Download the full version of the ebook at ebookname.

com

Signal Processing for Intell Sensor Systems with

MATLAB 2nd ed. Edition D. Swanson

https://ebookname.com/product/signal-processing-for-intell-
sensor-systems-with-matlab-2nd-ed-edition-d-swanson/

OR CLICK BUTTON

DOWNLOAD EBOOK

Download more ebook instantly today at https://ebookname.com

Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Discrete systems and digital signal processing with MATLAB

Second Edition Elali

https://ebookname.com/product/discrete-systems-and-digital-signal-
processing-with-matlab-second-edition-elali/

ebookname.com

Discrete Random Signal Processing and Filtering Primer

with MATLAB 1st Edition Alexander D. Poularikas

https://ebookname.com/product/discrete-random-signal-processing-and-
filtering-primer-with-matlab-1st-edition-alexander-d-poularikas/

ebookname.com

The Digital Signal Processing Handbook Wireless Networking

Radar Sensor Array Processing and Nonlinear Signal
Processing 2nd Edition Vijay K. Madisetti
https://ebookname.com/product/the-digital-signal-processing-handbook-
wireless-networking-radar-sensor-array-processing-and-nonlinear-
signal-processing-2nd-edition-vijay-k-madisetti/
ebookname.com

Industrial Safety and Health Management 6th Edition C. Ray

Asfahl

https://ebookname.com/product/industrial-safety-and-health-
management-6th-edition-c-ray-asfahl/

ebookname.com
The Academic Job Search Handbook Fifth Edition Julia
Miller Vick

https://ebookname.com/product/the-academic-job-search-handbook-fifth-
edition-julia-miller-vick/

ebookname.com

Engineering Mechanics Statics 10th Edition SOLUTION MANUAL

Russell C. Hibbeler

https://ebookname.com/product/engineering-mechanics-statics-10th-
edition-solution-manual-russell-c-hibbeler/

ebookname.com

The Colour of Betrayal Sebastian Foxley Medieval Mystery 4

1st Edition Mount

https://ebookname.com/product/the-colour-of-betrayal-sebastian-foxley-
medieval-mystery-4-1st-edition-mount/

ebookname.com

The Dream Hunter Dark Hunter 10 Hunter Legends 13 First

Edition Sherrilyn Kenyon

https://ebookname.com/product/the-dream-hunter-dark-hunter-10-hunter-
legends-13-first-edition-sherrilyn-kenyon/

ebookname.com

Basic electronics and linear circuits 2nd Edition N. N.

Bhargava

https://ebookname.com/product/basic-electronics-and-linear-
circuits-2nd-edition-n-n-bhargava/

ebookname.com
The Advertising Research Handbook 2nd Edition Revised and
Expanded Charles E. Young

https://ebookname.com/product/the-advertising-research-handbook-2nd-
edition-revised-and-expanded-charles-e-young/

ebookname.com
Signal
Processing
Processi g foro
Intelligent
Sensor Systems
ms
with MATLAB
T AB ®

Second Edition

David C. Swanson
www.itpub.net
®
www.itpub.net
Signal
Processing for
Intelligent
Sensor Systems
with MATLAB ®

Second Edition

David C. Swanson

Boca Raton London New York

CRC Press is an imprint of the

Taylor & Francis Group, an informa business
MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the
accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB® software or related products
does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular
use of the MATLAB® software.

CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2012 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group, an Informa business

No claim to original U.S. Government works

Version Date: 20110520

International Standard Book Number-13: 978-1-4398-7950-4 (eBook - PDF)

This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been
made to publish reliable data and information, but the author and publisher cannot assume responsibility for the valid-
ity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright
holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this
form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may
rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or uti-
lized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopy-
ing, microfilming, and recording, or in any information storage or retrieval system, without written permission from the
publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://
www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923,
978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.
Visit the Taylor & Francis Web site at
http://www.taylorandfrancis.com
and the CRC Press Web site at
http://www.crcpress.com

www.itpub.net
This book is dedicated to all who aspire to deeply understand signal processing
for sensors, not just enough to pass an exam or assignment, or to complete a
project, but deep enough to experience the joy of natural revelation. This takes
more than just effort. You have to love the journey. This was best said by one
of America’s greatest inventors, George Washington Carver, in the quote
“Anything will give up its secrets if you love it enough…”
www.itpub.net
Contents
Preface������������������������������������������������������������������������������������������������������������������������������������������� xiii
Acknowledgments���������������������������������������������������������������������������������������������������������������������������� xv
Author��������������������������������������������������������������������������������������������������������������������������������������������xvii

