多级预点火安全气囊的控制算法研究

发布时间：2018-09-17 06:46

【摘要】：汽车安全气囊是汽车被动安全系统中一个非常重要的装置，汽车发生碰撞时对保护乘员起着非常重要的作用。随着安全气囊技术的不断发展，出现了双气室、多气室等安全气囊技术。双气室安全气囊可以根据需要来调节打开的方式，可实现对气囊内部升压速率和峰值压力的调节，从而可以实现较单气室安全气囊对乘员更为完善的保护。点火控制算法是安全气囊系统的控制核心，控制算法的好坏，直接关系到安全气囊对乘员的保护效果以及气囊的安全可靠性。本文在传统安全气囊控制算法的基础上，研究驾驶员一侧正面双气室安全气囊的点火控制算法，建立基于模糊神经网络的安全气囊多级预测点火控制算法。其网络结构共分为上下两层，一层用于预测乘员头部位移，另一层用于对点火模式（一个气室点火、两个气室间隔10ms依次点火、两个气室同时点火和不点火的多级模式）的选择。双气室安全气囊只有满足两个条件，系统才会按照指定的点火模式点火：一个是根据安全气囊“127mm-30ms”点火规则，网络上层的纯神经网络预测驾驶人头部30ms后的位移是否达到127mm，达到则满足其中的一个条件；另一个是模糊神经网络层对点火做出点火决策，即作出是否需要点火，若点火按照哪一种方式点火，算法只有满足这两个条件才能按照某一种方式点火（单气室点火、两个气室间隔一定时间依次点火、两个气室同时点火）。网络经过MATLAB模糊与神经网络工具箱建模并训练后，，用样本数据进行检验，检验结果表明，本算法在现有数据的基础上，能够较准确地完成双气室安全气囊对点火时刻和点火条件的判定。结果表明，本控制算法对研究汽车双气室安全气囊的智能点火有一定的意义。
[Abstract]:Automobile airbag is a very important device in vehicle passive safety system, which plays a very important role in protecting occupants. With the development of airbag technology, double air chamber, multi-air chamber and other airbag technology appear. The double chamber airbag can adjust the opening mode according to the need, can realize the adjustment of the pressure rise rate and the peak pressure inside the airbag, thus can realize the more perfect protection to the occupant than the single air chamber airbag. The ignition control algorithm is the control core of airbag system. The quality of the control algorithm is directly related to the airbag's protective effect to the occupant and the safety reliability of the airbag. Based on the traditional airbag control algorithm, this paper studies the ignition control algorithm of the airbag on the front side of the driver, and establishes a multi-stage predictive ignition control algorithm for the airbag based on fuzzy neural network. The network structure is divided into two layers, one layer is used to predict the head displacement of the occupants, the other layer is used to ignite the fire mode (one chamber ignites, the 10ms between the two chambers is ignited sequentially). Selection of multi-stage modes of ignition and non-ignition for both chambers. Only if two conditions are satisfied, the system will ignite according to the specified ignition mode: one is based on the airbag "127mm-30ms" ignition rule, The pure neural network in the upper layer of the network predicts whether the displacement of the driver's head after 30ms is 127mm, and meets one of the conditions; the other is that the fuzzy neural network layer makes the decision to ignite the fire, that is, to make the decision on whether or not to light the fire. If the fire is ignited in which way, the algorithm can ignite the fire in a certain way only if the two conditions are satisfied (single chamber ignition, the interval between two chambers in turn, and the two chambers simultaneously). After modeling and training by MATLAB fuzzy and neural network toolbox, the network is tested with sample data. The results show that the algorithm is based on the existing data. It can accurately judge the ignition time and ignition condition of double air chamber airbag. The results show that this control algorithm has some significance for the intelligent ignition of double air chamber airbag.
【学位授予单位】：重庆理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.61;TP273

【参考文献】