基于贝叶斯推理的乘员约束系统参数识别  

Parameters Identification for Occupant Restraint System Based on Bayesian Inference

作　　者：卿宏军[1,2] 谢宝娣 张连怡 QING Hongjun;XIE Baoti;ZHANG Lianyi(College of Mechanical and Vehicle Engineering,Hunan University,Changsha 410082,China;Changzhou Institute of Machinery and Equipment of Hunan University,Changzhou 213000,China;Key Laboratory of Aerospace System Simulation of Beijing Simulation Center,Beijing 100854,China)

机构地区：[1]湖南大学机械与运载工程学院,湖南长沙410082 [2]常州湖南大学机械装备研究院,江苏常州213000 [3]北京仿真中心航天系统仿真重点实验室,北京100854

出　　处：《湖南大学学报（自然科学版）》2018年第8期42-47,共6页Journal of Hunan University:Natural Sciences

基　　金：国家自然科学基金资助项目(11572115);江苏省前瞻性联合研究项目(BY2014055);国家重点研发计划项目(2017YFC0820100);常州市应用基础研究计划项目(CJ20179009)~~

摘　　要：为了克服测量响应的不确定性给乘员约束系统参数识别带来的困难,利用马尔科夫链蒙特卡洛(Markov Chain Monte Carlo,MCMC)采样和近似模型构造技术,提出一种基于贝叶斯推理的乘员约束系统不确定性参数识别方法.该方法结合约束系统参数的先验分布和测量响应,通过马尔科夫链在未知参数联合概率密度空间进行抽样,从而获得了织带刚度缩放系数和质量流率缩放系数的后验边缘概率密度函数.识别结果表明,相比于传统确定性识别方法,基于贝叶斯推理的不确定性参数识别方法不仅能有效给出乘员约束系统参数的概率分布,而且能够保证参数寻优的全局收敛性.In order to overcome the difficulty for the parameters identification of occupant restraint system caused by the measured uncertainty,this paper proposed an uncertain identification method for the parameters of occupant restraint system based on Bayesian inference,which combined Markov Chain Monte Carlo（MCMC）sample and surrogate model.This method firstly obtains the prior distributions of identified parameters and measured responses,and then the MCMC sampling is applied to the joint probability density of unknown parameter.Then,the marginal posterior probability distributions of scale factor of webbing and rate of flow can be calculated.Compared with the traditional method of determined identification,the identified results show that the Bayes inference method for uncertain parameter identification not only obtains the probability distributions effectively,but also ensures the global convergence of identified parameter.

关 键 词：贝叶斯方法 参数识别 约束系统 MCMC 代理模型 

分 类 号：TH122[机械工程—机械设计及理论]