Part I Fundamentals of Digital Signal Processing

Chapter 1 Sampled Data Systems.................................................................................................. 3

1.1 A/D Conversion..................................................................................................3
1.2 Sampling Theory................................................................................................6
1.3 Complex Bandpass Sampling............................................................................. 9
1.4 Delta–Sigma Analog Conversion..................................................................... 12
1.5 MATLAB® Examples....................................................................................... 14
1.6 Summary, Problems, and References............................................................... 15
Problems...................................................................................................................... 16
References................................................................................................................... 17

Chapter 2 z-Transform.................................................................................................................. 19
2.1 Comparison of Laplace and z-Transforms........................................................ 19
2.2 System Theory.................................................................................................. 27
2.3 Mapping of s-Plane Systems to the Digital Domain........................................ 30
2.4 MATLAB® Examples....................................................................................... 39
2.5 Summary..........................................................................................................40
Problems...................................................................................................................... 41
References................................................................................................................... 41

Chapter 3 Digital Filtering........................................................................................................... 43

3.1 FIR Digital Filter Design................................................................................. 43
3.2 IIR Filter Design and Stability......................................................................... 47
3.3 Whitening Filters, Invertibility, and Minimum Phase..................................... 49
3.4 Filter Basis Polynomials................................................................................... 52
3.4.1 Butterworth Filters.............................................................................. 52
3.4.2 Chebyshev Type I Filters..................................................................... 55
3.4.3 Chebyshev Type II Filters................................................................... 56
3.4.4 Elliptical Filters................................................................................... 58
3.4.5 Bessel Filters....................................................................................... 59
3.4.6 High-Pass, Band-Pass, and Band-Stop Filter Transformations........... 59
3.4.7 MA Digital Integration Filter.............................................................. 59
3.5 MATLAB® Examples.......................................................................................60
3.6 Summary.......................................................................................................... 62
Problems...................................................................................................................... 63
References................................................................................................................... 63

ix
x Contents

Chapter 4 Digital Audio Processing............................................................................................ 65

4.1 Basic Room Acoustics...................................................................................... 65
4.2 Artificial Reverberation and Echo Generators................................................. 69
4.3 Flanging and Chorus Effects............................................................................ 72
4.4 Bass, Treble, and Parametric Filters................................................................. 74
4.5 Amplifier and Compression/Expansion Processors......................................... 76
4.6 Digital-to-Analog Reconstruction Filters.........................................................80
4.7 Audio File Compression Techniques................................................................ 82
4.8 MATLAB® Examples....................................................................................... 88
4.9 Summary.......................................................................................................... 91
Problems......................................................................................................................92
References...................................................................................................................92

Chapter 5 Linear Filter Applications........................................................................................... 95

5.1 State Variable Theory....................................................................................... 95
5.1.1 Continuous State Variable Formulation..............................................97
5.1.2 Discrete State Variable Formulation...................................................99
5.2 Fixed-Gain Tracking Filters........................................................................... 101
5.3 2D FIR Filters................................................................................................. 107
5.4 Image Upsampling Reconstruction Filters..................................................... 115
5.5 MATLAB® Examples..................................................................................... 117
5.6 Summary........................................................................................................ 119
Problems.................................................................................................................... 120
References................................................................................................................. 121

Part II Frequency Domain Processing

Chapter 6 Fourier Transform..................................................................................................... 127

6.1 Mathematical Basis for the Fourier Transform.............................................. 127
6.2 Spectral Resolution......................................................................................... 130
6.3 Fast Fourier Transform................................................................................... 135
6.4 Data Windowing............................................................................................. 138
6.5 Circular Convolution Issues........................................................................... 143
6.6 Uneven-Sampled Fourier Transforms............................................................ 146
6.7 Wavelet and Chirplet Transforms................................................................... 153
6.8 MATLAB® Examples..................................................................................... 162
6.9 Summary........................................................................................................ 165
Problems.................................................................................................................... 167
References................................................................................................................. 168

Chapter 7 Spectral Density........................................................................................................ 169

