基于在线LSSVM的陀螺漂移预测算法研究
详细信息 查看官网全文
- 英文论文题名：Research of Gyro Drift Prediction Based on Online LSSVM
- 论文作者：洪贝 ; 姜学鹏 ; 陈青华 ; 丁海强
- 英文论文作者：Hongbei ; Jiang Xuepeng ; Chen Qinghua ; Ding Haiqiang ; Naval Aeronautical and Astronautical University
- 年：2016
- 作者机构：海军航空工程学院;
- 论文关键词：非线性系统 ; 最小二乘支持向量机 ; 自适应辨识 ; 陀螺漂移预测 ; Hermite矩阵求逆
- 英文论文关键词：Nonlinear system ; least squares support vector machine ; adaptive identification ; gyro drift prediction ; hermite matrix inversion.
- 会议召开时间：2016-10-31
- 会议录名称：探索 创新 交流（第7集）——第七届中国航空学会青年科技论坛文集（上册）
- 英文会议录名称：Discovery,Innovation and Communication——Proceedings of 7th CSAA Science and Technology Youth Forum
- 语种：中文
- 分类号：V243
- 学会代码：ZGHU
- 会议名称：探索 创新 交流（第7集）——第七届中国航空学会青年科技论坛
- 会议地点：中国广东中山
- 主办单位：中国航空学会
- 学会名称：中国航空学会
- 页数：5
- 文件大小：298k
- 原文格式：O
- 会议级别：全国
摘要
为了预测陀螺漂移趋势,以某型陀螺的漂移角速度时间序列为对象,建立基于最小二乘支持向量机的非线性时间预测模型,提出了一种基于Hermite矩阵求逆引理的在线更新算法。首先介绍最小二乘支持向量机及其在线算法,针对在线更新时存在复杂的矩阵求逆运算,结合核扩展矩阵为实对称矩阵的特点,利用Hermite矩阵求逆引理递推求取核扩展矩阵。实测的陀螺漂移数据预测应用研究表明,模型在线更新的过程中,本文算法能充分利用历史的训练结果和核扩展矩阵的特点来减小模型计算复杂度,运算速度快、预测精度高。
For forecasting the gyro drift tendency,a prediction model based on LSSVM is established taking the time series of gyro's drift as study object,and a online algorithm for nonlinear system based on hermite matrix inversion is presented in this paper.Firstly,we introduced the mathematical model of regression LSSVM,and online learning algorithm.Based on the characteristic that the reproducing kernel matrix is Hermite and positive definite,we presented a new online learning algorithm by matrix block-inversion.The algorithm is applied to prediction research on real certain gyro drift data,and experiment results show that the algorithm fully utilizes the historical training results,reduces the storage space and calculation,resulting to a fast operation speed and a high prediction precision.
For forecasting the gyro drift tendency,a prediction model based on LSSVM is established taking the time series of gyro's drift as study object,and a online algorithm for nonlinear system based on hermite matrix inversion is presented in this paper.Firstly,we introduced the mathematical model of regression LSSVM,and online learning algorithm.Based on the characteristic that the reproducing kernel matrix is Hermite and positive definite,we presented a new online learning algorithm by matrix block-inversion.The algorithm is applied to prediction research on real certain gyro drift data,and experiment results show that the algorithm fully utilizes the historical training results,reduces the storage space and calculation,resulting to a fast operation speed and a high prediction precision.
引文
[1]蔡艳宁,胡昌华,等.基于支持向量回归机的陀螺漂移预测模型,中国惯性技术学报[J],2007,15(5).
[2]于旭东,张鹏飞,谢永元,等.基于CPSO-LSSVM的单轴旋转惯导系统轴向陀螺漂移辨识.系统工程与电子技术,2013,35(5):1049-1053.
[3]张伟,胡昌华,等.遗忘因子最小二乘支持向量机及在陀螺仪漂移预测中的应用研究,宇航学报[J],2007,28(2).
[4]Suykens J A K,Vandewalle J.Least Square Support Vector Machines Classifiers.Neural Processing Letters,1999,9(3):293-300.
