MATLAB Machine Learning Recipes: A Problem-

Solution Approach 3rd Edition Michael Paluszek

download

https://textbookfull.com/product/matlab-machine-learning-recipes-
a-problem-solution-approach-3rd-edition-michael-paluszek/

Download more ebook from https://textbookfull.com

We believe these products will be a great fit for you. Click
the link to download now, or visit textbookfull.com
to discover even more!

MATLAB Machine Learning Recipes: A Problem-Solution

Approach 2nd Edition Michael Paluszek

https://textbookfull.com/product/matlab-machine-learning-recipes-
a-problem-solution-approach-2nd-edition-michael-paluszek/

MATLAB Recipes: a problem-solution approach 2nd Edition

Michael Paluszek

https://textbookfull.com/product/matlab-recipes-a-problem-
solution-approach-2nd-edition-michael-paluszek/

MATLAB Machine Learning 1st Edition Michael Paluszek

https://textbookfull.com/product/matlab-machine-learning-1st-
edition-michael-paluszek/

Java 9 Recipes: A Problem-Solution Approach 3rd Edition

Josh Juneau

https://textbookfull.com/product/java-9-recipes-a-problem-
solution-approach-3rd-edition-josh-juneau/
Android Recipes A Problem-Solution Approach Dave Smith

https://textbookfull.com/product/android-recipes-a-problem-
solution-approach-dave-smith/

Raku Recipes: A Problem-Solution Approach J.J. Merelo

https://textbookfull.com/product/raku-recipes-a-problem-solution-
approach-j-j-merelo/

C Recipes - A Problem-Solution Approach 1st Edition

Shirish Chavan

https://textbookfull.com/product/c-recipes-a-problem-solution-
approach-1st-edition-shirish-chavan/

JavaScript Recipes: A Problem-Solution Approach 1st

Edition Russ Ferguson

https://textbookfull.com/product/javascript-recipes-a-problem-
solution-approach-1st-edition-russ-ferguson/

wxPython Recipes: A Problem - Solution Approach 1st

Edition Mike Driscoll

https://textbookfull.com/product/wxpython-recipes-a-problem-
solution-approach-1st-edition-mike-driscoll/
 MATLAB Machine
Learning Recipes
A Problem-Solution Approach

Third Edition

Michael Paluszek
Stephanie Thomas
MATLAB Machine Learning Recipes: A Problem-Solution Approach

Michael Paluszek Stephanie Thomas

Plainsboro, NJ Plainsboro, NJ
USA USA

ISBN-13 (pbk): 978-1-4842-9845-9 ISBN-13 (electronic): 978-1-4842-9846-6

https://doi.org/10.1007/978-1-4842-9846-6

Copyright © 2024 by Michael Paluszek and Stephanie Thomas

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material
is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, repro-
duction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic
adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed.
Trademarked names, logos, and images may appear in this book. Rather than use a trademark symbol with every
occurrence of a trademarked name, logo, or image we use the names, logos, and images only in an editorial fashion
and to the benefit of the trademark owner, with no intention of infringement of the trademark.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not
identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary
rights.
While the advice and information in this book are believed to be true and accurate at the date of publication, neither
the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may
be made. The publisher makes no warranty, express or implied, with respect to the material contained herein.
Managing Director, Apress Media LLC: Welmoed Spahr
Acquisitions Editor: Celestin Suresh John
Development Editor: Laura Berendson
Coordinating Editor: Mark Powers

Cover designed by eStudioCalamar

Cover image by Mohamed Nohassi@Unsplash.com
Distributed to the book trade worldwide by Apress Media, LLC, 1 New York Plaza, New York, NY 10004, U.S.A.
Phone 1-800-SPRINGER, fax (201) 348-4505, e-mail orders-ny@springer-sbm.com, or visit
www.springeronline.com. Apress Media, LLC is a California LLC and the sole member (owner) is Springer Science
+ Business Media Finance Inc (SSBM Finance Inc). SSBM Finance Inc is a Delaware corporation.
For information on translations, please e-mail booktranslations@springernature.com; for reprint, paperback, or
audio rights, please e-mail bookpermissions@springernature.com.
Apress titles may be purchased in bulk for academic, corporate, or promotional use. eBook versions and licenses are
also available for most titles. For more information, reference our Print and eBook Bulk Sales web page at http://
www.apress.com/bulk-sales.
Any source code or other supplementary material referenced by the author in this book is available to readers
on GitHub (https://github.com/Apress). For more detailed information, please visit https://www.apress.com/gp/
services/source-code.

Paper in this product is recyclable

To our families.
Contents

About the Authors XVII

About the Technical Reviewer XIX

Introduction XXI

1 An Overview of Machine Learning 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Elements of Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.2 Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.3 Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 The Learning Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Taxonomy of Machine Learning . . . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.5.1 Kalman Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.5.2 Adaptive Control . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.6 Autonomous Learning Methods . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6.1 Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.6.2 Decision Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6.3 Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.6.4 Support Vector Machines (SVMs) . . . . . . . . . . . . . . . . . . 16
1.7 Artificial Intelligence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.7.1 What Is Artificial Intelligence? . . . . . . . . . . . . . . . . . . . . 17
1.7.2 Intelligent Cars . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.7.3 Expert Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2 Data for Machine Learning in MATLAB 21

2.1 Introduction to MATLAB Data Types . . . . . . . . . . . . . . . . . . . . . 21
2.1.1 Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.1.2 Cell Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.1.3 Data Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1.4 Numerics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1.5 Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

V
CONTENTS

2.1.6 Datastore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.1.7 Tall Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.1.8 Sparse Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.1.9 Tables and Categoricals . . . . . . . . . . . . . . . . . . . . . . . . 30
2.1.10 Large MAT-Files . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.2 Initializing a Data Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3 mapreduce on an Image Datastore . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.4 Processing Table Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.5 String Concatenation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.6 Arrays of Strings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.7 Substrings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.7.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.7.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.7.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.8 Reading an Excel Spreadsheet into a Table . . . . . . . . . . . . . . . . . . . 44
2.8.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.8.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.8.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
2.9 Accessing ChatGPT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.9.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.9.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.9.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

VI
 CONTENTS

3 MATLAB Graphics 49
3.1 2D Line Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2 General 2D Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3 Custom Two-Dimensional Diagrams . . . . . . . . . . . . . . . . . . . . . . 54
3.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.4 Three-Dimensional Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.5 Draw a 3D Object with a Texture . . . . . . . . . . . . . . . . . . . . . . . . 59
3.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.6 General 3D Graphics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.7 Building a GUI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.7.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.7.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.7.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.8 Animating a Bar Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.8.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.8.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.8.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.9 Drawing a Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.9.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.9.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.9.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.10 Importing a Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.10.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.10.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.10.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.11 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

VII
CONTENTS

4 Kalman Filters 85
4.1 Gaussian Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.2 A State Estimator Using a Linear Kalman Filter . . . . . . . . . . . . . . . . 87
4.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.3 Using the Extended Kalman Filter for State Estimation . . . . . . . . . . . . 106
4.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.4 Using the UKF for State Estimation . . . . . . . . . . . . . . . . . . . . . . 111
4.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
4.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.5 Using the UKF for Parameter Estimation . . . . . . . . . . . . . . . . . . . . 117
4.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
4.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
4.6 Range to a Car . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

5 Adaptive Control 127

5.1 Self-Tuning: Tuning an Oscillator . . . . . . . . . . . . . . . . . . . . . . . 128
5.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
5.2 Implement MRAC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.3 Generating a Square Wave Input . . . . . . . . . . . . . . . . . . . . . . . . 140
5.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.4 Demonstrate MRAC for a Rotor . . . . . . . . . . . . . . . . . . . . . . . . 142
5.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
5.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
5.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142

VIII
 CONTENTS

5.5 Ship Steering: Implement Gain Scheduling for Steering Control of a Ship . . 145
5.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
5.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
5.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
5.6 Spacecraft Pointing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
5.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
5.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
5.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
5.7 Direct Adaptive Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.7.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.7.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.7.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
5.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

6 Fuzzy Logic 157

6.1 Building Fuzzy Logic Systems . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
6.2 Implement Fuzzy Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
6.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
6.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
6.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
6.3 Window Wiper Fuzzy Controller . . . . . . . . . . . . . . . . . . . . . . . . 169
6.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
6.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
6.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
6.4 Simple Discrete HVAC Fuzzy Controller . . . . . . . . . . . . . . . . . . . . 174
6.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
6.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
6.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
6.5 Variable HVAC Fuzzy Controller . . . . . . . . . . . . . . . . . . . . . . . . 180
6.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
6.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
6.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189

7 Neural Aircraft Control 191

7.1 Longitudinal Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
7.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
7.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
7.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

IX
CONTENTS

7.2 Numerically Finding Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . 198

7.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
7.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
7.3 Numerical Simulation of the Aircraft . . . . . . . . . . . . . . . . . . . . . . 200
7.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
7.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
7.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
7.4 Activation Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
7.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
7.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
7.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
7.5 Neural Net for Learning Control . . . . . . . . . . . . . . . . . . . . . . . . 204
7.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
7.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
7.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
7.6 Enumeration of All Sets of Inputs . . . . . . . . . . . . . . . . . . . . . . . 208
7.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
7.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
7.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
7.7 Write a Sigma-Pi Neural Net Function . . . . . . . . . . . . . . . . . . . . . 210
7.7.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
7.7.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
7.7.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
7.8 Implement PID Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
7.8.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
7.8.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
7.8.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
7.9 PID Control of Pitch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
7.9.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
7.9.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
7.9.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
7.10 Neural Net for Pitch Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . 223
7.10.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
7.10.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
7.10.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
7.11 Nonlinear Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.11.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.11.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.11.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.12 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

X
 CONTENTS

8 Introduction to Neural Nets 229

8.1 Daylight Detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
8.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
8.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
8.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
8.2 Modeling a Pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
8.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
8.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8.3 Single Neuron Angle Estimator . . . . . . . . . . . . . . . . . . . . . . . . . 235
8.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
8.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
8.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
8.4 Designing a Neural Net for the Pendulum . . . . . . . . . . . . . . . . . . . 240
8.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
8.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
8.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
8.5 XOR Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
8.6 Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
8.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254

9 Classification of Numbers Using Neural Networks 257

9.1 Generate Test Images with Defects . . . . . . . . . . . . . . . . . . . . . . . 257
9.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
9.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
9.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
9.2 Create the Neural Net Functions . . . . . . . . . . . . . . . . . . . . . . . . 262
9.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
9.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
9.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
9.3 Train a Network with One Output Node . . . . . . . . . . . . . . . . . . . . 267
9.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
9.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
9.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
9.4 Testing the Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
9.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
9.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
9.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
9.5 Train a Network with Many Outputs . . . . . . . . . . . . . . . . . . . . . . 273
9.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
9.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
9.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
9.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
XI
CONTENTS

