Search Results

Now that airbags are the accepted solution for protecting occupants in frontal impacts, and now that safety sells cars, it is natural to look closely at the second largest killer of automobile occupants, side impacts. This paper develops the theory of sensing side impacts based on the assumption that airbags will soon be used for side impact protection. The trade-offs between the various sensor technologies are discussed including electronic and mechanical sensors. For mechanical sensors, fluid damped, undamped and crush sensing switches are compared. Finally, the requirements for a successful predictive sensor will be presented.

This is the first paper in a new series to present a coherent theory of sensing frontal crashes, define the characteristics of future airbag sensor systems and to present examples of how this theory can be implemented. After summarizing the relevant conclusions from the authors' previous papers, this paper concludes that future systems should contain: crush zone sensors which sense relevant impacts to all portions of the vehicle front; an occupant position sensor as an input to the sensing system; and a mechanical safing/arming sensor having a long dwell. It is further concluded that cars should be designed so that only impacts involving the front of the vehicle need be sensed for the deployment of frontal protection airbags. This series of papers has the main goal of determining an overall theory of frontal crash sensing and the resulting desirable properties of sensor systems. A second goal is to give examples of how this theory can be realized in real sensor systems.

Crash sensors for use in deploying air bags operate in an environment of severe vibrations, not only along the longitudinal axis of the sensor, but also in the transverse vertical and lateral axes. These vibrations can have a detrimental effect on some crash sensor designs. Various methods using Fourier analysis have failed to provide a characterization of cross-axis vibrations. This is due, in part, to phase shifts in the cross-axis oscillations. A technique using shock spectrum analysis has been developed which can be used to characterize these vibrations. This work has led to the development of a specification for laboratory testing of the sensitivity of a sensor to cross-axis vibrations. This paper presents the underlying theoretical basis for the shock spectrum technique, the results of applying this technique to a library of crash data, and a recommended specification for laboratory sensor testing for sensors mounted in the crush zone and non-crush zone.

In two previous SAE papers (1,2) by the authors, supporting analysis was presented showing the difficulty in achieving a timely response to real-crash events using a single point sensor mounted in the non-crush zone of the vehicle (tunnel, cowl, etc.). The analysis demonstrated the propensity to deploy the air bag(s) late during certain of these events. If a vehicle occupant was not wearing a safety belt, the deceleration forces of the crash could place the occupant out of position and resting against the air bag when it was deployed. In another SAE paper (3) by H. J. Mertz et al, the authors demonstrated that animals, used as surrogates for humans, could be injured if positioned against an air bag at the time of deployment. Arguments are presented here to show that there is insufficient information in the crash pulse as sensed in the non-crush zone to deploy an air bag in time for the unbelted occupant.

In earlier SAE papers the authors demonstrated that the ideal crash sensor for use in the crush zone has a constant velocity change response, and various sensors are now in use which perform as acceleration integrators. Further study of sensor performance, however, has shown that crush zone sensors function by being struck by crushed material which is forced rearward during the crash. This observation has led to the design of an inexpensive sensor which measures crush instead of velocity. The crush of the vehicle is used as an accurate indicator of the severity of the crash. This paper presents the theoretical basis for a sensor which initiates air bag deployment when the crush of the vehicle exceeds a pre-selected amount. In a companion SAE paper, the authors have demonstrated that single point sensing in the passenger compartment may result in late air bag deployment on soft crashes, and, therefore, sensing in the crush zone is required.