摘要
建立正面碰撞乘员约束系统空间模型,通过实车碰撞试验验证了模型的正确性.结合均匀试验设计方法,分别建立3种人工神经网络拟合乘员伤害,网络结构经过优化和训练后使用遗传算法寻优,优化后的参数使得约束系统的保护性能得到显著改善.优化结果表明:RBF网络和GRNN网络对乘员约束系统的拟合效果较好,预测精度较高.
The occupant restraint system model of full front barrier crash was established and verified by real automotive crash test.Combining with the uniform design method,three types of artificial neural networks were established to simulate occupant injuries.After structure optimization and training of neural networks,the optimal solution was found by genetic algorithm,the performance of ORS with optimized parameters was dramatically improved.The result shows that RBF network and GRNN network can fit ORS very well and obtain an accurate prediction.
出处
《上海工程技术大学学报》
CAS
2010年第3期257-261,共5页
Journal of Shanghai University of Engineering Science
基金
上海高校知识创新工程(085)建设资助项目(JZ0901)
上海高校特聘教授(东方学者)岗位计划资助项目
关键词
乘员约束系统
神经网络
乘员伤害
优化
occupant restraint system（ORS）
neural network
occupant injury
optimization