10 Data Classification with Decision Trees 279

10.1 Generate Test Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
10.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
10.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
10.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
10.2 Drawing Trees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
10.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
10.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
10.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
10.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
10.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
10.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
10.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
10.4 Creating a Tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
10.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
10.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
10.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
10.5 Handmade Tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
10.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
10.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
10.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
10.6 Training and Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
10.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
10.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
10.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
10.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

11 Pattern Recognition with Deep Learning 303

11.1 Obtain Data Online for Training a Neural Net . . . . . . . . . . . . . . . . . 305
11.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.2 Generating Training Images of Cats . . . . . . . . . . . . . . . . . . . . . . 305
11.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
11.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
11.3 Matrix Convolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
11.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
11.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
11.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309
11.4 Convolution Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
11.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311

XII
 CONTENTS

11.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311

11.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
11.5 Pooling to Outputs of a Layer . . . . . . . . . . . . . . . . . . . . . . . . . . 312
11.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
11.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
11.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
11.6 Fully Connected Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
11.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
11.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
11.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
11.7 Determining the Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
11.7.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
11.7.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
11.7.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
11.8 Test the Neural Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
11.8.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
11.8.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
11.8.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
11.9 Recognizing an Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
11.9.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
11.9.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
11.9.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
11.10 Using AlexNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
11.10.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
11.10.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
11.10.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326

12 Multiple Hypothesis Testing 327

12.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
12.2 Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
12.2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
12.2.2 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
12.2.3 Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
12.2.4 Measurement Assignment and Tracks . . . . . . . . . . . . . . . . 333
12.2.5 Hypothesis Formation . . . . . . . . . . . . . . . . . . . . . . . . . 334
12.2.6 Track Pruning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
12.3 Billiard Ball Kalman Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
12.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
12.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
12.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
12.4 Billiard Ball MHT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342

XIII
CONTENTS

12.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342

12.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
12.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342
12.5 One-Dimensional Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
12.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
12.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
12.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
12.6 One-Dimensional MHT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
12.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
12.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
12.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
12.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351

13 Autonomous Driving with MHT 355

13.1 Automobile Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
13.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
13.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
13.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356
13.2 Automobile Radar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
13.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
13.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
13.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
13.3 Passing Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
13.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
13.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
13.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
13.4 Automobile Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
13.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
13.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
13.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
13.4.4 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
13.5 Automobile Simulation and the Kalman Filter . . . . . . . . . . . . . . . . . 367
13.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
13.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
13.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
13.6 Automobile Target Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . 371
13.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
13.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
13.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
13.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374

XIV
 CONTENTS

14 Spacecraft Attitude Determination 377

14.1 Star Catalog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
14.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
14.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
14.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
14.2 Camera Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
14.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
14.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
14.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 381
14.3 Celestial Sphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
14.3.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
14.3.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
14.3.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 383
14.4 Attitude Simulation of Camera Views . . . . . . . . . . . . . . . . . . . . . 384
14.4.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
14.4.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
14.4.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
14.5 Yaw Angle Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
14.5.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
14.5.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
14.5.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
14.6 Yaw Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
14.6.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
14.6.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
14.6.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
14.7 Attitude Determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
14.7.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
14.7.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
14.7.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
14.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399

15 Case-Based Expert Systems 401

15.1 Building Expert Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
15.1.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
15.1.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
15.1.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
15.2 Running an Expert System . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
15.2.1 Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
15.2.2 Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
15.2.3 How It Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
15.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410

XV
CONTENTS

A A Brief History 411

B Software for Machine Learning 419

Bibliography 431

Index 435

XVI
About the Authors

Michael Paluszek is President of Princeton Satellite Systems,

Inc. (PSS) in Plainsboro, New Jersey. Michael founded PSS
in 1992 to provide aerospace consulting services. He used
MATLAB to develop the control system and simulations for the
Indostar-1 geosynchronous communications satellite. This led to
the launch of Princeton Satellite Systems’ first commercial MAT-
LAB toolbox, the Spacecraft Control Toolbox, in 1995. Since
then, he has developed toolboxes and software packages for air-
craft, submarines, robotics, and nuclear fusion propulsion, result-
ing in Princeton Satellite Systems’ current extensive product line.
He is working with the Princeton Plasma Physics Laboratory on a compact nuclear fusion re-
actor for energy generation and space propulsion. He is also leading the development of new
power electronics for fusion power systems and working on heat engine–based auxiliary power
systems for spacecraft. Michael is a lecturer at the Massachusetts Institute of Technology.
Prior to founding PSS, Michael was an engineer at GE Astro Space in East Windsor, NJ.
At GE, he designed the Global Geospace Science Polar despun platform control system and led
the design of the GPS IIR attitude control system, the Inmarsat-3 attitude control systems, and
the Mars Observer delta-V control system, leveraging MATLAB for control design. Michael
also worked on the attitude determination system for the DMSP meteorological satellites. He
flew communication satellites on over 12 satellite launches, including the GSTAR III recovery,
the first transfer of a satellite to an operational orbit using electric thrusters.
At Draper Laboratory, Michael worked on the Space Shuttle, Space Station, and submarine
navigation. His Space Station work included designing Control Moment Gyro–based control
systems for attitude control.
Michael received his bachelor’s degree in Electrical Engineering and master’s and engi-
neer’s degrees in Aeronautics and Astronautics from the Massachusetts Institute of Technology.
He is the author of numerous papers and has over a dozen US patents. Michael is the au-
thor of MATLAB Recipes, MATLAB Machine Learning, Practical MATLAB Deep Learning:
A Projects-Based Approach, Second Edition, all published by Apress, and ADCS: Spacecraft
Attitude Determination and Control by Elsevier.

XVII
ABOUT THE AUTHORS

Stephanie Thomas is Vice President of Princeton Satellite Sys-

tems, Inc. in Plainsboro, New Jersey. She received her bache-
lor’s and master’s degrees in Aeronautics and Astronautics from
the Massachusetts Institute of Technology in 1999 and 2001.
Stephanie was introduced to the PSS Spacecraft Control Tool-
box for MATLAB during a summer internship in 1996 and has
been using MATLAB for aerospace analysis ever since. In her
over 20 years of MATLAB experience, she has developed many
software tools, including the Solar Sail Module for the Spacecraft
Control Toolbox, a proximity satellite operations toolbox for the
Air Force, collision monitoring Simulink blocks for the Prisma
satellite mission, and launch vehicle analysis tools in MATLAB and Java. She has developed
novel methods for space situation assessment, such as a numeric approach to assessing the
general rendezvous problem between any two satellites implemented in both MATLAB and
C++. Stephanie has contributed to PSS’ Spacecraft Attitude and Orbit Control textbook, featur-
ing examples using the Spacecraft Control Toolbox, and written many software User’s Guides.
She has conducted SCT training for engineers from diverse locales such as Australia, Canada,
Brazil, and Thailand and has performed MATLAB consulting for NASA, the Air Force, and
the European Space Agency. Stephanie is the author of MATLAB Recipes, MATLAB Machine
Learning, and Practical MATLAB Deep Learning published by Apress. In 2016, she was named
a NASA NIAC Fellow for the project “Fusion-Enabled Pluto Orbiter and Lander.” Stephanie
is an Associate Fellow of the American Institute of Aeronautics and Astronautics (AIAA) and
Vice Chair of the AIAA Nuclear and Future Flight Propulsion committee. Her ResearchGate
profile can be found at https://www.researchgate.net/profile/Stephanie-Thomas-2.

XVIII
About the Technical Reviewer

Joseph Mueller took a new position as Principal Astrodynamics

Engineer at Millennium Space Systems in 2023, where ground-
breaking capabilities in space are being built, one small satellite
at a time, and he is honored to be a part of it.
From 2014 to 2023, he was a senior researcher at Smart In-
formation Flow Technologies, better known as SIFT. At SIFT, he
worked alongside amazing people, playing in the sandbox of in-
credibly interesting technical problems. His research projects at
SIFT included navigation and control for autonomous vehicles,
satellite formation flying, space situational awareness, and robotic
swarms.
Joseph is married and is a father of three, living in Champlin, MN.
His Google Scholar profile can be found at https://scholar.google.com/citations?hl=en&
user=breRtVUAAAAJ and his ResearchGate profile at www.researchgate.net/profile/Joseph-
Mueller-2.

XIX
Introduction

Machine Learning is becoming important in every engineering discipline. For example:

1. Autonomous cars: Machine learning is used in almost every aspect of car control systems.

2. Plasma physicists use machine learning to help guide experiments on fusion reactors.
TAE Technologies has used it with great success in guiding fusion experiments. The
Princeton Plasma Physics Laboratory (PPPL) has used it for the National Spherical Torus
Experiment to study a promising candidate for a nuclear fusion power plant.

3. It is used in finance for predicting the stock market.

4. Medical professionals use it for diagnoses.

5. Law enforcement and others use it for facial recognition. Several crimes have been solved
using facial recognition!

6. An expert system was used on NASA’s Deep Space 1 spacecraft.

7. Adaptive control systems steer oil tankers.

There are many, many other examples.

While many excellent packages are available from commercial sources and open source
repositories, it is valuable to understand how these algorithms work. Writing your own al-
gorithms is valuable both because it gives you insight into the commercial and open source
packages and also because it gives you the background to write your custom Machine Learning
software specialized for your application.
MATLAB had its origins for that very reason. Scientists who needed to do operations
on matrices used numerical software written in FORTRAN. At the time, using computer lan-
guages required the user to go through the write-compile-link-execute process which was time-
consuming and error-prone. MATLAB presented the user with a scripting language that allowed
the user to solve many problems with a few lines of a script that executed instantaneously. MAT-
LAB has built-in visualization tools that helped the user better understand the results. Writing
MATLAB was a lot more productive and fun than writing FORTRAN.
The goal of MATLAB Machine Learning Recipes: A Problem-Solution Approach is to help
all users harness the power of MATLAB to solve a wide range of learning problems. The book
has something for everyone interested in Machine Learning. It also has material that will allow
people with an interest in other technology areas to see how Machine Learning, and MATLAB,
can help them solve problems in their areas of expertise.

