MATLAB在卡尔曼滤波器中应用的理论与实践Kalman
MATLAB在卡尔曼滤波器中应用的理论与实践KalmanKALMAN FILTERINGTheory and Practice Using MATLABThird editionMOHINDER S GREWALCalifornia State University at FullertonANGUS P. ANDREWSRockwell Science Center (retired)WILEYA JOHN WILEY & SONS, INC. PUBLICATIONCopyright 2008 by John Wiley sons, Inc. All rights reservedPublished by John Wiley sons, InC, Hoboken, New JerseyPublished simultaneously in CanadaNo part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or byany means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permittedunder Section 107 or 108 of the 1976 United States Copyright Act, without either the prior writtenpermission of the Publisher, or authorization through payment of the appropriate per-copy fee to theCopyright Clearance Center, Inc, 222 Rosewood Drive, Danvers, MA 01923,(978)750-8400, fax(978)750-4470,oronthewebatwww.copyright.com.RequeststothePublisherforpermissionshouldbe addressed to the Permissions Department, John Wiley Sons, Inc, lll River Street, Hoboken, NJ07030,(201)748-6011,fax(201)748-6008,oronlineathttp://www.wiley.com/go/permissionimit of liability Disclaimer of Warranty: While the publisher and author have used their best efforts inpreparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability orfitness for a particular purpose. No warranty may be created or extended by sales representatives orwritten sales materials. The advice and strategies contained herein may not be suitable for your situationYou should consult with a professional where appropriate. Neither the publisher nor author shall be liablefor any loss of profit or any other commercial damages, including but not limited to special, incidentalconsequential, or other damagesFor general information on our other products and services or for technical support, please contact ourCustomer Care Department within the United States at(800)762-2974, outside the United States at(317)572-3993 or fax(317)572-4002Wiley also publishes its books in a variety of electronic formats. Some content that appears in print maynot be available in electronic format. For more information about wiley products, visit our web site atwww.wiley.comLibrary of Congress Cataloging- in-Publication DataGrewal. Mohinder sKalman filtering: theory and practice using MATLAB/Mohinder S. GrewalAngus p. andrews. 3rd edIncludes bibliographical references and indexISBN978-0-470-17366-4( cloth)1. Kalman filtering. 2. MATLAB. I. Andrews, Angus P. II. TitleQA402.3.G69520086298312—dc22200803733Printed in the United States of america10987654321CONTENTSPrefaceAcknowledgmentsXIIIList of abbreviationsXV1 General Information1.1 On Kalman Filtering1.2 On Optimal Estimation Methods, 51. 3 On the notation Used In This book 231. 4 Summary, 25Problems. 262 Linear Dvnamic Systems2. 1 Chapter focus, 312.2 Dynamic System Models, 362. 3 Continuous Linear Systems and Their Solutions, 402.4 Discrete Linear Systems and Their Solutions, 532.5 Observability of Linear Dynamic System Models, 552.6 Summary, 61Problems. 643 Random Processes and Stochastic Systems3.1 Chapter Focus, 673.2 Probability and random Variables (rvs), 703.3 Statistical Properties of RVS, 78CONTEN3.4 Statistical Properties of Random Processes(RPs),803.5 Linear rp models. 