7.1 Spectral Density Derivation........................................................................... 169
7.2 Statistical Metrics of Spectral Bins................................................................ 172
7.2.1 Probability Distributions and PDFs.................................................. 173
7.2.2 Statistics of the NPSD Bin................................................................ 175
7.2.3 SNR Enhancement and the Zoom FFT............................................. 176

www.itpub.net
Contents xi

7.2.4 Conversion of Random Variables...................................................... 177

7.2.5 Confidence Intervals for Averaged NPSD Bins................................ 179
7.2.6 Synchronous Time Averaging........................................................... 180
7.2.7 Higher-Order Moments..................................................................... 181
7.2.8 Characteristic Function..................................................................... 182
7.2.9 Cumulants and Polyspectra............................................................... 182
7.3 Transfer Functions and Spectral Coherence................................................... 188
7.4 Intensity Field Theory.................................................................................... 199
7.4.1 Point Sources and Plane Waves.........................................................200
7.4.2 Acoustic Field Theory.......................................................................200
7.4.3 Acoustic Intensity.............................................................................. 203
7.4.4 Structural Intensity............................................................................206
7.4.5 Electromagnetic Intensity..................................................................208
7.5 Intensity Display and Measurement Techniques............................................209
7.5.1 Graphical Display of the Acoustic Dipole........................................209
7.5.2 Calculation of Acoustic Intensity from Normalized
Spectral Density................................................................................ 213
7.5.3 Calculation of Structural Intensity for Compressional and
Bending Waves.................................................................................. 215
7.5.4 Calculation of the Poynting Vector................................................... 217
7.6 MATLAB® Examples..................................................................................... 218
7.7 Summary........................................................................................................ 219
Problems.................................................................................................................... 220
References................................................................................................................. 221

Chapter 8 Wavenumber Transforms.......................................................................................... 223

8.1 Spatial Transforms.......................................................................................... 223
8.2 Spatial Filtering and Beamforming................................................................ 225
8.3 Image Enhancement Techniques.................................................................... 233
8.4 JPEG and MPEG Compression Techniques...................................................240
8.5 Computer-Aided Tomography........................................................................ 243
8.6 Magnetic Resonance Imaging........................................................................ 249
8.7 MATLAB® Examples..................................................................................... 254
8.8 Summary........................................................................................................ 258
Problems.................................................................................................................... 261
References................................................................................................................. 261

Part III Adaptive System Identification and Filtering

Chapter 9 Linear Least-Squared Error Modeling...................................................................... 265

9.1 Block Least Squares....................................................................................... 265
9.2 Projection-Based Least Squares..................................................................... 269
9.3 General Basis System Identification............................................................... 271
9.3.1 Mechanics of the Human Ear............................................................ 273
9.3.2 Least-Squares Curve Fitting............................................................. 275
9.3.3 Pole–Zero Filter Models.................................................................... 276
9.4 MATLAB® Examples..................................................................................... 279
9.5 Summary........................................................................................................280
xii Contents

Problems....................................................................................................................280
References................................................................................................................. 281

Chapter 10 Recursive Least-Squares Techniques........................................................................ 283

10.1 RLS Algorithm and Matrix Inversion Lemma..............................................284
10.1.1 Matrix Inversion Lemma..................................................................284
10.1.2 Approximations to RLS.................................................................... 286
10.2 LMS Convergence Properties......................................................................... 287
10.2.1 System Modeling Using Adaptive System Identification.................. 287
10.2.2 Signal Modeling Using Adaptive
Signal-Whitening Filters................................................................... 291
10.3 Lattice and Schur Techniques........................................................................ 295
10.4 Adaptive Least-Squares Lattice Algorithm.................................................... 301
10.4.1 Wiener Lattice...................................................................................307
10.4.2 Double/Direct Weiner Lattice........................................................... 310
10.5 MATLAB® Examples..................................................................................... 312
10.6 Summary........................................................................................................ 314
Problems.................................................................................................................... 315
References................................................................................................................. 316

Chapter 11 Recursive Adaptive Filtering..................................................................................... 317

11.1 Adaptive Kalman Filtering............................................................................. 318
11.2 IIR Forms for LMS and Lattice Filters.......................................................... 332
11.3 Frequency Domain Adaptive Filters.............................................................. 347
11.4 MATLAB® Examples..................................................................................... 353
11.5 Summary........................................................................................................ 355
Problems.................................................................................................................... 357
References................................................................................................................. 357

Part IV Wavenumber Sensor Systems

Chapter 12 Signal Detection Techniques..................................................................................... 363