XXI
INTRODUCTION

Using the Included Software

This textbook includes a MATLAB toolbox that implements the examples. The toolbox
consists of

1. MATLAB functions

2. MATLAB scripts

3. HTML help

The MATLAB scripts implement all of the examples in this book. The functions encapsulate the
algorithms. Many functions have built-in demos. Just type the function name in the command
window, and it will execute the demo. The demo is usually encapsulated in a subfunction. You
can copy out this code for your demos and paste it into a script. For example, type the function
name PlotSet into the command window, and the plot in Figure 1 will appear.
1 >> PlotSet

cos
1
A
B
0.5

0
y

-0.5

-1
0 100 200 300 400 500 600 700 800 900 1000
x

sin
1

0.5

0
y

-0.5

-1
0 100 200 300 400 500 600 700 800 900 1000
x

Figure 1: Example plot from the function PlotSet.m

XXII
 INTRODUCTION

If you open the function, you will see the demo:

1 %%% PlotSet>Demo
2 function Demo
3
4 x = linspace(1,1000);
5 y = [sin(0.01*x);cos(0.01*x);cos(0.03*x)];
6 disp('PlotSet: One x and two y rows')
7 PlotSet( x, y, 'figure title', 'PlotSet Demo',...
8 'plot set',{[2 3], 1},'legend',{{'A' 'B'},{}},'plot title',{'cos','
sin'});

You can use these demos to start your scripts. Some functions, like right-hand-side functions
for numerical integration, don’t have demos. If you type a function name at the command line
that doesn’t have a built-in demo, you will get an error as in the code snippet below.
1 >> RHSAutomobileXY
2 Error using RHSAutomobileXY (line 17)
3 A built-in demo is not available.

The toolbox is organized according to the chapters in this book. The folder names are
Chapter_01, Chapter_02, etc. In addition, there is a General folder with functions that support
the rest of the toolbox. In addition, you will need the open source package GLPK (GNU Linear
Programming Kit) to run some of the code. Nicolo Giorgetti has written a MATLAB mex
interface to GLPK that is available on SourceForge and included with this toolbox. The interface
consists of

1. glpk.m

2. glpkcc.mexmaci64, or glpkcc.mexw64, etc.

3. GLPKTest.m

which are available from https://sourceforge.net/projects/glpkmex/. The second item is the

mex file of glpkcc.cpp compiled for your machine, such as Mac or Windows. Go to www.gnu.
org/software/glpk/ to get the GLPK library and install it on your system. If needed, download
the GLPKMEX source code as well and compile it for your machine, or else try another of the
available compiled builds.

XXIII
CHAPTER 1

An Overview of Machine Learning

1.1 Introduction
Machine Learning is a field in computer science where data is used to predict, or respond to,
future data. It is closely related to the fields of pattern recognition, computational statistics, and
artificial intelligence. The data may be historical or updated in real time. Machine learning is
important in areas like facial recognition, spam filtering, content generation, and other areas
where it is not feasible, or even possible, to write algorithms to perform a task.
For example, early attempts at filtering junk emails had the user write rules to determine
what was junk or spam. Your success depended on your ability to correctly identify the attributes
of the message that would categorize an email as junk, such as a sender address or words in the
subject, and the time you were willing to spend to tweak your rules. This was only moderately
successful as junk mail generators had little difficulty anticipating people’s handmade rules.
Modern systems use machine learning techniques with much greater success. Most of us are
now familiar with the concept of simply marking a given message as “junk” or “not junk” and
take for granted that the email system can quickly learn which features of these emails identify
them as junk and prevent them from appearing in our inbox. This could now be any combination
of IP or email addresses and words and phrases in the subject or body of the email, with a variety
of matching criteria. Note how the machine learning in this example is data driven, autonomous,
and continuously updating itself as you receive emails and flag them. However, even today, these
systems are not completely successful since they do not yet understand the “meaning” of the
text that they are processing.
Content generation is an evolving area. By training engines over massive data sets, the
engines can generate content such as music scores, computer code, and news articles. This has
the potential to revolutionize many areas that have been exclusively handled by people.
In a more general sense, what does machine learning mean? Machine learning can mean
using machines (computers and software) to gain meaning from data. It can also mean giving
machines the ability to learn from their environment. Machines have been used to assist humans
for thousands of years. Consider a simple lever, which can be fashioned using a rock and a length
of wood, or an inclined plane. Both of these machines perform useful work and assist people,
but neither can learn. Both are limited by how they are built. Once built, they cannot adapt to
changing needs without human interaction.

© The Author(s), under exclusive license to APress Media, LLC, part of Springer Nature 2024 1
M. Paluszek, S. Thomas, MATLAB Machine Learning Recipes,
https://doi.org/10.1007/978-1-4842-9846-6 1
CHAPTER 1 OVERVIEW

Machine learning involves using data to create a model that can be used to solve a problem.
The model can be explicit, in which case the machine learning algorithm adjusts the model’s
parameters, or the data can form the model. The data can be collected once and used to train a
machine learning algorithm, which can then be applied. For example, ChatGPT scrapes textual
data from the Internet to allow it to generate text based on queries. An adaptive control system
measures inputs and command responses to those inputs to update parameters for the control
algorithm.
In the context of the software we will be writing in this book, machine learning refers to
the process by which an algorithm converts the input data into parameters it can use when
interpreting future data. Many of the processes used to mechanize this learning derive from
optimization techniques and, in turn, are related to the classic field of automatic control. In
the remainder of this chapter, we will introduce the nomenclature and taxonomy of machine
learning systems.

1.2 Elements of Machine Learning

This section introduces key nomenclature for the field of machine learning.

1.2.1 Data
All learning methods are data driven. Sets of data are used to train the system. These sets may
be collected and edited by humans or gathered autonomously by other software tools. Control
systems may collect data from sensors as the systems operate and use that data to identify pa-
rameters or train the system. Content generation systems scour the Internet for information. The
data sets may be very large, and it is the explosion of data storage infrastructure and available
databases that is largely driving the growth in machine learning software today. It is still true that
a machine learning tool is only as good as the data used to create it, and the selection of training
data is practically a field in itself. Selection of data for many systems is highly automated.

NOTE When collecting data for training, one must be careful to ensure that the time
variation of the system is understood. If the structure of a system changes with time, it may be
necessary to discard old data before training the system. In automatic control, this is sometimes
called a forgetting factor in an estimator.

1.2.2 Models
Models are often used in learning systems. A model provides a mathematical framework for
learning. A model is human-derived and based on human observations and experiences. For
example, a model of a car, seen from above, might be that it is rectangular with dimensions that
fit within a standard parking spot. Models are usually thought of as human-derived and provide
a framework for machine learning. However, some forms of machine learning develop their
models without a human-derived structure.

2
 CHAPTER 1 OVERVIEW

1.2.3 Training
A system which maps an input to an output needs training to do this in a useful way. Just as
people need to be trained to perform tasks, machine learning systems need to be trained. Train-
ing is accomplished by giving the system an input and the corresponding output and modifying
the structure (models or data) in the learning machine so that mapping is learned. In some ways,
this is like curve fitting or regression. If we have enough training pairs, then the system should
be able to produce correct outputs when new inputs are introduced. For example, if we give
a face recognition system thousands of cat images and tell it that those are cats, we hope that
when it is given new cat images it will also recognize them as cats. Problems can arise when you
don’t give it enough training sets, or the training data is not sufficiently diverse, for instance,
identifying a long-haired cat or hairless cat when the training data is only of short-haired cats.
A diversity of training data is required for a functioning algorithm.
Supervised Learning
Supervised learning means that specific training sets of data are applied to the system. The
learning is supervised in that the “training sets” are human-derived. It does not necessarily
mean that humans are actively validating the results. The process of classifying the systems’
outputs for a given set of inputs is called “labeling.” That is, you explicitly say which results are
correct or which outputs are expected for each set of inputs.
The process of generating training sets can be time-consuming. Great care must be taken
to ensure that the training sets will provide sufficient training so that when real-world data is
collected, the system will produce correct results. They must cover the full range of expected
inputs and desired outputs. The training is followed by test sets to validate the results. If the
results aren’t good, then the test sets are cycled into the training sets, and the process is repeated.
A human example would be a ballet dancer trained exclusively in classical ballet technique.
If they were then asked to dance a modern dance, the results might not be as good as required
because the dancer did not have the appropriate training sets; their training sets were not suffi-
ciently diverse.
Unsupervised Learning
Unsupervised learning does not utilize training sets. It is often used to discover patterns in data
for which there is no “right” answer. For example, if you used unsupervised learning to train
a face identification system, the system might cluster the data in sets, some of which might be
faces. Clustering algorithms are generally examples of unsupervised learning. The advantage
of unsupervised learning is that you can learn things about the data that you might not know in
advance. It is a way of finding hidden structures in data.

3
CHAPTER 1 OVERVIEW

Semi-supervised Learning
With this approach, some of the data are in the form of labeled training sets, and other data are
not [12]. Typically, only a small amount of the input data is labeled, while most are not, as the
labeling may be an intensive process requiring a skilled human. The small set of labeled data is
leveraged to interpret the unlabeled data.
Online Learning
The system is continually updated with new data [12]. This is called “online” because many of
the learning systems use data collected while the system is operating. It could also be called
recursive learning. It can be beneficial to periodically “batch” process data used up to a given
time and then return to the online learning mode. The spam filtering systems collect data from
emails and update their spam filter. Generative deep learning systems like ChatGPT use massive
online learning.

1.3 The Learning Machine

Figure 1.1 shows the concept of a learning machine. The machine absorbs information from
the environment and adapts. The inputs may be separated into those that produce an immediate
response and those that lead to learning. In some cases, they are completely separate. For ex-
ample, in an aircraft, a measurement of altitude is not usually used directly for control. Instead,
it is used to help select parameters for the actual control laws. The data required for learning
and regular operation may be the same, but in some cases, separate measurements or data will
be needed for learning to take place. Measurements do not necessarily mean data collected by
a sensor such as radar or a camera. It could be data collected by polls, stock market prices, data
in accounting ledgers, or any other means. Machine learning is then the process by which the
measurements are transformed into parameters for future operation.