883.6 Shaping Filters and State Augmentation, 953.7 Mean and Covariance propagation, 993.8 Relationships between Model Parameters, 1053.9 Orthogonality principle 1143.10 Summary, 118Problems. 1214 Linear Optimal Filters and Predictors1314.1 Chapter Focus, 1314.2 Kalman Filter. 1334.3 Kalman-Bucy filter, 1444.4 Optimal Linear Predictors, 1464.5 Correlated noise Sources 1474.6 Relationships between Kalman-Bucy and wiener Filters, 1484.7 Quadratic Loss Functions, 1494.8 Matrix Riccati Differential Equation. 1514.9 Matrix Riccati Equation In Discrete Time, 1654.10 Model equations for Transformed State Variables, 1704.11 Application of Kalman Filters, 1724.12 Summary, 177Problems. 1795 Optimal Smoothers5.1 Chapter Focus, 1835.2 Fixed-Interval Smoothing, 1895.3 Fixed-Lag Smoothing, 2005.4 Fixed-Point Smoothing, 2135.5 Summary, 220Problems. 226 Implementation Methods2256. 1 Chapter Focus, 2256.2 Computer Roundoff, 2276.3 Effects of roundoff errors on Kalman filters 2326.4 Factorization Methods for Square-Root Filtering, 2386. 5 Square-Root and UD Filters, 2616.6 Other Implementation Methods, 2756.7 Summary, 288Problems. 2897 Nonlinear Filtering2937.1 Chapter Focus, 2937.2 Quasilinear Filtering, 296CONTENTS7.3 Sampling Methods for Nonlinear Filtering, 3307.4 Summary, 345Problems. 3508 Practical Considerations3558.1 Chapter Focus. 3558.2 Detecting and Correcting Anomalous behavior, 3568.3 Prefiltering and Data Rejection Methods, 3798.4 Stability of Kalman Filters, 3828. 5 Suboptimal and reduced- Order Filters, 3838.6 Schmidt-Kalman Filtering, 3938.7 Memory, Throughput, and wordlength Requirements, 4038.8 Ways to Reduce Computational requirements 4098.9 Error Budgets and Sensitivity Analysis, 4148.10 Optimizing Measurement Selection Policies, 4198.11 Innovations analysis, 4248.12 Summary, 425Problems. 4269 Applications to Navigation4279.1 Chapter focus, 4279.2 Host vehicle dynamics, 4319.3 Inertial Navigation Systems(INS), 4359. 4 Global Navigation Satellite Systems(GNSS), 4659.5 Kalman Filters for GNSS. 4709.6 Loosely Coupled GNSS/INS Integration, 4889.7 Tightly Coupled GNSS /INS Integration, 4919. 8 Summary, 507Problems. 508Appendix A MATLAB Software511A 1 Notice. 511A 2 General System Requirements, 511A 3 CD Directory Structure, 512A 4 MATLAB Software for Chapter 2, 512A. 5 MATLAB Software for Chapter 3, 512A6 MATLAB Software for Chapter 4, 512A. 7 MATLAB Software for Chapter 5, 513A 8 MATLAB Software for Chapter 6, 513A 9 MATLAB Software for Chapter 7, 514A10 MATLAB Software for Chapter 8, 515A 11 MATLAB Software for Chapter 9, 515A 12 Other Sources of software 516CONTENAppendix b A Matrix Refresher519B. 1 Matrix Forms. 519B 2 Matrix Operations, 523B 3 Block matrix Formulas. 527B 4 Functions of Square Matrices, 531B 5 Norms. 538B6 Cholesky decomposition, 541B7 Orthogonal Decompositions of Matrices, 543B 8 Quadratic Forms, 545B 9 Derivatives of matrices. 