of reversed 3-point belt geometry (seat integrated), inboard torso side-support and buckle pretensioner were evaluated regarding the ability to better restrain the upper torso to the seat to prevent headto- interior impacts.

The method of evaluation, proposed and used in this paper, was a new sled test method simulating full-scale tripped rollovers along the longitudinal axis during the initial phase of tripping, the airborne phase and the first ground impact. The roof was assumed in the tests to be able to withstand the ground impact. Since car occupants normally are seated with a certain kyfosis and may straighten and elongate their spine, standard HIII ATDs were modified with 3D-flexible lumbar spines and used in both front seat positions.

As a result, the rollover sled test method worked properly with good repeatability. While the head of the non-leading side (far-side) dummy impacted the inner roof in the standard 3-point belt configuration, the seat integrated 3-point belt with reversed geometry and buckle pretensioner showed ability to restrain the torso from moving inboard and towards the roof during the rollover tests.