[5]Suykens J A K,Vandewalle J,Moor B D.Optimal Control by Least Squares Support Vector Machines.Neural Networks,2001,14:23-35.
[6]J A K Suykens,J D Brabanter,L Lukas,J Vandewalle.Weighted Least Squares Support Vector Machines;Robustness and Sparse Approximation.Neurocompting,2002,48(1-4):85-105.
[7]Fabian Ojeda,Johan A K Suykens,Bart De Moor.Low Rank Updated LS-SVM Classifiers for Fast Variable Selection.Neural Networks,2008,21:437-449.
[8]张浩然,汪小东.回归最小二乘支持向量机的增量和在线式学习算法.计算机学报,2006,29(3):400-406.
[9]范玉刚,李平,宋执环.动态加权最小二乘支持向量机.控制与决策,2006,21(10):1129-1133.
[10]蔡艳宁。胡昌华,汪洪桥,等.基于自适应动态无偏LSSVM的故障在线监测.系统仿真学报,2009,21(13):4129-4134.
[11]杨志峰,雷虎民,董飞垚,等.基于LS_SVM的导弹在线误差补偿逆控制.系统工程与电子技术,2010,32(6):1314-1317.
[12]赵洋,陈国顺,马飒飒,等.基于OS.LSSVM的电子装备在线故障预测模型研究.计算机测量与控制.2012.20(11):2903-2908.
[13]田中大,高宪文,李琨.基于KPCA与LSSVM的网络控制系统时延预测方法.系统工程与电子技术.2013,35(6):1281-1285.
[14]王健康,张海波,黄向华,等.基于在线滚动LS.SVR的涡轴发动机混合预测控制.航空学报,2012,33(10):1755-1764.
[15]曹葵康,沈海斌.基于多样本的在线支持向量回归算法.浙江大学学报(理学版),2011,38(4):405-408.
[2]于旭东,张鹏飞,谢永元,等.基于CPSO-LSSVM的单轴旋转惯导系统轴向陀螺漂移辨识.系统工程与电子技术,2013,35(5):1049-1053.
[3]张伟,胡昌华,等.遗忘因子最小二乘支持向量机及在陀螺仪漂移预测中的应用研究,宇航学报[J],2007,28(2).
[4]Suykens J A K,Vandewalle J.Least Square Support Vector Machines Classifiers.Neural Processing Letters,1999,9(3):293-300.
[5]Suykens J A K,Vandewalle J,Moor B D.Optimal Control by Least Squares Support Vector Machines.Neural Networks,2001,14:23-35.
[6]J A K Suykens,J D Brabanter,L Lukas,J Vandewalle.Weighted Least Squares Support Vector Machines;Robustness and Sparse Approximation.Neurocompting,2002,48(1-4):85-105.
[7]Fabian Ojeda,Johan A K Suykens,Bart De Moor.Low Rank Updated LS-SVM Classifiers for Fast Variable Selection.Neural Networks,2008,21:437-449.
[8]张浩然,汪小东.回归最小二乘支持向量机的增量和在线式学习算法.计算机学报,2006,29(3):400-406.
[9]范玉刚,李平,宋执环.动态加权最小二乘支持向量机.控制与决策,2006,21(10):1129-1133.
[10]蔡艳宁。胡昌华,汪洪桥,等.基于自适应动态无偏LSSVM的故障在线监测.系统仿真学报,2009,21(13):4129-4134.
[11]杨志峰,雷虎民,董飞垚,等.基于LS_SVM的导弹在线误差补偿逆控制.系统工程与电子技术,2010,32(6):1314-1317.
[12]赵洋,陈国顺,马飒飒,等.基于OS.LSSVM的电子装备在线故障预测模型研究.计算机测量与控制.2012.20(11):2903-2908.
[13]田中大,高宪文,李琨.基于KPCA与LSSVM的网络控制系统时延预测方法.系统工程与电子技术.2013,35(6):1281-1285.
[14]王健康,张海波,黄向华,等.基于在线滚动LS.SVR的涡轴发动机混合预测控制.航空学报,2012,33(10):1755-1764.
[15]曹葵康,沈海斌.基于多样本的在线支持向量回归算法.浙江大学学报(理学版),2011,38(4):405-408.