Measurements (Learning)

Learning

Parameters Actions

Machine
Environment
Actions

Measurements (Immediate Use)

Figure 1.1: A learning machine that senses the environment and stores data in memory

4
 CHAPTER 1 OVERVIEW

Note that the machine produces output in the form of actions. A copy of the actions may
be passed to the learning system so that it can separate the effects of the machine’s actions
from those of the environment. This is akin to a feedforward control system, which can result
in improved performance.
A few examples will clarify the diagram. We will discuss a medical example, a security
system, and spacecraft maneuvering.
A doctor might want to diagnose diseases more quickly. They would collect data on tests
on patients and then collate the results. Patient data might include age, height, weight, historical
data like blood pressure readings and medications prescribed, and exhibited symptoms. The
machine learning algorithm would detect patterns so that when new tests were performed on a
patient, the machine learning algorithm would be able to suggest diagnoses or additional tests to
narrow down the possibilities. As the machine learning algorithm was used, it would, hopefully,
get better with each success or failure. Of course, the definition of success or failure is fuzzy. In
this case, the environment would be the patients themselves. The machine would use the data
to generate actions, which would be new diagnoses. This system could be built in two ways.
In the supervised learning process, test data and known correct diagnoses are used to train the
machine. In an unsupervised learning process, the data would be used to generate patterns that
might not have been known before, and these could lead to diagnosing conditions that would
normally not be associated with those symptoms.
A security system might be put into place to identify faces. The measurements are camera
images of people. The system would be trained with a wide range of face images taken from
multiple angles. The system would then be tested with these known persons and its success rate
validated. Those that are in the database memory should be readily identified, and those that are
not should be flagged as unknown. If the success rate was not acceptable, more training might
be needed, or the algorithm itself might need to be tuned. This type of face recognition is now
common, used in Mac OS X’s “Faces” feature in Photos, face identification on the new iPhone
X, and Facebook when “tagging” friends in photos.
For precision maneuvering of a spacecraft, the inertia of the spacecraft needs to be known.
If the spacecraft has an inertial measurement unit that can measure angular rates, the inertia
matrix can be identified. This is where machine learning is tricky. The torque applied to the
spacecraft, whether by thrusters or momentum exchange devices, is only known to a certain
degree of accuracy. Thus, the system identification must sort out, if it can, the torque scaling
factor from the inertia. The inertia can only be identified if torques are applied. This leads to
the issue of stimulation. A learning system cannot learn if the system to be studied does not
have known inputs, and those inputs must be sufficiently diverse to stimulate the system so that
the learning can be accomplished. Training a face recognition system with one picture will not
work.

5
CHAPTER 1 OVERVIEW

1.4 Taxonomy of Machine Learning

In this book, we take a larger view of machine learning than is normal. Machine learning as de-
scribed earlier is the collecting of data, finding patterns, and doing useful things based on those
patterns. We expand machine learning to include adaptive and learning control. These fields
started independently but now are adapting technology and methods from machine learning.
Figure 1.2 shows how we organize the technology of machine learning into a consistent
taxonomy. You will notice that we created a title that encompasses three branches of learning;
we call the whole subject area “Autonomous Learning.” That means learning without human
intervention during the learning process. This book is not solely about “traditional” machine
learning. Other, more specialized books focus on any one of the machine learning topics. Op-
timization is part of the taxonomy because the results of optimization can be discoveries, such
as a new type of spacecraft or aircraft trajectory. Optimization is also often a part of learning
systems.

Autonomous
Learning

Control Machine Learning

State Inductive
Estimation Learning

Pattern
Adaptive Recognition Expert
Control Systems

Data Mining
System

Fuzzy Logic

Optimal
Control Optimization

Figure 1.2: Taxonomy of machine learning. The dotted lines show connections between branches

6
 CHAPTER 1 OVERVIEW

There are three categories under Autonomous Learning. The first is Control. Feedback con-
trol is used to compensate for uncertainty in a system or to make a system behave differently
than it would normally behave. If there was no uncertainty, you wouldn’t need feedback. For
example, if you are a quarterback throwing a football at a running player, assume for a moment
that you know everything about the upcoming play. You know exactly where the player should
be at a given time, so you can close your eyes, count, and just throw the ball to that spot. As-
suming the player has good hands, you would have a 100% reception rate! More realistically,
you watch the player, estimate the player’s speed, and throw the ball. You are applying feedback
to the problem. As stated, this is not a learning system. However, if now you practice the same
play repeatedly, look at your success rate, and modify the mechanics and timing of your throw
using that information, you would have an adaptive control system, the second box from the top
of the control list. Learning in control takes place in adaptive control systems and also in the
general area of system identification.
System identification is learning about a system. By system, we mean the data that rep-
resents anything and the relationships between elements of that data. For example, a particle
moving in a straight line is a system defined by its mass, the force on that mass, its velocity,
and its position. The position is related to the velocity times time, and the velocity is related and
determined by the acceleration, which is the force divided by the mass.
Optimal control may not involve any learning. For example, what is known as full-state
feedback produces an optimal control signal but does not involve learning. In full-state feed-
back, the combination of model and data tells us everything we need to know about the system.
However, in more complex systems, we can’t measure all the states and don’t know the param-
eters perfectly, so some form of learning is needed to produce “optimal” or the best possible re-
sults. In a learning system, optimal control would need to be redefined as the system learns. For
example, an optimal space trajectory assumes thruster characteristics. As a mission progresses,
the thruster performance may change, requiring recomputation of the “optimal” trajectory.
System identification is the process of identifying the characteristics of a system. A sys-
tem can, to a first approximation, be defined by a set of dynamical states and parameters. For
example, in a linear time-invariant system, the dynamical equation is

ẋ = Ax + Bu
. (1.1)

where A and B are matrices of parameters, u is an input vector, and x is the state vector. System
identification would find A and B. In a real system, A and B are not necessarily time invariant,
and most systems are only linear to a first approximation.
The second category is what many people consider true Machine Learning. This is mak-
ing use of data to produce behavior that solves problems. Much of its background comes from
statistics and optimization. The learning process may be done once in a batch process or con-
tinually in a recursive process. For example, in a stock buying package, a developer might have
processed stock data for several years, say before 2008, and used that to decide which stocks
to buy. That software might not have worked well during the financial crash. A recursive pro-
gram would continuously incorporate new data. Pattern recognition and data mining fall into
this category. Pattern recognition is looking for patterns in images. For example, the early AI

