简易跟踪与卡尔曼滤波指南

5星 · 超过95%的资源需积分: 33 75 浏览量更新于2024-07-25 1 收藏 3.83MB PDF 举报

"《追踪与卡尔曼滤波轻松入门》(Tracking and Kalman Filtering Made Easy)是Eli Brookner所著的一本专业书籍，由John Wiley & Sons, Inc.出版于1998年。该书主要针对那些对跟踪技术与卡尔曼滤波方法感兴趣的读者，系统地介绍了这两个在信息技术、信号处理和导航等领域广泛应用的关键概念和技术。卡尔曼滤波是一种数学工具，最初由Rudolf E. Kalman在20世纪50年代提出，用于估计动态系统的状态，特别是在存在噪声和不确定性的情况下。它被广泛用于自动驾驶、航空航天、机器人学、无线通信等多个领域，以提高数据融合的精确性和实时性能。书中详细阐述了卡尔曼滤波的基本原理，包括滤波器的设计、预测步骤、更新步骤以及如何处理观测噪声和过程噪声，使复杂问题变得直观易懂。本书不仅包含了理论知识，还包含了许多实际应用案例，帮助读者理解这些理论在实际场景中的应用。此外，作者作为咨询科学家，可能提供了深入的工业级见解和实践经验，使得学习者能够更好地将理论应用于工程实践中。版权方面，John Wiley & Sons, Inc.强调所有权利受到保护，未经许可，不得进行任何形式的复制、存储、传输或解码，这符合1976年美国版权法的规定。读者在引用或使用书中的内容时，应确保遵守法律并尊重作者的知识产权。《追踪与卡尔曼滤波轻松入门》是一本极具价值的参考书，适合初学者和专业人士，无论是希望深入了解这一领域的研究人员，还是寻求解决实际问题工程师，都能从中受益匪浅。通过阅读这本书，读者可以掌握追踪技术的核心理念，并熟练运用卡尔曼滤波来优化数据处理和决策过程。"

processing orthonormal transformation methods to the DOLP approach used in

Section 5.3 for obtaining a polynomial ﬁt to data. The two methods are shown

to be essentially identical. The square-root Kalman ﬁlter, which is less sensitive

to round-off errors, is discussed in Section 14.5.

Up until now the deterministic part of the target model was assumed to be

time invariant. For example, if a polynomial ﬁt of degree m was used for the

target dynamics, the coefﬁcients of this polynomial ﬁt are constant with time.

Chapter 15 treats the case of time-varying target dynamics.

The Kalman and Bayes ﬁlters developed up until now depend on the

observation scheme being linear. This is not always the situation. For example,

if we are measuring the target range R and azimuth angle  but keep track of the

target using the east-north x, y coordinates of the target with a Kalman ﬁlter,

then errors in the measurement of R and  are not linearly related to the

resulting error in x and y because

x ¼ R cos  ð1Þ

and

y ¼ R sin  ð2Þ

where  is the target angle measured relative to the x axis. Section 16.2 shows

how to simply handle this situation. Basically what is done is to linearize

Eqs. (1) and (2) by using the ﬁrst terms of a Taylor expansion of the inverse

equations to (1) and (2) which are

R ¼

ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ

þ y

ð3Þ

 ¼ tan

ð4Þ

Similarly the equations of motion have to be linear to apply the Kalman–

Bayes ﬁlters. Section 16.3 describes how a nonlinear equation of motion can be

linearized, again by using the ﬁrst term of a Taylor expansion of the nonlinear

equations of motion. The important example of linearization of the nonlinear

observation equations obtained when observing a target in spherical coordinates

(R, , ) while tracking it in rectangular (x, y, z) coordinates is given. The

example of the linearization of the nonlinear target dynamics equations

obtained when tracking a projectile in the atmosphere is detailed. Atmospheric

drag on the projectile is factored in.

In Chapter 17 the technique for linearizing the nonlinear observation

equations and dynamics target equations in order to apply the recursive Kalman

and Bayes ﬁlters is detailed. The application of these linearizations to a

nonlinear problem in order to handle the Kalman ﬁlter is called the extended

Kalman ﬁlter. It is also the ﬁlter Swerling originally developed (without the

xviii PREFACE

ACKNOWLEDGMENT

I would like to thank Fred Daum (Raytheon Company), who ﬁrst educated me

on the Kalman ﬁlter and encouraged me throughout this endeavor. I also would

like to thank Erwin Taenzer (formally of the Raytheon Company), from whose

many memos on the g–k and g–h–k ﬁlters I ﬁrst learned about these ﬁlters. I am

indebted to Barbara Rolinski (formerly of the Raytheon Company), who helped

give birth to this book by typing a good part of its ﬁrst draft, including its many

complicated equations. This she did with great enthusiasm and professionalism.

She would have ﬁnished it were it not that she had to leave to give birth to her

second child. I would also like to thank Barbara Rolinski for educating her

replacement on the typing of the text with its complicated equations using the

VAX Mass-11 Word Processor. I would like to thank Lisa Cirillo (formerly of

the Raytheon Company), Barbara Rolinski’s ﬁrst replacement, for typing the

remainder of the ﬁrst draft of the book. I am most appreciative of the help of

Richard P. Arcand, Jr. (Raytheon Company), who helped Barbara Rolinski and

Lisa Cirillo on difﬁcult points relative to the use of the VAX Mass-11 for typing

the manuscript and for educating Ann Marie Quinn (Raytheon Company) on the

use of the Mass-11 Word Processor. Richard Arcand, Jr. also meticulously made

the second-draft corrections for the ﬁrst part of the book. I am most grateful to

Ann Marie Quinn for retyping some of the sections of the book and making the

legion of corrections for the many successive drafts of the book. Thanks are also

due the Ofﬁce Automation Services (Raytheon Company) for helping to type

the second draft of the second part of the book. Sheryl Evans (Raytheon

Company) prepared many of the ﬁgures and tables for the book and for that I

am grateful. I am grateful to Richard Arcand, Jr. for converting the text to

Microsoft Word on the MAC. I am extremely grateful to Joanne Roche, who

completed the horrendous task of retyping the equations into Microsoft Word

xxi

剩余491页未读，继续阅读

logimaigicyang

粉丝: 2

简易跟踪与卡尔曼滤波指南

轻松掌握卡尔曼滤波追踪：《Tracking and Kalman Filtering Made Easy》简介

IMM Kalman Tracking 示例：交互式多模型算法实现

高效目标跟踪研究：基于改进Mean Shift与粒子滤波的快速方法

TRACKING AND KALMAN FILTERING

Target Tracking with Kalman Filtering, KNN and LSTMs.pdf

Robust On-line Beat Tracking with Kalman Filtering and probabilistic data association

Vehicle Tracking Algorithm Based on Partial Gabor Feature Matching and Kalman Filtering

Target Tracking Based on Mean Shift and Improved Kalman Filtering Algorithm

Trajectory tracking control of wheeled mobile manipulator based on fuzzy neural network and extended Kalman filtering

Real-time tracking of multiple persons by Kalman filtering...

最新资源