12.1 Rician PDF.....................................................................................................364
12.1.1 Time-Synchronous Averaging........................................................... 365
12.1.2 Envelope Detection of a Signal in Gaussian Noise........................... 367
12.2 RMS, CFAR Detection, and ROC Curves..................................................... 374
12.3 Statistical Modeling of Multipath................................................................... 381
12.3.1 Multisource Multipath....................................................................... 382
12.3.2 Coherent Multipath............................................................................ 383
12.3.3 Statistical Representation of Multipath............................................. 385
12.3.4 Random Variations in Refractive Index............................................ 388
12.4 MATLAB® Examples..................................................................................... 391
12.5 Summary........................................................................................................ 392
Problems.................................................................................................................... 394
References................................................................................................................. 394

www.itpub.net
Contents xiii

Chapter 13 Wavenumber and Bearing Estimation....................................................................... 397

13.1 Cramer–Rao Lower Bound............................................................................. 398
13.2 Bearing Estimation and Beam Steering.........................................................403
13.2.1 Bearings from Phase Array Differences...........................................403
13.2.2 Multiple Angles of Arrival................................................................407
13.2.3 Wavenumber Filters........................................................................... 410
13.3 Field Reconstruction Techniques................................................................... 418
13.4 Wave Propagation Modeling.......................................................................... 428
13.5 MATLAB® Examples..................................................................................... 436
13.6 Summary........................................................................................................ 438
Problems.................................................................................................................... 439
References................................................................................................................. 439

Chapter 14 Adaptive Beamforming and Localization................................................................. 441

14.1 Array “Null-Forming”.................................................................................... 443
14.2 Eigenvector Methods of MUSIC and MVDR................................................ 447
14.3 Coherent Multipath Resolution Techniques...................................................460
14.3.1 Maximal Length Sequences.............................................................. 462
14.4 FMCW and Synthetic Aperture Processing................................................... 472
14.5 MATLAB® Examples..................................................................................... 476
14.6 Summary........................................................................................................ 478
Problems....................................................................................................................480
References................................................................................................................. 481

Part V Signal Processing Applications

Chapter 15 Noise Reduction Techniques..................................................................................... 485

15.1 Electronic Noise............................................................................................. 485
15.2 Noise Cancellation Techniques...................................................................... 497
15.3 Active Noise Attenuation................................................................................504
15.4 MATLAB® Examples..................................................................................... 519
15.5 Summary........................................................................................................ 520
Problems.................................................................................................................... 521
References................................................................................................................. 522

Chapter 16 Sensors and Transducers........................................................................................... 523

16.1 Simple Transducer Signals............................................................................. 524
16.2 Acoustic and Vibration Sensors..................................................................... 530
16.2.1 Electromagnetic Mechanical Transducer.......................................... 530
16.2.2 Electrostatic Transducer.................................................................... 537
16.2.3 Condenser Microphone.....................................................................546
16.2.4 Micro-Electromechanical Systems................................................... 549
16.2.5 Charge Amplifier............................................................................... 550
16.2.6 Reciprocity Calibration Technique................................................... 552
xiv Contents

16.3 Chemical and Biological Sensors................................................................... 555

16.3.1 Detection of Small Chemical Molecules........................................... 556
16.3.2 Optical Absorption Chemical Spectroscopy..................................... 558
16.3.3 Raman Spectroscopy......................................................................... 560
16.3.4 Ion Mobility Spectroscopy................................................................ 562
16.3.5 Detecting Large Biological Molecules..............................................564
16.4 Nuclear Radiation Sensors............................................................................. 566
16.5 MATLAB® Examples..................................................................................... 569
16.6 Summary........................................................................................................ 570
Problems.................................................................................................................... 572
References................................................................................................................. 572

Chapter 17 Intelligent Sensor Systems........................................................................................ 575

