The Research On Prediction Algorithm For Airbag Actuation Based On Artificial Neural Network

Airbag System, as one of the countermeasures of vehicle passive safety, has got fast development and popularization. In order to improve the reliability of airbag deployment, it is very necessary to study the ignition control theory. This paper analyzes the published ignition algorithms and puts forward a new one, which is based on artificial neural network, the reliability of which is also validated by experiment and simulation.This paper researches the design of the electronic control system of airbag and points out that the design of ignition algorithm is the key technology of an electronic control system. Through analyzing the ignition control theory, this paper concludes two key factors as the criterions of appraising the airbag's ignition: ignition term and optimal ignition moment. This paper also introduces published ignition algorithms and analyzes these ignition algorithms through a collision curve of the car. Most of the algorithms only use one index of the criterions to control the ignition of airbag. Under some condition, it maybe causes inaccurate inflation. This paper puts forward that it can be avoided if using two indexes in ignition control. One index is used to judge whether the airbag should inflate, another index is to make the optimal ignition moment. On the basis of that, this paper presents a new ignition algorithm based on artificial neural network. The algorithm uses two indexes to control the ignition term and optimal ignition moment of airbag respectively, which can ensure the airbag inflating accurately. The two indexes are collision intensity of automobile and the displacement of occupant's head. This paper designs a new experimentation which associates trolley experimentation with computer simulation. Through the experimentation, large numbers of collision curves can be obtained. The artificial neural network model of the algorithm is trained by the collision curves. Finally, the new ignition algorithm is validated. The result proves that the new ignition algorithm presented by this paper is accurate and can increase the reliability of the airbag significantly.