7
Random documents with unrelated
content Scribd suggests to you:
this on this side, and nothing on that side, which should I take? I
lose my resolution,—I don't know; I feel that, by a person to whom I
had no objection in any other shape, I might perhaps be
superinduced to do as others have done, and to make a sacrifice, for
the sake of spending our lives in that kind of domestic combination
which binds people together more than anything else ever can. I am
weak on that point, I know; but then, the home affections, as Mr.
Longstaff says, constitute a very worthy and amiable weakness.”
Miss Sowersoft uttered this last sentence in such a peculiar tone of
self-satisfied depreciation, as evidently proved that she considered
herself a much more eligible subject, on account of that identical
weakness which she had verbally condemned, than she would have
been if wholly free from it.
“Well, meesis,” replied Mr. Palethorpe, with considerate
deliberation, “I should have no objection to our union, if it so
happened that we were not doing very well as we are at present;
and, while we are making a little money to put by every week, I
think it is as well just now to let good alone. I should like—”
“Oh, you misunderstand me!” exclaimed Miss Sowersoft; “I did not
make any allusions to you in particular. Oh, no! I have had very
many most excellent offers, and could have them now for that
matter; but then, you see, I was only just saying, as the thought
came across my mind, that there is something to be said against
being married, and something against keeping single. I remember
the time when I could not bear the very thoughts of a man about
me; but, somehow, as one gets older we see so much more of the
world, and one's ideas change almost as much as one's bodies;
really, I am as different as another woman to what I once was.
Somehow, I don't know how, but so it happens—Ah!” shrieked Miss
Sowersoft, interrupting herself in the demonstration of this very
metaphysical and abstruse point in her discourse, “take hold of me,
dear,—take hold of me! I've trod on a toad, I believe!”
At the same time she threw her arms up to Mr. Palethorpe for
protection; and, very accidentally, of course, they chanced to alight
round that worthy's neck. A round dozen of rough-bearded kisses,
which even he, stoic as he was, could not refrain from bestowing
upon her, in order to revive and restore her spirits, smacked loudly
on the dusky air, and set poor little Colin a-laughing in spite of
himself.
“Who the deuce is that!” earnestly whispered the farming-man.
“There's somebody under the brook bank!” and, as he instantly
disengaged Miss Sowersoft from his arms, he rushed round the
holly-bushes, and caught fast hold of Colin, just as that unlucky lad
was making a speedy retreat across the rivulet into the opposite
orchard. “What! it is you, you young divel, is it?” exclaimed he in a
fury, as he dragged the boy up the sloping bank, and bestowed upon
him sundry kicks, scarcely inferior to those of a vicious horse, with
his heavy, clench-nailed, quarter-boots. “You 're listening after your
meesis, now, are you? Dang your meddling carcass! I 'll stop your
ears for you!”
And bang went his ponderous fist on Colin's organs of
Secretiveness and Acquisitiveness, until his head sung again
throughout, like a seething caldron.
“That's right!” cried Miss Sowersoft; “make him feel; drag him up;
my face burns with shame at him; I'm as hot as a scarlet-fever, I am
—a young scoundrel!”
And Colin was pulled up on to the level of the garden, more like a
half-killed rat than a half-grown human being.
“We'll know how this is, meesis,” said Mr. Palethorpe, when he had
fairly landed his cargo. “I 'll see to the bottom of it before he goes
into th' house. He sha'n't have a chance of being backed up in his
impudence as he was t'other night.”
“Take him into the thrashing-barn,” advised Miss Sowersoft, “and
we can have him there in private.”
Colin now found breath to put in a protest against the bill of
indictment which they were preferring against him.
“I was not listening,” said he; “I was only writing a letter to my
mother, I 'm sure!”
 “What! at dark hour?” ejaculated Palethorpe with a laugh. “Come
along, you young liar! you shan't escape that way.” Accordingly he
dragged the lad up the garden, and behind the house, into the
spacious barn, of which Miss Sowersoft had spoken: and, while that
innocent lady went to procure a lantern, her favourite held him
tightly by the collar; save when, occasionally, to beguile the time
until her return, he regaled him with a severe shake, and an
additional curse or two upon his vagabond and mischievous carcass.
“Do you think he knows anything about it?” asked Miss Sowersoft
aside to Palethorpe, as she entered the barn, and the dim light of
her horn-lantern summoned to view the spectral appearances—
rather than the distinct objects themselves—of various implements
of husbandry, and of heaps of thrashed wheat and straw scattered
around.
“Well, I don't know; but I should think not much,” said he.
“I hope not,” rejoined his mistress, “or it will get into everybody's
mouth. But we will question him very closely; we 'll have it out of
him by hook or by crook.”
She then held a broken side of the lantern a little above Colin's
face, in order to cast the better light upon it; and proceeded to
question the culprit.
“Now, before I ask you a single question, promise to tell me the
truth, and nothing but the truth. Now, mark; I shall know whether
you speak the truth or not, so it will be of no use to try to deceive
me. Tell me whether you heard me and Mr. Palethorpe talking in the
garden; and whether you saw him pick me up so very kindly when I
slipped down; and then tell me for what purpose you were standing
behind those trees? No falsehoods, now. The truth, nothing else.
Take care; because if you say anything untrue I shall know it
directly; and then woe be to you for your trouble?”
“I always do tell truth,” replied Colin, crying, “without being
frightened into it that way. I'm sure I had only been writing a letter
to my mother and Fanny; and I stood there because I did not want
anybody to catch me.”
 “And why did not you want anybody to catch you?”
“Why, because I didn't,” answered Colin.
“Because you didn't!” exclaimed Mr. Palethorpe, as he emerged
from out the shadow of Miss Sowersoft's figure; “what answer is
that, you sulky ill-looking whelp? Give meesis a proper answer, or I 'll
send my fist in your face in a minnit!”
Miss Sowersoft put her hand on Palethorpe's arm to keep him
back,—not so much to prevent him carrying his threat into
execution, as because his interference seemed to imply a doubt of
her own abilities in worming all she wanted to know out of the boy
before her.
“But why didn't you?” she asked again, more emphatically.
“Because they might want to read my letter.”
“Oh,—there's something in it not to be seen, is there?” continued
the inquisitor, as her cheeks reddened with fears of she knew not
what.
“It is all truth, every word of it!” contended Colin.
“Ay, ay, my lad, we must see about that. I cannot let you send a
whole pack of falsehoods over to Bramleigh, and make as much
mischief in my family as your mother made in Mr. Longstaff's. It is
needful to look after your doings. Is the letter in your pocket?”
Having received an answer in the affirmative, she directed
Palethorpe to search him for it; an operation which that amiable
individual very soon concluded by drawing the desired document
from his trowsers.
“Oh, this is it, is it?” said Miss Sowersoft, as she partly opened it
to assure herself. “Well, well,” folding it up again: “we'll read this by
and by. Now, what did you hear us talking about? If you say
anything shameful, now, and we shall know whether it is true or not
directly that we hear it,—if you do not say something—a—. You
know what Scripture tells you, always to speak well of your mistress
and master. Be careful, now. What did we say?”
 “Please, 'um,” replied Colin, “you said, that when people get
married they strike a balance between them; and that if one thing
was on one side, and nothing on the other, you should lose your
resolution, and make a sacrifice of the little you possess, whatever it
is.”
“Oh, you little wretch!” ejaculated Maria. “Go on with your lies, go
on! and you shall have it on your shoulders when you have done.
What else, you vile toad?”
Colin stood mute.
“What next, I say!” stormed the lady, with a furious stamp of the
right foot.
“Why, then, mum,” added Colin, “I heard Palethorpe kiss you.”
“Kiss me!—kiss me, you young rascal!” and the face of Miss
Sowersoft became as red as the gills of one of her own turkey-cocks
at the discovery. “If you dare to say such a thing as that again, I 'll
strip the very skin off your back,—I will, you caitiff! Kiss me, indeed!
A pretty tale to tell as ever I heard!”
“I'm sure it's true,” blubbered the boy; “for I heard it ever so many
times.”
“Oh!” exclaimed the virtuous Miss Sowersoft, “so we have got it
out of you at last. What!—your mother has set you to watch your
mistress, has she? That's all her schooling, is it? But Mr. Palethorpe
shall learn you to spy about this house,—He shall, you dog!”
That worthy was now about to pounce upon his victim, but was
again arrested by his mistress.
“Stop! stop!—we have not done yet,” pulling the letter before
mentioned from her bosom; “there is a pretty budget here, I 'll be
bound to say. After such as this, we may expect anything. There is
nothing too bad for him.”
While Palethorpe held the culprit fast by one hand, and the lantern
in the other, he and Miss Sowersoft enjoyed the high gratification of
perusing together the said letter which follows:—
“Dear Mother and Fanny,
 “As I promised to write if they would not let me come on Sunday,
which they did not do, I take this opportunity after tea to tell you all
about it. I like this house very well, and have caught fourteen rats
with traps of my own setting, besides helping Abel to shoot
forwards, which he fired at, and I looked on while. I can harness a
horse and curry him down already. But when I first got here I did
not think I should like it at all, as Palethorpe flew at me like a yard-
dog because I spoke to him, and Miss Sowersoft was mangling, and
as cross as patch. I did think of coming home again; but then I said
to myself, 'Well, I'll lay a penny if I do, mother will send me back; so
it will be of no use, and I shall have my walk for nothing.' I do not
like mistress a bit. When she was at our house, she told you a pack
of the biggest fibs in the world. I never beard of a bigger fibber than
she is in my life; for all the good victuals she made such a bother
about are made up for Palethorpe. He is like a master-pig in a sty,
because he crunches up the best of everything. Mistress seems very
fond of him, though; for after we had had a shindy the first night,
and Palethorpe made my nose bleed, I went to bed, and saw her tie
her nightcap on his head, and feed him with a posset. I could not
help laughing, he looked such a fool. Then I heard her courting him
as plain as sunshine; for she tries as hard as she can to get him to
marry her; but I would not have her, if I were him, she is so very
mean and pretending. But then he is a savage idle fellow himself:
and as Abel said to him, said he, 'You never touch plough nor bill-
hook once a-week,'—no more he does. Our mistress backs him up in
it, and that is the reason. I shall come over as soon as I can, as I
want to see you and Fanny very much indeed.
“Yours affectionately,
“Colin Clink.”
At all events the murder was now out, and no mistake. The letter
dropped from Miss Sowersoft's hand, and she almost fainted in Mr.
Palethorpe's arms, as she faintly sighed, “Oh!—he 'll be the death of
me!”
When Miss Sowersoft was somewhat recovered, Palethorpe turned
in great wrath towards Colin, uttering a more fearful asseveration
than I can repeat, that if he could make no better use than that of
his eyes when he went to bed, he would knock them out of his head
for him. Seizing the boy ferociously by the nape of the neck with one
hand, and a portion of his clothes with the other, he lifted him from
the ground, like a dog by head and tail, and carried him straight into
the yard, dashing him violently into the horse-trough, very much to
the satisfaction of the indignant Miss Sower-soft, who had suddenly
recovered on beholding this spectacle, and followed her favourite
with the lantern. While Palethorpe held him down in the trough, Miss
Sowersoft proceeded with great alacrity to pump upon him very
vigorously until her arms were tired.
The boy's cries soon brought several of the domestics of the
establishment together. Sally rushed out of her kitchen inquiring
what Colin had done to be ducked.
“Spying after the secrets of other people!” exclaimed the wrathful
Mr. Palethorpe.
“Spying!” echoed the maid.
“Yes, spying!” added Miss Sowersoft, in corroboration of
Palethorpe's statement. “We have caught him out, according to his
own confession, in spying after the secrets of everybody about the
premises, and sending it all in writing to his mother!”
“Ay! I'd souse him well!” observed Sally, who began to fear that
some of her own secret interviews with Abel had very probably been
registered in black and white, for the edification of the good people
of Bramleigh.
“What has he been a-gate of?” asked Abel, who had come up just
in time to catch the end of the above conversation.
“Oh, he's been watching you come into the dairy when I was
there!” added Sally, accompanying her remark with a broad simper,
and a sly blushing glance at Abel, which caused Abel to shuffle on
his feet, and dangle his legs about, as though at a loss what to do
with them.
“Then a sheep-washing will do him no harm for sheep's eyes,”
rejoined Abel, rounding off his sharp-pointed wit with a broad laugh.
 When the ducking was concluded, they drove him, bruised,
drenched, and weeping, into the kitchen. Old George, who had been
a distant and silent spectator of the scene, stood at the door as he
entered.
“Ay, poor boy!” said he, pityingly, as the child passed by him,
“they'd more need to nurse him by the fireside than half drown him
this way. It's sad wages—sad wages, indeed, for a nest-babe like
him! But they don't heed what I say. I'm an old man, and have no
right to speak.”
Miss Sowersoft seized the earliest opportunity she could to place
Colin's letter upon the fire, which she did with a spoonful of salt
upon it, in order that its flames should be of the same colour as its
contents.
In the mean time Colin had shuffled off his mortal coil of wet
clothes, and in a moist skin gone silently off to bed. At supper-time
old George carried him up the pint of warm ale which had been
served out for himself. Colin accepted it, less because he relished it,
than because he knew not how at that moment to refuse the hand
by which it was offered; and within ten minutes afterwards,
notwithstanding all his troubles, he fell into a sound state of repose.
 CHAPTER XIV.
The benefits of being soused in a horse-trough.—Some farther
specimens of Miss Sowersoft's moral excellence.—An unlooked-for
discovery is partially made, which materially concerns Miss Fanny
Woodruff and Dr. Rowel.