17.1 Automatic Target Recognition Algorithms.................................................... 578
17.1.1 Statistical Pattern Recognition.......................................................... 578
17.1.2 Adaptive Neural Networks................................................................ 583
17.1.3 Syntactic Pattern Recognition........................................................... 590
17.2 Signal and Image Features............................................................................. 598
17.2.1 Basic Signal Metrics.......................................................................... 599
17.2.2 Pulse-Train Signal Models................................................................ 601
17.2.3 Spectral Features...............................................................................602
17.2.4 Monitoring Signal Distortion............................................................603
17.2.5 Amplitude Modulation......................................................................605
17.2.6 Frequency Modulation......................................................................607
17.2.7 Demodulation via Inverse Hilbert Transform...................................609
17.3 Dynamic Feature Tracking and Prediction.................................................... 618
17.4 Intelligent Sensor Agents................................................................................ 630
17.4.1 Internet Basics................................................................................... 631
17.4.2 IP Masquerading/Port Forwarding................................................... 632
17.4.3 Security versus Convenience............................................................. 632
17.4.4 Role of the DNS Server..................................................................... 633
17.4.5 Intelligent Sensors on the Internet.................................................... 633
17.4.6 XML Documents and Schemas for Sensors..................................... 636
17.4.7 Architectures for Net-Centric Intelligent Sensors............................. 639
17.5 MATLAB® Examples.....................................................................................640
17.6 Summary........................................................................................................640
Problems.................................................................................................................... 642
References................................................................................................................. 643

www.itpub.net
Preface
The second edition of Signal Processing for Intelligent Sensor Systems enhances many of the unique
­features of the first edition with more answered problems, web access to a large collection of
MATLAB® scripts used throughout the book, and the addition of more audio engineering, transduc-
ers, and sensor networking technology. All of the key algorithms and development methodologies
have been kept from the first edition, and hopefully all of the typographical errors have been fixed.
The addition of a chapter on Digital Audio processing reflects a growing interest in digital surround
sound (5.1 audio) techniques for entertainment, home theaters, and virtual reality systems. Also,
new sections are added in the areas of sensor networking, use of meta-data architectures using
XML, and agent-based automated data mining and control. This later information really ties large-
scale networks of intelligent sensors together as a network of thin file servers. Intelligent algorithms,
either resident in the sensor/file-server nodes, or run remotely across the network as intelligent
agents, can then provide an automated situational awareness. The many algorithms presented in
Signal Processing for Intelligent Sensor Systems can then be applied locally or network-based to
realize elegant solutions to very complex detection problems.
It was nearly 20 years ago that I was asked to consider writing a textbook on signal processing
for sensors. At the time I typically had over a dozen textbooks on my desk, each with just a few
small sections bookmarked for frequent reference. The genesis of this book was to bring together
all these key subjects into one text, summarize the salient information needed for design and appli-
cation, and organize the broad array of sensor signal processing subjects in a way to make it acces-
sible to engineers in school as well as those practicing in the field. The discussion herein is somewhat
informal and applied and in a tone of engineer-to-engineer, rather than professor-to-student. There
are many subtle nuggets of critical information revealed that should help most readers quickly
master the algorithms and adapt them to meet their requirements. This text is both a learning
resource and a field reference. In support of this, every data graph in the text has a MATLAB
m-script in support of it and these m-scripts are kept simple, commented, and made available to
readers for download from the CRC Press website for the book (http://www.crcpress.com/product/
isbn/9781420043044). Taylor & Francis Group (CRC Press) acquired the rights to the first edition
and have been relentless in encouraging me to update it in this second edition. There were also a
surprising number of readers who found me online and encouraged me to make an updated second
edition. Given the high cost of textbooks and engineering education, we are excited to cut the price
significantly, make the book available electronically online, as well as for “rent” electronically which
should be extremely helpful to students on a tight budget. Each chapter has a modest list of solved
problems (answer book available from the publisher) and references for more information.
The second edition is organized into five parts, each of which could be used for a semester course
in signal processing, or to supplement a more focused course textbook. The first two parts,
“Fundamentals of Digital Signal Processing” and “Frequency Domain Processing,” are appropriate
for undergraduate courses in Electrical and/or Computer Engineering. Part III “Adaptive System
Identification and Filtering” can work for senior-level undergraduate or a graduate-level course, as
is Part IV on “Wave Number Sensor Systems” that applies the earlier techniques to beamforming,
image processing, and signal detection systems. If you look carefully at the chapter titles, you will
see these algorithm applications grouped differently from most texts. Rather than organizing these
subjects strictly by application, we organize them by the algorithm, which naturally spans several
applications. An example of this is the recursive least-squares algorithm, projection operator sub-
space decomposition, and Kalman filtering of state vectors, which all share the same basic recursive
update algorithm. Another example is in Chapter 13 where we borrow the two-dimensional FFT

xv
xvi Preface

usually reserved for image processing and compression and use it to explain available beam pattern
responses for various array shapes.
Part V of the book covers advanced signal processing applications such as noise cancellation,
transducers, features, pattern recognition, and modern sensor networking techniques using XML
messaging and automation. It covers the critical subjects of noise, sensors, signal features, pattern
matching, and automated logic association, and then creates generic data objects in XML so that all
this information can be found. The situation recognition logic emerges as a cloud application in the
network that automatically mines the sensor information organized in XML across the sensor nodes.
This keeps the sensors as generic websites and information servers and allows very agile develop-
ment of search engines to recognize situations, rather than just find documents. This is the current
trend for sensor system networks in homeland security, business, and environmental and demo-
graphic information systems. It is a nervous system for the planet, and to that end I hope this contri-
bution is useful.

