In the early 1980's a series of tests was conducted simulating rear-end crashes. The tests demonstrated that a conventional automotive bucket seat adequately retains an unbelted dummy on the seat for rear-end impacts up to 6.4 m/s and 9.5 g severity. For this severity of impact the total rearward rotation of the seatback is less than 60° from the vertical and is associated with a normal acceleration of the dummy's chest into the seatback of up to 10 g. The tangential acceleration of the dummy, which may induce riding up the seat, was generally less than the normal component so that the occupant was prevented from sliding up the deflected seatback. The bucket seat provided adequate containment and control of occupant displacements for each of the initial seatback angles of 9°, 22°, and 35°. Thus, the performance of the seat was not significantly effected by the initial seatback angle when the impact was less than 6.4 m/s (9.5 g) and the total deflection of the seatback was less than 60°.

The bucket seat did not retain the dummy on the seat for rear-end impact simulations of 9.6 m/s and 15.5 g, irrespective of the initial seatback angle of 9°, 22°, or 35°. In these tests there was a total seatback deflection of greater than 60°. The tangential acceleration of the chest was comparable to the normal component of acceleration inducing deformation of the seatback. In this case there was sufficient tangential acceleration to allow the dummy to slide up the seatback and induce extension of the neck greater than 70° as the head rides above the integrated head restraint.

In a limited series of tests, the ability of the seat to adequately retain an occupant surrogate in rear-end impacts was directly associated with the severity of deceleration. Although the seatback may dynamically deform 20°-30° during occupant loading, the tangential acceleration of the occupant is not sufficient to overcome the normal compression of the occupant into the seat at a seatback angle less than 60°. However, in those simulations where the seatback angle was greater than 60°, there was enough tangential acceleration of the occupant to permit riding up the seatback with the potential for secondary impacts of the occupant with the automotive interior.

We also observed that extension of the neck occurs as the anthropomorphic dummy's head rides above the integrated headrest of the seatback. This occurs late in the impact simulation (100-120 ms) near the end of the whole body displacement of the occupant into and deflecting the seatback. More importantly, as the anthropomorphic dummy leaves the seatback, the head may be extended as much as 70° when the surrogate moves toward rearward components of the car interior.