O
N the following morning Palethorpe arose, and finding Colin
still asleep, was proceeding, whip in hand, to help him up
according to custom, when, as he turned down the clothes
that almost enveloped the child's head, the unusual
appearance of his countenance arrested the man's attention as well
as his hand. His veins were swollen with rapid bounding blood, and
his heart thumped audibly in its place, and with doubly accelerated
motion, as though eagerly hastening to beat out its appointed
number of pulsations, and leave the little harassed life it contained
again free from the pains and vexations of this lower world.
Something like remorse passed for a moment over the man's dark
countenance as he gazed. What had they done to him?—what was
amiss? He covered the boy carefully up again, and hastened down
stairs to communicate the news to Miss Sowersoft.
“Oh,—it's all nonsense!” she exclaimed, on hearing all that Mr.
Palethorpe had to say about it. “The lad's got a bit of a cold,—that's
all. I 'll make him a basin of milk, with a little of that nice feverfew
out of the garden boiled in it, and then if you wake him up, and let
him take that, it will stick to his ribs, and do him an amazing deal of
good.”
But as there was no hurry about such a matter, Miss Sowersoft
very leisurely took her own breakfast before she set about carrying
her very charitable project into execution. When the milk, with some
sprigs of feverfew boiled in it, was ready, Sally was sent up stairs
with it. She found Colin awake, but weak and ill; and, much to her
surprise, on presenting him with a lump of bread and the basin of
milk, which more closely resembled a light green wash for stencilling
walls, than any true Christian dish, he could neither touch nor bear
the sight of either.
“La!” cried Sally, “why, I never heard anything like it, as neither to
eat nor drink! Come, cram a bit down your throat with your finger,
and see if it will not get you an appetite. Why, I can eat and drink
very well, and why shouldn't you? Come, come, don't be soft, and
refuse what Gor-amighty sends you, while it lies in your power to get
it. I'm sure this milk is very nice, indeed.”
In corroboration of her statement she took a sip. But Colin shook
his head feebly and heavily, and declared it would do him no good.
He could take nothing,—he wanted nothing, but to be left alone,
that he might think and wish, and weep as he thought and wished
that he were but once more at home, or that his mother or Fanny
were but with him.
Shortly after Sally had returned below stairs, and communicated
the astounding intelligence that Colin would take neither bit nor sup,
Miss Sowersoft herself crept up stairs. She assured him he had
plenty of colour in his face; that there could not be anything
particularly amiss with him; advised him against putting on
pretences of sickness, lest he should be struck with sickness in
reality as a judgment on him, like the children that mocked the
prophet Elijah, and were eaten up by bears; and concluded by
insinuating, that if he were tickled with a whip-thong, he would in all
probability be a great deal better directly.
“Send me home!” bitterly ejaculated Colin, bursting into tears.
“Put me in a cart, and send me home!—I want to go home!—I must
go home!—Mother'!—Fanny!—Oh, come to me!—I shall die—I shall
die!”
Miss Sowersoft felt rather alarmed; but reflecting that there was
nothing like showing a little spirit and resolution when young folks
took such whims as those into their heads, she severely taunted him
with being home-sick and mother-sick; told him that neither she nor
Fanny, if they were present, could do more for him than she could;
and threatened that, if he did not leave off that hideous noise, which
was disgraceful to a great lad of his age, she would tie a stocking
round his mouth, and stop him that way. There being no great
consolation in all this, it is not surprising that our hero made such
slight application of it, that, for the matter of any difference it made
in him, Miss Sowersoft might just as well have tied her stocking
across her own mouth, or stuffed it in, which ever she might prefer,
as have given utterance to it. She was therefore constrained to
submit to the lad's own way, and to confess in her own mind that
there really was something more amiss with him than at first she
had believed.
By mid-day he had become a great deal worse; and in the
afternoon, as his disorder still rapidly increased, Mr. Palethorpe was
despatched on horseback to Bramleigh, for the purpose of consulting
Dr. Rowel.
About six o'clock in the evening he returned home, bringing with
him a packet of white powders in little blue papers, tied together
much in the fashion of that little pyrotechnic engine of mischief
usually denominated a cracker.
Certain fears which had by this time crept over the mind of Miss
Sowersoft caused her to be more than usually charitable and eager
in her inquiries after the doctor's opinion about Colin: but the
answers she received were neither very conclusive nor very
satisfactory. She was, in fact, obliged to seek for consolation, for the
present, in the belief, which she struggled hard to impress firmly
upon herself, that the boy's illness had arisen wholly in consequence
of his sitting on the ground so late in the evening to write his letter;
and that his subsequent sousing in the horse-trough had no
connexion whatever with it; as he might very easily have fallen
accidentally into a river instead, and received no more harm from it
than he had from the aforesaid pumping.
Daring several subsequent days the boy continued in such a state
as filled his mistress with continual apprehensions lest her house
should eventually be troubled with his corpse. About his death,
considering that event solely by itself, she cared very little; he might
live or die, just as his constitution inclined him, for aught she would
choose between the two; only, in case he should not survive, it
would annoy her very much indeed to have all the trouble of getting
another body's corpse prepared for the ground, without in all
likelihood ever receiving from Mrs. Clink a single halfpenny in return
for it. She mentioned her apprehensions to Mr. Palethorpe, who
replied that it was all silly childishness to allow herself to be imposed
on by her own good feelings, and that to talk about humanity would
never do for folks so far north as they were. On this unquestioned
authority Miss Sowersoft would inevitably have acted that very day,
and removed our hero, at any risk, to Bramleigh, in order to give him
a chance of dying comfortably at home, had not fortune so ordered
it, that, while preparations were being made for taking him from a
bed of fever into an open cart which stood ready in the yard, Dr.
Rowel chanced to ride up, and at once put his veto upon their
proceedings. Not that the doctor would by any means have
purposely ridden half the distance for the sake of such a patient; but
as chance not unfrequently favours those whom their own species
despise, it happened that his professional assistance had that
afternoon been required in the case of a wealthy old lady in the
neighbourhood; and, as the doctor's humanity was not, at all events,
so very short-legged as not to be able to carry him one quarter of a
mile when it lay in his way, he took Snitterton Lodge in his circuit,
for the sake of seeing Master Colin.
It will readily be supposed that during these few days, (as the boy
had not made his appearance at home on the previous Sunday,
according to conditional promise,) both his mother and Fanny had
almost hourly been expecting to hear from him. Nor had various
discussions on the cause of his silence been by any means omitted.
Mrs. Clink attributed it to the fact of his having found everything so
very pleasant at Snitterton Lodge, that he really had had neither
time nor inclination to wean himself for a few short hours from the
delights with which he was surrounded; but Fanny, whose mind had
been dwelling ever since his departure upon the dismal forebodings
with which Miss Sowersoft's appearance had filled it, expressed to
Mrs. Clink her full belief that something had happened to Colin, or
he would never have neglected either to come himself, or to write,
as he had promised.
“I am sure,” she continued, very pensively, “it has made me so
uneasy all this last week, that I have dreamed about him almost
every night. Something has happened to him, I am as certain as if I
had seen it; for I can trust to Colin's word just as well as though he
had taken his oath about it. However, I will walk over this afternoon
and see; for I shall never rest until I know for a certainty.”
“Walk, fiddlesticks!” exclaimed Mrs. Clink. “If you go over there in
that suspicious manner, as though you fancied they had murdered
him, it is a hundred to one but you will affront Miss Sowersoft, and
get Colin turned out of a situation that may be the making of him.
Stay where you are—do; and if you cannot make anything, do not
mar it by interfering in a matter that you know nothing about. I have
had trouble enough with him one way or another, without his being
brought back on my hands, when he is as comfortable, I dare say,
as he possibly can be.”
Though the latter remark was evidently intended to apply to
Fanny's supposed injudicious solicitude for Colin's welfare, the girl
passed it by without observation. She hurried her day's work
forwards, in order to gain the necessary time for making her
projected visit; and at about the middle of the afternoon suddenly
disappeared from the eyes of Mrs. Clink, without informing her
previously touching her place of destination.
While Dr. Rowel was yet in attendance on Colin, Fanny arrived and
introduced herself to Miss Sowersoft, as she was employing herself
in picking the pips off a handful of cowslips which lay in her lap. On
seeing Fanny thus unexpectedly, and under circumstances which she
felt would require some very ingenious explanation or evasion, her
countenance seemed to darken as though a positive shadow had
been cast upon it. A struggle between her real feelings and her
consciousness of the necessity to disguise them ensued; and in the
course of a few brief seconds the darkness of her countenance
passed away, and she affected to salute her unwelcome visitor with
much cordiality.
In reply to Fanny's inquiry respecting Colin, Miss Sowersoft stated
that he was improving very nicely under Mr. Palethorpe's tuition,
although they had had some trouble to make him do as he was bid;
that he had enjoyed the most extraordinary good health until a few
days ago, when he took a little cold, which had made him rather
poorly.
“There!—I was sure of it!” cried Fanny, interrupting her; “I said so
to his mother before I came away. I knew there was something
amiss, or he would have written to us before now. And how did he
take such a cold, Miss Sowersoft?”
“Take cold!—why, you know there are a hundred different ways of
taking cold, and it is impossible sometimes for even a person himself
to say how he took it. I am sure Palethorpe gets tremendous colds
sometimes, and how he gets them is a perfect miracle. But, on my
word, cold is so insinuating, that really, as I say sometimes, there is
not a part but it will find its way to at one time or another.”
“Yes—but where is Colin now?—because I shall want to see him
before I go back.”
“Oh, he is somewhere about the house,” replied Miss Sowersoft,
with an unprecedented degree of effrontery; “but your seeing him is
not of the least consequence. It cannot cure his cold; and as for
anything else, it would very likely make him all the more
discontented when you were gone again. If you take my advice, you
would not see him, especially when I can tell you everything just the
same as though you saw it yourself.”
At this moment the foot of the doctor, as he groped his way down
stairs, was overheard by the speaker. She started up instantly, and
endeavoured to hurry Fanny out of the room before that professional
gentleman should enter it; but her manoeuvre failed, and before
Miss Sowersoft could caution him to be silent the doctor remarked,
in a sufficiently loud tone to be heard distinctly by both, that unless
the boy was taken great care of, there was little chance left of his
recovery.
“What boy?” exclaimed Fanny, rushing forward. “What is he so ill