MATLAB® is a registered trademark of The MathWorks, Inc. For product information, please
contact:

The MathWorks, Inc.

3 Apple Hill Drive
Natick, MA 01760-2098 USA
Tel: 508 647 7000
Fax: 508-647-7001
E-mail: info@mathworks.com
Web: www.mathworks.com

www.itpub.net
Acknowledgments
I am professionally indebted to all the research sponsors who supported my colleagues, students,
and me over the years on a broad range of sensor applications and network automation. It was
through these experiences and by teaching that I obtained the knowledge behind this textbook. The
Applied Research Laboratory at The Pennsylvania State University is one of the premier engineer-
ing laboratories in the world, and my colleagues there will likely never know how much I have
learnt from them and respect them. A special thanks goes to Mr. Arnim Littek, a great engineer in
the beautiful country of New Zealand, who thought enough of the first edition to send me a very
detailed list of typographical errors and suggestions for this edition. There were others, too, who
found me through the Internet, and I really loved the feedback which served as an inspiration to
write the second edition. Finally to my wife Nadine, and children Drew, Anya, Erik, and Ava, your
support means everything to me.

xvii
www.itpub.net
Author
David C. Swanson has over 30 years of experience with sensor electronics and signal processing
algorithms and 15 years of experience with networking sensors. He has been a professor in the
Graduate Program in Acoustics at The Pennsylvania State University since 1989 and has done
extensive research in the areas of advanced signal processing for acoustic and vibration sensors
including active noise and vibration control. In the late 1990s, his research shifted to rotating equip-
ment monitoring and failure prognostics, and since 1999 has again shifted into the areas of chemi-
cal, biological, and nuclear detection. This broad range of sensor signal processing applications
culminates in his book Signal Processing for Intelligent Sensor Systems, now in its second edition.
Dr. Swanson has written over 100 articles for conferences and symposia, dozens of journal articles
and patents, and three chapters in books other than his own. He has also worked in industry for
Hewlett-Packard and Textron Defense Systems, and has had many sponsored industrial research
projects. He is a fellow of the Acoustical Society of America, a board-certified member of the
Institute of Noise Control engineers and a member of the IEEE. His current research is in the areas
of advanced biomimetic sensing for chemicals and explosives, ion chemistry signal processing, and
advanced materials for neutron detection. Dr. Swanson received a BEE (1981) from the University
of Delaware, Newark, and an MS (1984) and PhD (1986) from The Pennsylvania State University,
University Park, where he currently lives with his wife and four children. Dr. Swanson enjoys music,
football, and home brewing.

xix
www.itpub.net
Part I
Fundamentals of Digital
Signal Processing
It was in the late 1970s that the author first learned about digital signal processing as a freshman
electrical engineering student. Digital signals were a new technology and generally only existed
inside computer programs and as hard disk files on cutting edge engineering projects. At the time,
and reflected in the texts of that time, much of the emphasis was on the mathematics of a sampled
signal, and how sampling made the signal different from the analog signal equivalent. Analog signal
processing is very much a domain of applied mathematics, and looking back over 40 years later, it
is quite remarkable how the equations we process easily today in a computer program were imple-
mented eloquently in analog electronic circuits. Today there is little controversy about the equiva-
lence of digital and analog signals except perhaps among audio extremists/purists. Our emphasis in
this part is on explaining how signals are sampled, compressed, and reconstructed, how to filter
signals, how to process signals creatively for images and audio, and how to process signal informa-
tion “states” for engineering applications. We present how to manage the nonlinearity of converting
a system defined mathematically in the analog s-plane to an equivalent system in the digital z-plane.
These nonlinearities become small in a given low-frequency range as one increases the digital
sample rate of the digital system, but numerical errors can become a problem if too much oversam-
pling is done. There are also options for warping the frequency scale between digital and analog
systems.
We present some interesting and useful applications of signal processing in the areas of audio
signal processing, image processing, and tracking filters. This provides for a first semester course to
cover the basics of digital signals and provide useful applications in audio and images in addition to
the concept of signal kinematic states that are used to estimate and control the dynamics of a signal
or system. Together these applications cover most of the signal processing people encounter in
everyday life. This should help make the material interesting and accessible to students new to the
field while avoiding too much theory and detailed mathematics. For example, we show frequency
response functions for digital filters in this part, but we do not go into spectral processing of signals
until Part II. This also allows some time for MATLAB® use to develop where students can get used
to making m-scripts and plots of simple functions. The application of fixed-gain tracking filters on
a rocket launch example will make detailed use of signal state estimation and prediction as well as
computer graphics in plotting multiple functions correctly. Also, using a digital photograph and
2 Signal Processing for Intelligent Sensor Systems with MATLAB®

