-
一个基于VB.net的实时曲线控件
一个基于VB.net的实时曲线控件，里面有源代码，可以直接用。
- 2020-12-02下载
- 积分：1
-
光伏电池仿真模型
光伏电池仿真模快，效果不错，值得借鉴，matlab平台，效果很不错，值得下载。
- 2020-12-01下载
- 积分：1
-
模板匹配技术
更深入理解模板匹配技术，能够独立根据已知算法（相关匹配（Correlation Matching）、基于Hausdorff距离匹配方法 及考虑对场景图象距离变换（Distance Transform）的Hausdorff距离匹配方法）在MATLAB下编程实现相关的模板匹配技术，并通过结果，对比不同算法的优缺点。1）利用①相关匹配（Correlation Matching）、②基于Hausdorff距离匹配方法 及③考虑对场景图象距离变换（Distance Transform）的Hausdorff距离匹配方法,实现模板目标在场景图象中的定位。2）对于每个模板分别给出最后的目标定位位置坐标（左
- 2020-11-27下载
- 积分：1
-
ViBe背景建模Matlab代码
ViBe背景建模Matlab代码实现，只完成了比较基础的部分，但与官网的实现方式差不多，速度比较快。
- 2020-12-06下载
- 积分：1
-
STM32F429 原理图 &&PCB
【实例简介】STM32F429IGT6完整PCB工程文件，包括PCB,SCH,LIB，四层板，可直接打样，亲测稳定可用，兼容原子，野火底板可用AD软件打开
- 2021-11-10 00:34:18下载
- 积分：1
-
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
-
VINS论文推倒及代码解析
VINS 的功能模块可包括五个部分:数据预处理、初始化、后端非线性优化、闭环检测及闭环优化。代码中主要开启了四个线程,分别是:前端图像跟踪、后端非线性优化(其中初始化和 IMU 预积分在这个线程中)、闭环检测、闭环优化。、总体框架Measurement PreprocessingInitializationCamera(30hz)Feature Detectionnd rackerVisual-lnertialInitializedis- onlySfMAlignmentIMU (100hMU Pre-integrationLocal Visual-Inertial: OldestSliting WindowNewestNonli+、 Keyframe?OptimizationBundle Adjustment II Loop detectionwith RelocalizationStates from Loop ClosureFealure retrievel oop Deleted二二1---11------22===Global Pose Graph4-DoF Pose Graph OptimizationKeyframe DatabaseOptimization图1VINS框架ⅵINS的玏能模块可包括五个部分:数据预处理、初始化、后端非线性优化、闭环检测及闭环优化。代码中主要开启了四个线稈,分别是:前端图像跟踪、后端非线性伉化(其中初始化和IMU预积分在这个线程中)、闭环检测、闭环优化各个功能模块的作用上要有:1.I图像和MU预处理●图像:提取图像 Harris角点,利用金字塔光流跟踪相邻帧,通过 RANSAC去除异常点,最后将跟踪到的特征点push到图像队列中,并通知后端进行处理●IU:将IMU数据进行积分,得到当前时刻的位置、速度和旋转(PVQ),同时计算在后端优化中将用到的相邻帧的预积分增量,及预积分误差的 Jacobian矩阵和协方差项。1.2初始化首先,利用SFM进行纯视觉佔计滑窗內所有帧的位姿及3D点逆深度,最后与IMU预积分进行对齐求解初始化参数1.3后端滑窗优化将视觉约束、IMU约束和闭环约束放在·个大的目标函数中进行非线性优化,求解滑窗内所有帧的PVQ、bias等。L M States in the sliding windowIMU:k States from loop clos1Camera:冷 MU measurements>visual measurements★ Catur图2滑窗优化示意图14闭环检测和优化利用D)BoW进行闭环检测,当检测成功后进行重定位,最后对整个相机轨迹进行闭环优化。