as that? For God's sake let me see him!”
Concluding from the direction in which the doctor had come that
Colin was somewhere in the regions above, she flew rather than
walked up stairs, without waiting for an invitation or a conductor,
and soon threw her arms in an ecstasy of grief upon his neck.
“Oh, Colin! God has sent me on purpose to save you! Do be
better, and you shall go home again very soon.”
But Colin could only put up his pallid arms in an imploring action,
and cry for very joy, as he gazed in the face of one of those only two
who had occupied his das and night thoughts, and been the
unconscious subjects of his unceasing and most anxious wishes.
The trouble of this first meeting being over, some more quiet
conversation ensued; and, although almost too ill and weak to be
allowed to talk, Colin persisted in stating briefly to the horror-
stricken Fanny the kind of reception he had met with on his arrival,
his treatment afterwards, the taking of his letter from him, and the
brutal conduct which had caused his present illness. The girl stood
silent, merely because she knew not what to think, what to believe,
what to doubt; and was besides utterly lost for words to express
properly her strangely mingled thoughts. It was almost impossible—
incredible! Why could they do it? There was no cause for it—there
could be no cause for it. Human nature, and especially human
nature in the shape of woman, was incapable of anything so
infamous. Yet Colin was sensible—he had told an intelligible tale;
and, most true of all, there he lay, a mere vision of what he was so
brief a time ago,—a warranty plain and palpable that grievous wrong
had been endured. Her brain was absolutely bewildered—she looked
like one hovering on the doubtful boundary between sense and
insanity. She cast her eyes around for surety—on the bed—at him, A
burst of tears, as of a spring that for the first time breaks its bounds,
succeeded,—and then another and another, as she fell on her knees
and buried her face in the clothes that covered him.
By and by, the doctor and Miss Sowersoft were present in the
room with her. Fanny raised her head and beheld Colin's mistress
attempting, in the presence of the doctor, to do the attentive, by
adjusting the sheet about the boy's neck to keep off the external air.
“Do not touch him!” exclaimed Fanny, springing to her feet; “he
shall have nothing from your hands!”
“Ay!” cried the doctor: “young woman, what now, what now?”
“What now? Sir, you may well say what now! I have heard all
about it—Colin has told me all. Miss Sowersoft has nearly killed him,
and now wants to show, because you are here, how kind and good
she is!”
So saying, Fanny resolutely set about making the arrangement
which Miss Sowersoft had contemplated with her own hands.
“Why—what—who is this young woman?” asked the doctor,
somewhat astonished at the unexpected scene which had just
passed before him.
“Nobody!” replied Miss Sowersoft; “she is only Mrs. Clink's servant,
and a pert impudent hussy, too, as you have heard.”
At the same time she looked in the doctor's face, and
endeavoured to smile contemptuously, though it “came off” in such
a manner as would inevitably have frightened anybody less
accustomed than was Dr. Rowel to witness the agonies of the human
countenance.
“Yes, sir,” added Fanny, “I am only a servant; but I am a woman,
whether servant or mistress. I nursed this lad when I was but six
years old myself, and have taken care of him ever since. She shall
not drown him, though she thinks she will!”
“Me drown him!” exclaimed Miss Sowersoft in feigned amazement.
“Yes,” replied Fanny, “you drown him. If you had not half
murdered him in that trough, he would never have been here now.”
 “Do let us go down stairs, doctor,” observed Miss Sowersoft; “such
rubbish as this is not worth hearing.” And she made her way towards
the door.
“Where is that letter?” cried Fanny eagerly, fearful lest the lady to
whom she addressed herself should escape.
“Pshaw! nonsense! don't catechise me!” replied Miss Sowersoft, as
she tripped down stairs; while the doctor, half in soliloquy and half
addressing Miss Sowersoft, remarked, in allusion to Fanny, “She's a
damsel of some spirit too!” Then addressing the girl herself, “Are you
the little girl I saw at Mrs. Clink's when this boy was born?”
“Yes, sir, I am,” answered Fanny, as her passion sunk almost to
nothing, and she blushed to be so questioned.
“Ah, indeed!” cried Doctor Rowel. “Well, I should not have thought
it. Why, you are quite a fine young woman now. Dear-a-me! I had
quite lost sight of you. I could not have believed it. Humph!” And the
doctor surveyed her fair proportions with something of
astonishment, and a great deal of satisfaction. To think that from
such a little pale, half-fed, unhappy thing of work and thought
beyond her years as she then was, there should have sprung up the
full-sized, the pretty featured, and naturally genteel-looking girl now
before him! But then, he had not that benefit which the reader
enjoys, of reflecting how worldly circumstances, how poverty and
plenty, sway the tempers of mankind; and that, as Mistress Clink's
circumstances improved, so had Fanny improved likewise; and from
seven or eight years old upwards, Fanny had enjoyed a much more
comfortable home than, on his first introduction to her, might
reasonably have been expected.
Doctor Rowel resumed his conversation.
“And how came you to be put to service so very early? for you had
not, if I remember rightly, either health or strength to recommend
you.”
Colin's eyes as he lay were fixed, as it might have been the eyes
of a picture, on the doctor's countenance.
 “I don't know, I'm sure, sir,” replied Fanny: but after a few
moments' hesitation, added, “I suppose it was because I had no
friends.”
“No friends!” the doctor repeated,—“why, where's your father and
mother?”
“I never knew them, sir.”
“Indeed! never knew them!”
“No, sir!” and Fanny sobbed at the very recollection of her
childhood's helplessness.
“Humph!” ejaculated the doctor; “you scarcely seem to have been
born for a servant. Where did Mrs. Clink find you?”
“I do not know, sir. She never told me.”
“Ah!—oh! oh!—well! It's odd she never told you. So you do not
know either who your father, or your mother, or your friends were?”
“No, sir,—I do not. But I remember———”
“Well,—go on,—you remember,—what do you remember? where
did you come from? Do you know that?”
“I think, from Leeds, sir.”
“Leeds!” exclaimed the doctor; “and what else do you remember?”
“I can remember, sir,—though I can but just remember it,—that
my father was taken away from me once, and I never saw him
again.”
“And, what's your name?” continued the doctor in evident
excitement.
“Fanny Woodruff,” she replied.
The doctor's features became pale and rigid, and his eyes were
fixed upon her almost immoveably.
“God bless my soul!” he slowly ejaculated, as he rose to leave the
room; “she should have been lost, or dead!”
But he turned again when at the head of the stairs.
“Now, young woman,—if you can keep a secret,—tell nobody, not
even your mistress, what has passed. Take no notice; and perhaps I
may do something for you. But I thought we had seen the last of
your face seventeen years ago!”
Fanny and Colin were left alone.
“He knows something about me!” was the first thought that arose
in Fanny's mind. But she did not utter it, and only asked very softly,
if Colin had heard what the doctor said.
“Yes,” he replied, “and I shall never forget it.”
“But, say nothing,” added the girl: “he promised to do something
for me. I wonder what it is!”
“So do I,” added Colin; “something worth having, I dare say.”
Thus they talked till evening. Colin said how much better he felt
since she had been with him; and Fanny declared she would not
leave him again for another day, until he was well; and, when he
was well, then she would get him away from such unfeeling people,
even though she had to go down on her knees to beg another
situation for him elsewhere.
When, some little time afterwards, Fanny went down stairs, and
informed the mistress of the house of her resolution to stay and
attend on Colin until he was better, that amiable creature replied, “I
think you won't then. We have not any room to spare. As if I was
going to keep beds at liberty, to accommodate any trunnion that
may think fit to cram herself into my house! We've plenty of work on
our hands without having to wait on other people's servants. What
do you say, Palethorpe?”
“Well, I don't know, meesis,” replied Mr. Palethorpe; “it seems as if
Mr. Rowel was understood to say he was very bad, and must be
waited on pretty constantly.”
“I'm sure I sha'n't wait on him neither constantly nor
inconstantly!” very pertly exclaimed Miss Sowersoft; and certainly
giving a very ingenious turn to her own views, as soon as she found
which way her lover's needle pointed; “I'm not going to trot up and
down stairs a thousand times a day for the sake of such a thing as a
plough-lad. Them may wait on him that likes him, if he is to be
waited on; but I'm positive I shan't, nor anybody else that belongs
to me!”
This conclusion left, without another word, the field wholly open
to Fanny; and as Miss Sowersoft, on concluding her speech, bounced
off into the dairy, not another word was needed.
Whatever might be the views entertained by the lady of the house
touching the treatment most proper for Colin, there still were
individuals amongst that rude community whose feelings were of a
somewhat more catholic kind than those of their mistress; so that
Fanny found no difficulty in procuring a volunteer, in the person of
Abel, to go over to Bramleigh for the purpose of informing Mistress
Clink how affairs stood, and of bringing back such few needful
articles as Fanny might require during her stay at the farm.
All that night she passed a sleepless watch by the side of Colin's
bed, beguiling the hours not devoted to immediate attendance on
him, partly by looking over the little books which had come from
home in his box, but more by employing her mind in the creation of
every possible description of fanciful supposition touching her own
origin, her history, her parents, and the knowledge which the doctor
appeared to have of her earliest life. What was it?—what could it be?
and, what could he mean by enjoining her to mention nothing of all
this to any second person? In her he had unexpectedly found one
whom he had known a baby, and had believed to be dead, or lost in
the vast crowds of poverty long ago. Had she been born to better
things than surrounded her now? Had she been defrauded of her
rights? And, did the doctor bid her be silent because he might have
to employ stratagem in order to recover them again? Perhaps she
was born—nay! she knew not what she was born; nor dare she trust
herself to think, scarcely; though, certain it is that a visionary world
of ladies and gentlemen, and fine things, and wealth to set Colin up
in the world and to make his mother comfortable, and to exalt
herself over all the petty enemies by whom they were now
surrounded, passed in pleasant state before her prolific imagination:
while, it is equally certain, that—blushing, though unseen and in
secret, at the very consciousness—a prouder feeling sprung up in
her bosom, and she began to feel as though she must be more
genteel, and more particular, and less like a common servant, than
she had hitherto been.
Such were the golden fancies, and the pretty resolves that
crowded round her brain that night. Neither, as a honest chronicler
of human nature, would I take upon me to assert that she did not
once or twice during these reveries rise to contemplate her features
in the glass, and to adjust her hair more fancifully, and wonder—if it
should be so—what kind of looking lady she should make. Truly, it
was a pretty face that met her eyes in the mirror. As Colin woke up
from a partial slumber, and raised his head slightly from the pillow,
to ascertain what had become of his guardian, the reflection of her
countenance as she was “looking the lady,” chanced to catch his eye:
and, though he smiled as he gently sunk down again, he thought
that that face would never again pass from before him.
 CHAPTER XV.
Fanny is deceived by the doctor.—A scene in Rowel's
“Establishment for the Insane” at Nabbfield.

