The two known systems for restraining vehicle occupants in a crash situation are the seat belt and the air bag. A series of head-on crash tests against a barrier indicated that these restraining devices present widely different G waveforms for the chest of a dummy seated behind the steering wheel as the crash pulse becomes more severe. The dummy chest, when restrained by a seat belt, showed a gentle G waveform, while with an air bag, it presented another waveform with unusual spikes.
An analysis of the test findings reveals that spikes in chest G are always accompanied by similar spikes in G for the pelvis. It was found that this is attributable to a substantial shock when the dummy's upper leg contacts the molded pelvic structure. This suggests that the existing dummy might have some structural deficiency. This contact occurs when the restraining force for the upper torso of the dummy is much less than that for the pelvis.
An air bag was used to restrain the driver's upper torso and a knee restraint to hold the pelvis. The restraining force for the upper torso was limited by the characteristics of the bag. Therefore, use of a softer knee bolster appeared to be the only available alternative for keeping the upper and lower restraining forces well-balanced.
Another series of crash tests was conducted on an improved air bag system using softer knee bolsters. The findings indicate that the improved system can allow gentle G waveforms even with severe crash pulses despite the shortcoming of the existing dummy.