546Bibliography549Index565PREFACEThis book is designed to provide familiarity with both the theoretical and practicalaspects of Kalman filtering by including real-world problems in practice as illustrativeexamples. The material includes the essential technical background for Kalman filter-ing and the more practical aspects of implementation: how to represent the problem ina mathematical model, analyze the performance of the estimator as a function ofsystem design parameters, implement the mechanization equations in numericallystable algorithms, assess its computational requirements, test the validity of resultsitor the filteThetant attributes ofthe subject that are often overlooked in theoretical treatments but are necessary forapplication of the theory to real-world problemsIn this third edition, we have included important developments in the implemen-tation and application of Kalman filtering over the past several years, including adaptations for nonlinear filtering, more robust smoothing methods, and develelopingapplications in navigationWe have also incorporated many helpful corrections and suggefrom ourreaders, reviewers, colleagues, and students over the past several years for theoverall improvement of the textbookAll software has been provided in MatLab so that users can take advantage ofits excellent graphing capabilities and a programming interface that is very close tothe mathematical equations used for defining Kalman filtering and its applicationsSee Appendix a for more information on MATLAB softwareThe inclusion of the software is practically a matter of necessity because Kalmanfiltering would not be very useful without computers to implement it. It provides aMATLAB is a registered trademark of The Mathworks, IncEFACEbetter learning experience for the student to discover how the Kalman filter works byobserving it in actionThe implementation of Kalman filtering on computers also illuminates some of thepractical considerations of finite-wordlength arithmetic and the need for alternativealgorithms to preserve the accuracy of the results. If the student wishes to applywhat she or he learns, then it is essential that she or he experience its workingsand failings--and learn to recognize the differenceThe book is organized as a text for an introductory course in stochastic processes atthe senior level and as a first-year graduate-level course in Kalman filtering theory andapplicationIt can also be used for self-instruction or for purposes of review by practi-cing engineers and scientists who are not intimately familiar with the subject. Theorganization of the material is illustrated by the following chapter-level dependencygraph, which shows how the subject of each chapter depends upon material in otherchapters. The arrows in the figure indicate the recommended order of study. Boxesabove another box and connected by arrows indicate that the material represented bythe upper boxes is background material for the subject in the lower boxAPPENDIX B: A MATRIX REFRESHERGENERAL INFORMATION2. LINEAR DYNAMIC SYSTEMSRANDOM PROCESSES AND STOCHASTIC SYSTEMS4. OPTIMAL LINEAR FILTERS AND PREDICTORS5. OPTIMAL SMOOTHERS6. IMPLEMENTATIONMETHODS7. NONLINEAR8. PRACTICAL9. APPLICATIONSFILTERINGCONSIDERATIONSTO NAVIGATIONAPPENDIX A: MATLAB SOFTWAREChapter l provides an informal introduction to the general subject matter by wayof its history of development and application. Chapters 2 and 3 and Appendix b coverthe essential background material on linear systems, probability, stochastic processesand modeling. These chapters could be covered in a senior-level course in electricalcomputer, and systems engineeringChapter 4 covers linear optimal filters and predictors, with detailed examples ofapplications. Chapter 5 is a new tutorial-level treatment of optimal smoothing