U预积分VisionIMUVision图3MU预积分示意图21当前时刻pVQ的连续形式将第k唢和第kl帧之间的所有IMU进行积分,可得第kHI帧的位置、速度和旋转(PVQ),作为视觉估计的初始值,这里的旋转采用的四元数。v△t+k+1∈[k,k+1]rW(at-ba ) -owletbk JtE[k, k+1]n(,-bdt∈[k,k+1]其中,a2和O为ⅠMU测量的加速度和角速度,是在Body自身坐标系, world坐标系是IMU所在的惯导系,上式的旋转公式推导可参考附录10.1。22当前时刻PVQ的中值法离散形式公式(1)给出的是连续吋刻的相机当前PVR的达代公式,为了跟代码致,下面给出基于中值法的公式,这与 Estimator:; processIMg(O函数中的Ps]、Rs]和Vs是一致的,IMU积分出来的第j时刻的物理量可以作为第j帧图像的初始值。tr t+a26t(2)ka,St其中q(a1-ba)-g"+q:+1(a+1-ba)(a;+o;+1)2.3两帧之间PVQ增量的连续形式通过观察公式(1)可知,IvU的预积分需要依赖与第k帧的ν和R,当我们在后端进行非线性优化时,需要迭代更新第κ唢的ν和R,这将导致我们需要根据每次迭代后值重新进行积分,这将非常耗吋。因此,我们考虑将优化变量从第k帧到第κ+1帧的IU预积分项中分离开来,通过对公式(1)左右两侧各乘Rb,可化简为:R(+p2k-=2△)+ak+1b其中DtElk, k+1t∈[k,k+1R k(at-bar)ldt)Wendtt∈[kk+1这样我们就得到了连续时刻的MU预积分公式,可以发现,上式得到的MU预积分的值只与不同时刻的a2和o相关。这里我们需要重新讨论下公式(5)预积分公式,以ab,为例,我们发现它是与MU的bias相关的,而bias也是我们需要优化的变量,这将导致的问题是,当每次迭代时,我们得到一个新的bias,又得根据公式(巧5)重新对第k帧和第k+1帧之间的IMU预积分,非常耗时。这里假设预积分的变化量与bias是线性关系,可以写成:ab,+/6n6ba+/16b+8 8ba +p(6)k+1sb24两帧之间PVQ增量的欧拉法离散形式面给出离散时刻的IMU预积分公式,首先按照论文中采用的欧拉法,给出第i个MU时刻与第i1个IMU时刻的变量关系为b+1k+的6t+元R(P)(1+R(P")(a2-bbn)δt25两帧之间PⅤQ增量的中值法离散形式卜面给出代码中采用的基」中值法的IMU预积分公式,这与 Estimator: processIMUO函数中的 Integration Base: push backo上是一致的。注意这里跟公式(2)是不一样的,这里积分出来的是前后两顿之间的IU增量信息,而公式(2)给出的当前帧时刻的物理量信息+1+B k St +=a, &tbb+1Bi + au1其中a,=slqilai-bai)+qiDi t aitl2.6连续形式下PVQ增量的误差、协方差及 JacobianIMU在每个吋刻积分出来的值是有误差的,下面我们对误差进行分析。首先我们直接给出在t时刻误差项的导数为:sa00016a000000-82(066hkR;0006|=00-(a-bh)0-1192k|+|000|mLL000016ba00101n000018b000F+ozk+ Gt其中:F25×15,G215×2,62x1,n12×,上式推导可参考附录102。下面我们讨论它的作用,将其可以简写为:6之k=F62z+Gtnt根据导数定义可知:62b=1m24-6262+8=62+628t=(+F6t)6z+(Gt6t)nt(11)这里我们对公式(1)的IMU误差运动方程再说明,将上式和EKF对比可知,上式恰好给出了如EKF一般对非线性系统线性化的过程,这里的意义是表示下一个时刻的IMU测量误差与上一个时刻的成线性关系,这样我们根据当前时刻的值,可以预测出下一个时刻的均值和协方差,而公式(1)给出的是均值预测,协方差预测公式如下Pb+6=(1+Ft)P(+Fl6t)7+(G,t)Q(G18t)ot(12)上式给出了协方差的选代公式,初始值Pk=0。