基于WLS-SVM的加速度计动态模型参数辩识  

Accelerometer dynamic model parametric identification using WLS-SVM

作　　者：王建林[1] 郭永奇 魏青轩[1] 孙桥[2] 胡红波[2] WANG Jianlin;GUO Yongqi;WEI Qingxuan;SUN Qiao;HU Hongbo(College of Information Science and Technology,Beijing University of Chemical Technology,Beijing 100029,China;National Institute of Metrology,Beijing 100029,China)

机构地区：[1]北京化工大学信息科学与技术学院,北京100029 [2]中国计量科学研究院,北京100029

出　　处：《振动与冲击》2018年第19期239-244,253,共7页Journal of Vibration and Shock

基　　金：国家重大科学仪器设备开发专项项目(2012YQ090208)

摘　　要：提高加速度计动态模型参数辨识精度,对研究和改善加速度计动态特性有重要作用。针对加速度计的非线性影响其动态模型参数辨识精度的问题,提出了一种基于加权最小二乘(WLS)和支持向量机(SVM)的加速度计动态模型参数辩识方法,该方法针对包含线性部分和非线性项的加速度计二阶非线性动态模型,利用WLS辩识加速度计动态模型的线性部分参数,并采用SVM估计加速度计动态模型的非线性特性,通过迭代和最小化所构建的误差准则函数,实现加速度计动态模型参数最优辨识。仿真实验和加速度计绝对法冲击激励校准实验表明,该方法能够减小加速度计非线性对动态模型参数辩识精度的影响,所得加速度计动态模型参数辨识结果具有较高的精度。It is important to improve parametric recognition accuracy of accelerometer dynamic model for studying and improving accelerometer dynamic characteristics.Aiming at the problem of nonlinearity of accelerometer affecting its dynamic model parametric identification accuracy,an accelerometer dynamic model parametric identification method using the weighted least squares(WLS)and the support vector machine(SVM)was proposed.Aiming at the second order nonlinear dynamic model of accelerometer containing a linear part and nonlinear terms,parameters of the linear part of the model were identified using WLS,and the nonlinear characteristics of the model were estimated using SVM.Then the constructed error criterion function was iterated and minimized to realize the optimal identification for accelerometer dynamic model parameters.The simulation tests and the accelerometer calibration tests under shock excitation based on the absolute method showed that the proposed method can be used to reduce influences of nonlinearity of accelerometer on its dynamic model parametric identification accuracy;the identified results for accelerometer dynamic model parameters have a higher accuracy.

关 键 词：加速度计 非线性动态模型 支持向量机 加权最小二乘 参数辩识 

分 类 号：TH824[机械工程—仪器科学与技术]