- 2020-12-01下载
- 积分：1
转子系统固有频率的传递矩阵计算方法及其MATLAB实现
文章介绍了计算多自由度转子系统固有频率的传递矩阵法，以及用于实现该算法的Prohl法和Riccati 法的推导过程。利用Matlab 强大的绘图计算功能和改进的Riccati 传递矩阵法所具有的良好的数值稳定性，避免了传统的Prohl 传递矩阵法在计算过程中的丢根现象，提高了整个转子系统分析运算的精度。并用Matlab 对各算法的数值稳定性进行了分析。190其中112,21,2对应于(3)式的矩阵各项。将式(6)展开,得:}+1=11M}+12引入如下的 Riccal变换式中[]就是ca传递矩阵,它是一个2×2阶的待定矩阵,把式(8)代人式(7)式中得这就是 Riccait递推公式。由起始截面的边界条件(门1=0,(e小)≠0固有初始条件[S]=[0]。代人式(9)就可依次递推[S,[,.S对末端截面N+1有:由边界条件{门}x1-{0},{e≠0故得(10)式有解的条件是:+和PωM/法一样,在感兴趣的范围内按一定的步长选定试算频率计算出剩余量S-值,就可以画出剩余量随a变化的曲线,曲线与横坐标交点所对应的转速就是转子的各界临界转速。在PmM的传递矩阵法中,是用r阶的矩阵递推来求剩余量△(o2)。在Bceb的传递矩阵法中是用r/2阶的矩阵国递推来求剩余量S×+1,由于与的递推式中含有逆矩阵,使得剩余量曲线经常会出现异号无穷型奇点。因而在常见的转子动力系统中,剩余量曲线的根和奇点的位置十分接近。在实际转子系统中,临界转速值与奇点值间的间隔可能少于10/m,因此这种方法的丢根现象不可避免。参考PnoM方法中剩余量△(a2)无奇点的事实,可以对 riccati方法中的剩余量加以改造。由式(10)得+1n{%+12]1{}依次类推{}[]+∏[2+21{12在满足相同边界条件时应有△1=[]L21320064事实上(12)式就是(5)式,只是在数值计算中,它们是按不同的方法递推而得到的。因此在数值上它们的精度也不同。当PmM法出现数值不稳定时,(13)式所示的剩余量仍然保持相当的精度。由于剩余量(o2)随0变化的曲线不存在奇点,因此以作剩余量的曲线也不存在奇点。由于(12)式中un+ux]在进行S的递推过程中都已求得,所以在计算时也不会增加太多的工作量,但却可以克服丟根的缺点。事实上(13)式是把(11)式的异号无穷型奇点变为同号无穷型奇点,这样只有当跨过一个真正的根时才变号。枚除了两个临界转速值非常接近的情况,即当两临界转速的差小于所选步长时,一般不会发生漏根。三利用MmMh编制PmM/法、Bicn法及改进的kKRiccati法的程序对各算法结果进行分析。运用算例:如图转子系统简化模型,其数据如下1转子系统简化模型2.94t=588t(=236)1.3m(=1,2,,6)29592×10(kN·m)(i=1,2,)6)支承简化为如图模型相应参数为1.9600×106kN.m-1;2.7048×10kN·m=3.5771(=1,2)编制Maab程序运行待如下表所示的各阶频率。从表1可以看出在 Protel法的计算结果中,小于1058239rad/s固有频率共计算出了7个, Ricca算法计算出了13个固有频率,而改进了的ieai算法在消除奇点干扰后可以计算出17个固有频率。从而明显的看出改进的Racm法可以很好的避免计算过程中的丢根,在数值上具有很好的稳定性。计算细果慧裝protel算法(rads)Riccati算法(rads)改进的 Riccati算法(rad/s190.812100.815208249197.895197.895445924208.245208.24522.9655646.410445.9256832.610458.175458.1751058.239539925539925580.l659646.415574.265759.225580.165832.615646.415987.0057150451058.23583261516987.0051058.235利用a的绘图功能我们可以直观的从图中分析岀各算法的漏根现黎如图2、图3、图4所示:1912P法计算恩有单率输出固像1eg法计算回有率出四像t”改进计算有率始步入从图2、图3、图4可以看出在530到580的频率区间上,前两条曲线与0轴只有一个交点即所求固有频率为539925a/s的点,第三条曲线在相同的区间上与0轴的交点为三个,显然改进的 mccall方法找回了漏掉的根550.225ad/s和574265rad/s。利用 Matlab程序绘图我们还可以绘出改进的 Riccati方法把异号无穷型奇点转化成了同号的无穷型奇点的情况,如图5、图6所示。从图中区间987ras到1090rad/s的曲线可以明显的看出图5曲线以0轴为对称轴倒置后即得到图6在此区间的曲线线形,从而改进了 Riccati算法,在曲线中,只有在跨过个真正的根时剩余量才变号。所以除了两个临界转速之差小于所选步长的情况除外,一般改进后的riccati算法不会发生丢根c算利0改进的热计算有明p1m0p三41000100在计算多自由度转子系统固有频率的传递矩阵法中,我们可以利用 Matlab编程实现Ph/法、 riccati法以及改进的Riea法对于系统固有频率的计算,利用Maab的绘图功能对各算法的结果进行直观的分析,从而明显的看出各算法的漏根情况。本文对于计算复杂的多自由度系统固有频率具有参考意义,也可用于复杂系统低阶固有频率的粗算。同时 Matlab的矩阵运算功能在传递矩阵法中也得到了充分的利用(Electromechanical Engineering Dept, Sichuan University of Science Engineering, Zigong 643000, China)This article introduced the transfer matrix method about the natural frequency calculation of themuulti- degrees freedom rotor system, as well as inferential reasoning process about Prohl law and thericcatilayUSing formidable cartography and computation function of the Matlab as well as the good value stability aboutimproved riccati law it avoided the losing of the natural frequency and enhanced the precision ofentire rotorsystem further analyze. The value stability of various algorithms areanalyzed with Matlab in the paper toorotor system; natural frequency; transfer matrix method; Matlab
- 2020-12-04下载
- 积分：1
|
联系客服（QQ：3324461878）