P
OOR girl! What pains she takes—if not to “curse herself,” at
least to form that paradise out of the chaos of her own
thoughts, which her supposed benefactor, the physician, never
intended to realize. She was deceived, utterly and deeply
deceived; and deceived, too, by the very means which the doctor
had recommended to her apparently for the attainment of success.
For, great as some of our modern diplomatists have incontestably
been considered in their noble and polite art, I much question
whether the man more capable of aspiring to higher honours in it
than Doctor Rowel of Nabbfield, is not yet to be born.
As the doctor rode homewards, after his interview with Fanny, he
several times over, and with inexpressible inward satisfaction,
congratulated and complimented himself upon having achieved such
a really fine stroke of policy at a very critical moment, as no other
man living could, he verily believed, have at all equalled. Within the
space of a few brief moments he had, to his infinite astonishment,
discovered, in the person of a serving girl, one whom he himself had
endeavoured, while she was yet an infant, to put out of the way;
and upon whose father he had perpetrated one of the most
atrocious of social crimes, for the sole purpose of obtaining the
management of his property while he lived, and its absolute
possession on his decease. He had ascertained that the girl retained
some indistinct recollection of the forcible arrest and carrying away
of her parent, of which he himself had been the instigator; and thus
suddenly he found himself placed in a position which demanded
both promptitude and ingenuity in order to secure his own safety
and the permanency of all he held through this unjust tenure. Since
any discovery by Fanny of what had passed between them would
inevitably excite public question and inquiry, the very brilliant idea
had instantaneously suggested itself to his mind that—as in-the girl's
continued silence alone lay his own hopes of security—no project
could be conceived more likely to prove successful in obtaining and
preserving that silence, than that of representing it as vital to her
own dearest interest to keep the subject deeply locked for the
present in her own bosom. This object, he flattered himself, he had
already succeeded in achieving, without exciting in the mind of
Fanny herself the least suspicion of his real and ultimate purpose. At
the same time he inwardly resolved not to stop here, but to resort to
every means in his power calculated still more deeply to bind the
unsuspecting young woman to the preservation of that silence upon
the subject, which, if once broken, might lead to the utter overthrow
of a system which he had now maintained for many years.
Elated with the idea of his own uncommon cleverness, he
cantered along the York road from the moor with corresponding
briskness; turned down a green lane to the left, cleared several
fences and a pair of gates in his progress, and reached within sight
of his “Establishment for the Insane” at Nabbfield, as the last light of
another unwished-for and unwelcome sun shot through the barred
and grated windows of the house, and served dimly to show to the
melancholy habitants of those cells the extent of their deprivations
and their misery.
Far advanced as it was in the evening, the doctor had not yet
dined; his professional duties, together with some other causes
already explained, having detained him beyond his usual hour.
Nevertheless, for reasons best known to himself, but which, it may
be supposed, the events of the afternoon had operated in producing,
the doctor had no sooner dismounted, and resigned his steed to the
care of a groom, who appeared in waiting the instant that the clatter
of his hoofs sounded on the stones of the yard, than, instead of
retiring to that removed portion of the building, in which, for the
purpose of being beyond reach of the cries of those who were kept
in confinement, his own private apartments were situated, he
demanded of one of the keepers the key of a particular cell. Having
obtained it,—
“Shall I attend you, sir?” asked the man.
“No, Robson. James is harmless. I will see him into his cell myself
to-night.”
“He is in the patient's yard, sir,” replied the keeper.
“Very well—very well. Wait outside; and, if I want assistance, I will
call you.”
The man retired, while Doctor Rowel proceeded down a long and
ill-lighted passage, or corridor, in which were several angular turns
and windings; and when nearly lost in the gloom of the place, he
might have been heard to draw back a heavy bolt, and raise a
spring-latch like an iron bar, which made fast the door that opened
upon the yard, or piece of ground to which the keeper had alluded.
It was just at that brief but peculiar time at the turn of day and
night, which every observer of Nature must occasionally have
remarked, when the light of the western atmosphere, and that of a
rayless moon high up the southern heaven, mingle together in
subdued harmony, and produce a kind of illumination, issuing from
no given spot, but pervading equally the whole atmosphere,—like
that which we might imagine of a fairy's palace,—without any
particular source, neither wholly of heaven nor of earth, but
partaking partially of each.
The passage-door was thrown back, and the doctor stood upon its
threshold. A yard some forty feet square, surrounded by a wall about
six yards high, and floored with rolled gravel, like the path of a
garden, was before him. Near the centre stood a dismal-looking
yewtree, its trunk rugged, and indented with deep natural furrows,
as though four or five shoots had sprung up together, and at last
become matted into one; its black lines of foliage, harmonizing in
form with the long horizontal clouds of the north-west quarter, which
now marked the close approach of night. Nothing else was to be
seen. As the eye, however, became somewhat more accustomed to
the peculiar dusky light which pervaded this place, the figure of a
man standing against the tree-trunk became visible; with his arms
tightly crossed upon his breast, and bound behind him as though
they had almost grown into his sides; and his hair hanging long
upon his shoulders, somewhat like that of a cavalier, or royalist, of
the middle of the seventeenth century.
The doctor raised his voice, and called, in a lusty tone,
“Woodruff!”
The patient returned no answer, nor did he move.
“James Woodruff!” again shouted the doctor.
A slight turn of the head, which as quickly resumed its previous
attitude, was the only response made to the doctor's summons.
Finding that he could not call this strange individual to him, Doctor
Rowel stepped across the yard, and advanced up to him.
“James,” said he mildly, “it is time you were in your cell.”
The man looked sternly in his face, and replied, “I have been
there some thousands of times too often already.”
“Never heed that,” answered Rowel. “You must go to rest, you
know.”
“Must go—ay? Ah! and so I must. I am helpless. But, had I one
hand free—only one hand—nay, with one finger and thumb, I would
first put you to rest where you should never wake again! When am I
to go free?”
“Will you go to your room?” said the doctor, without regarding his
question.
“I ask again,” cried the alleged madman, “as I have asked every
day past counting, when am I to be loosed of this accursed place?
How long is this to last?”
“Only until you are better,” remarked, with deep dissimulation, this
worthy member of the faculty.
“Better!” exclaimed Woodruff, with rising passion, as he tugged to
loosen his arms from the jacket which bound him, though as
ineffectually as a child might have tugged at the roots of an oak
sapling. “I could curse you again and doubly for that word, but that I
have cursed till language is weak as water, and words have no more
meaning. I am sick of railing. Better! Till I am better! Thief!—liar!—
villain!—for you are all these, and a thousand more,—I am well. You
know it. Sound in mind and body,—only that these girths have
crippled me before my time. How am I mad? I can think, reason,
talk, argue,—hold memory of past life. I remember, villain! when you
and your assassins seized me; stole my child from me; swore that I
was mad; and brought me here, now seventeen years ago; and all in
order that you might rob me of my property!—I remember that. Is
that madness? I remember, before that, that I married your sister.
Was it not so? I remember that she died, and left me a little pattern
of herself, that called you uncle. Was not that so? Where is that
child? What has become of her? Or are you a murderer besides? All
this I remember: and I know now that I have power of will, and
aptness to do all that man's mind is called to do. How, then, am I
mad? Oh! for one hand free! One hand and arm. Only one! Give me
that half chance to struggle with you. Let us end it so, if I am never
to go free again. Take two to one; and if you kill me, you shall stand
free of the scaffold; for I will swear with my last breath that you did
it in self-defence. Do that. Let me have one grapple—a single gripe—
and, if you can master me, why God forgive you!”
The doctor smiled, as in contempt of the impotent ravings and
wild propositions of his brother-in-law; for such, it is almost needless
to state, James Woodruff was. But the alleged maniac continued his
discourse.
“Then, as you are such a rank, arrant coward, give me my whole
liberty; let me go beyond this house, and I will never touch you. I
will not ruffle a hair of your accursed head. Do that, and I will leave
you to God for the reward of all you have done to me and mine. Set
me free! Untie my limbs, and let me out this night! It is dark.
Nobody can tell where I came from. Let me go, and I will never
mention your name in complaint, nor lift a hand against you. Think,
man,—do but think! To spend seventeen years of nights in that
dungeon, and seventeen years of days on this speck of ground! To
you who have been at liberty to walk, and breathe freely, and see
God's creation, it may be idle; but I have seen nothing of seventeen
springs but their light skies; nor of summers, but their heat and their
strong shadows; nor of autumn, but the random leaves which the
wind whirled over into this yard; nor of winter, but its snow and
clouds. I want to be upon the green earth,—the grass,—amongst the
fields. I want to see my wife's grave again!—some other human face
than yours I—and—and—Man,—if you be man,—I want to find my
daughter!”
He flung himself on the ground, and groaned as in utter despair.
The doctor was accustomed to witness these fits of frenzy, and
therefore paid no farther attention now than consisted in an effort to
raise the man again upon his feet, and a renewed solicitation to him
to retire into his room.
“No,” said he; “I have something to speak of yet. I have come to
another determination. In my mind, villain! there has been
seventeen years of rebellion against your wrong; and I have sworn,
and have kept my oath till now, that you should never compel me to
give up my rights, in virtue of my wife, to you. But time has outworn
the iron of my soul: and seventeen years of this endurance cannot
be set against all the wealth of the world. What is it to me? To dig
the earth, and live on roots; but to be free with it; to go and come
as I list; to be at liberty, body and limb! This would be paradise
compared with the best palace that ever Mammon built in hell. Now,
take these straps from off me, and set me free. Time is favourable.
Take me into your house peaceably and quietly, and I will make over
to you all I have, as a free gift. What you have stolen, you shall
keep. Land, houses, gold, everything; I will not retain of them a
grain of sand, a stone, or a sparkle of metal. But let me out! Let me
see this prison behind me!”
“It would be the act of a lunatic, and of no effect,” replied the
doctor.
“How lunatic? To give that which is of no use to me for that which
is dearer than life? Besides, I am sane—sound of mind.”
 “No,” interrupted the doctor, “you are wrong on one question. Your
disease consists in this very thing. You fancy I keep you confined in
order to hold your property myself.”
“Fancy you do!” savagely exclaimed Woodruff, stamping the
ground with rage; “this contradiction is enough to drive me mad. I
know it! You know it. There is no fancy in the case. It is an excuse, a
vile pretence, a lie of seventeen years' standing. It was a lie at first.
Will you set me free?”
“It cannot be,” said the doctor; “go to your room.”
“It shall be!” replied Woodruff; “I will not go.”
“Then I must call assistance,” observed Rowel, as he attempted to
approach the door at which he had entered.
“You shall not!” replied the patient, placing himself in front of the
doctor, as though resolutely bent on preventing his approach to the
door, although he had not the least use of his arms, which might
have enabled him to effect his purpose.
“Stand aside, fool!” Rowel exclaimed, as he threw out his right
arm in order to strike off the intruder. But Woodruff anticipated him;
and, by a sudden and dexterous thrust of his foot in a horizontal
line, knocked the doctor's legs from under him, and set him
sprawling on the ground. Woodruff fell upon him instantly, in order
to keep him down, and to stifle the loud cries of “Robson! Robson!”
which were now issuing in rapid succession from the doctor's larynx.
At the same time a tremendous struggle, rendered still more
desperate by the doctor's fears, took place on the ground; during
which the unhappy Woodruff strove so violently to disengage his
hands from the ligatures of the waistcoat which bound him, that the
blood gushed copiously from his mouth and nostrils. His efforts were
not altogether unavailing. He partly disengaged one hand; and, with
a degree of activity and energy only to be accounted for from the
almost superhuman spirit which burned within him, and for which
his antagonist, with all his advantages, was by no means an equal
match, he succeeded in planting his forefinger and thumb, like the
bite of a crocodile upon the doctor's throat.
 Original Size

“Swear to let me free, or I 'll kill you!” he exclaimed.

“Yes,—y—e—s,—I sw—ear!” gurgled through the windpipe of Dr.
Rowel as he kicked and plunged like a horse in a bog to shake off his
foe. The light of a lamp flashed upon them, and Robson rushed into
the yard.
“Let me out!” again demanded Woodruff.
“I will; I will!” replied the doctor.
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com