two-dimensional low- and high-pass filters provide an interesting introduction to image processing
using simple digital filters. Over 40 years ago, one could not imagine teaching signal processing
fundamentals while covering such a broad range of applications. However, any cell phone today has
all of these applications built in, such as sampling, filtering, and compression of the audio signal,
image capture and filtering, and even a global positioning system (GPS) for estimating location,
speed, and direction.

www.itpub.net
 1 Sampled Data Systems
Figure 1.1 shows a basic general architecture that can be seen to depict most adaptive signal process-
ing systems. The number of inputs to the system can be very large, especially for image processing
sensor systems. Since an adaptive signal processing system is constructed using a computer, the
inputs generally fall into the categories of analog “sensor” inputs from the physical world and digital
inputs from other computers or human communication. The outputs also can be categorized into
digital information, such as identified patterns, and analog outputs that may drive actuators (active
electrical, mechanical, and/or acoustical sources) to instigate physical control over some part of the
outside world. In this chapter, we examine the basic constructs of signal input, processing using
digital filters, and output. While these very basic operations may seem rather simple compared to
the algorithms presented later in the text, careful consideration is needed to insure a high-fidelity
adaptive processing system. Figure 1.1 also shows how the adaptive processing can extract the
salient information from the signal and automatically arrange it into XML (eXtensible Markup
Language) databases, which allows broad use by network processes. Later in the book we will dis-
cuss this from the perspective of pattern recognition and web services for sensor networks. The next
chapter will focus on fundamental techniques for extracting information from the signals.
Consider a transducer system that produces a voltage in response to some electromagnetic or
mechanical wave. In the case of a microphone, the transducer sensitivity would have units of
volts/Pascal. For the case of a video camera pixel sensor, it would be volts per lumen/m 2, while
for an infrared imaging system the sensitivity might be given as volts per degree Kelvin. In any
case, the transducer voltage is conditioned by filtering and amplification in order to make the best
use of the analog-to-digital converter (ADC) system. While most adaptive signal processing sys-
tems use floating-point numbers for computation, the ADC converters generally produce fixed-
point (integer) digital samples. The integer samples from the ADC are further converted to
floating-point format by the signal processor chip before subsequent processing. This relieves the
algorithm developer from the ­problem of controlling numerical dynamic range to avoid underflow
or overflow errors in fixed-point processing unless lesser expensive fixed-point processors are
used. If the processed signals are to be output, then floating-point samples are simply reconverted
to ­integer and an analog voltage is produced using a digital-to-analog converter (DAC) system
and ­filtered and amplified.

1.1 A/D CONVERSION

Quite often, adaptive signal processing systems are used to dynamically calibrate and adjust input
and output gains of their respective ADC and DAC devices. This extremely useful technique requires
a clear understanding of how most data acquisition systems really work. Consider a generic succes-
sive approximation 8-bit ADC as seen in Figure 1.2. The operation of the ADC actually involves an
internal DAC that produces an analog voltage for the “current” decoded digital output. A DAC sim-
ply sums the appropriate voltages corresponding to the bits set to 1. If the analog input to the ADC
does not match the internal DAC output, the binary counter counts up or down to compensate. The
actual voltage from the transducer must be sampled and held constant (on a capacitor) while the
successive approximation completes. On completion, the least significant bit (LSB) of the digital
output number will randomly toggle between 0 and 1 as the internal D/A analog output voltage
converges about the analog input voltage. The “settling time” for this process increases with the
number of bits quantized in the digital output. The shorter the settling time, the faster the digital