其中,Q为表示噪声项的对角协方差矩阵000003000另外根据(11)式可获得诀差项的 Jacobian的迭代公式:(I+F26t)(14)其中 Jacobian的初始值为bk=12.7离散形式的PVQ增量误差分析我们首先直接给出PVQ增量误差在离散形式下的矩阵形式,为了与代码一致,我们修改下变量顺序,这和代码中 midPointIntegration(函数是一致的。(但不知为何计算的V中与前四个噪声项相关的差个负号?)1t fo660f106t‖loeBk+1=0f211f20016bδb0[6b102001rnot000kRkotk+1(15006t0n0000δt其中,推导可参考附录10.3:stE(ak-ba)02-4B+1(kk+121k+1b.)6t|6t2(Rr+ rk+18t2Stn=71=Rk+1(a+1-b)6tWr+ wf1=Ik+11+Gb。)δt-Rn+1(ak+121=-2配+1Stl st21(RK+Ruts)4rula1RrotstStR+1(a1R,+114/+11t28离散形式的PVQ增量误差的 Jacobian和协方差将公式(15)简写为:k+1F15×158215×1+V15×13Q则 Jacobian的迭代公式为k+15×15=F/k(16)其中, Jacobian的初始值为/k=l。这里计算出来的k+1只是为了给后面提供对bias的acoblar。协方差的迭代公式为P+15×15=FPFr+vQv(17)其中,初始值P=0。Q为表示噪声项的对角协方差矩阵:00000000aa000Q18×180a00(18)000000三、后端非线性优化31状态向量状态向量共包括滑动窗口内的n+l1个所有相机的状态(包括位置、朝向、速度、加速度计bias和陀螺仪bias)、 Camera到IMU的外参、m+1个3D点的逆深度X=[xr=pw,vb bpc,q3.2目标函数吗+(喻,2)+2(19)其中三个残差项即误差项分别为边缘化的先验信息、IMU测量残差、视觉的重投影残差。三种残差都是用马氏距离表示。根据《十四讲》中高斯牛顿法,若要计算目标函数的最小值,可以理解为,当优化变量有一个增量后,目标函数值最小,以IU残差为例,可写成如下所示:nin lre2bk, X+8Xrk x)+HSⅩDk+1oXk+1k+1其中HB,为B关于 XIK Jacobian,将上式展开并令关于6X的导数为0,可得增量δx的计算公式:H k 8X=k+1TB那么,公式(28)可写成+∑+∑Tk∑1rc上式中,B为MU预积分噪声项的协方差,P为vual观测的噪声协方差。当MU的噪声协方差P越大时,其信息矩阵Pk,将越小,意味着该MU观测越不可信,换句话说,因MU噪声较大,越不可信IMU预积分数据,而更加相信 visual观测。注意,这里的IMU和vsua协方差的绝对值没有意义,因为考虑得是两者的相对性可将上式继续简化为:(Ap+AB +Acox=bp +bB +bc其中,Ap,AB和Ac为 Hessian矩阵,上述方程称之为增量方程。33MU约束1)残差:两帧之间的PVQ和bias的变化量的差△tx+k+1bk qbk+1bR+1 xyz+g"△t)-Bk(20)sbbbb其中各增量关于bias的 Jacobian可从公式(16)的大 Jacobian中的相应位置获得。上面与代码中 Integration base: evaluateD对应,2)优化变量pb, 0W, Svb ,8ba:,bor Opb,, 80W ,Swb,, bakr, Sba3)Jacobian:计算 Jacobian时,残差对应求偏导对象分别为p6e,6vB,6h,ba],6b,6b
- 2020-12-07下载
- 积分：1
-
800W 手机镜头 ZEMAX数据
4pcs 非球面 参数齐全 可以设计参考之用
- 2020-12-03下载
- 积分：1
-
自制有源隔离RS232转RS485 RS422
【实例简介】自制有源高速光耦隔离RS232转RS485 RS422的电路图
- 2021-11-18 00:50:21下载
- 积分：1
-
android studio 简单获取天气案例
通过网络数据请求获取简单的天气预报,通过解析网络数据显示到UI界面上
- 2020-12-06下载
- 积分：1