3
4 Signal Processing for Intelligent Sensor Systems with MATLAB®

XML
database
Input Extracted
sensing ADC information
system Adaptive
signal Web
processing services
Output system
control DAC Commands and
actuator configuration

FIGURE 1.1 A generic architecture for an adaptive signal processing system, including sensor inputs, ­control
outputs, and information formatting in XML databases for access through the Internet.

output sample rate may be. The toggling of the LSB as it approximates the analog input signal leads
to a low level of uniformly distributed (between 0 and 1) random noise in the digitized signal. This
is normal, expected, and not a problem as long as the sensor signal strengths are sufficient enough
such that the quantization noise is small compared to signal levels. It is important to understand how
transducer and data acquisition systems work so that the adaptive signal processing algorithms can
exploit and control their operation.
While there are many digital coding schemes, the binary number produced by the ADC is usu-
ally coded in either offset binary or in two’s complement formats [1]. Offset binary is used for either
all-positive or all-negative data such as absolute temperature. The internal DAC in Figure 1.2 is set
to produce a voltage Vmin that corresponds to the number 0, and Vmax for the biggest number or 255
(11111111), for the 8-bit ADC. The largest number produced by an M-bit ADC is therefore 2M − 1.
The smallest number, or LSB, will actually be wrong about 50% of the time due to the approxi­mation
process. Most data acquisition systems are built around either 8-, 12-, 16-, or 24-bit ADCs giving
maximum offset binary numbers of 255, 4095, 65535, and 16777215, respectively. If a “noise-less”
signal corresponds to a number of, say 1000, on a 12-bit A/D, the signal-to-noise ratio (SNR) of the
quantization is 1000:1, or approximately 60 dB.
Signed numbers are generally encoded in two’s complement format where the most significant
bit (MSB) is 1 for negative numbers and 0 for positive numbers. This is the normal “signed integer”
format in programming languages such as “C.” If the MSB is 1 indicating a negative number, the

Digital output

Internal
8-bit
8-bit
counter
DAC

Analog a
If a > b: count down
input If b > a: count up
b

Sample and hold Comparator

FIGURE 1.2 A generic successive approximation type 8-bit ADC showing the internal DAC converter to
compare the counter result to the input voltage.

www.itpub.net
Sampled Data Systems 5

magnitude of the negative binary number is found by complementing (changing 0–1 or 1–0) all of
the bits and adding 1. The reason for this apparently confusing coding scheme has to do with the
binary requirements of logic-based addition and subtraction circuitry in all of today’s computers
[2,3]. The logical simplicity of two’s complement arithmetic can be seen when considering that
the sum of 2 two’s complement numbers, N1 and N2, is done exactly the same as for offset binary
numbers, except any carryover from the MSB is simply ignored. Subtraction of N1 from N2 is done
simply by forming the two’s complement of N1 (complementing the bits and adding 1), and then
­adding the two numbers together ignoring any MSB carryover. An 8-, 12-, 16-, or 24-bit two’s
complement ADC with numbers over ranges of (+127, −128), (+2047, −2048), (+32767, −32768), and
(+8388607, −8388608), respectively.
Table 1.1 shows two’s complement binary for a 3-bit ±3.5 V A/D and shows the effect of sub­
tracting the number +2 (010 or +2.5 V) from each of the possible 3-bit numbers. Note that the
complement of +2 is (101) and adding 1 gives the “two’s complement” of (110), which is equal to
numerical −2 or −1.5 V in Table 1.1.
As can be seen in Table 1.1, the numbers and voltages with an asterisk are rather grossly in error.
This type of numerical error is the single most reason to use floating-point rather than fixed-point
signal processors. It is true that fixed-point signal processor chips are very inexpensive, lower power,
and faster at fixed-point arithmetic. However, a great deal of attention must be paid to insuring that
no numerical errors of the type in Table 1.1 occur in a fixed-point processor. Fixed-point processing
severely limits the numerical dynamic range of the adaptive algorithms used. In particular, algo-
rithms involving many divisions, matrix operations, or transcendental functions such as logarithms
or trigonometric functions are generally not good candidates for fixed-point processing. All the
subtractions are off by at least 0.5 V, or half the LSB. A final point worth noting from Table 1.1 is
that while the analog voltages of the ADC are symmetric about 0 V, the coded binary numbers are
not, giving a small numerical offset from the two’s complement coding. In general, the design of
analog circuits with nearly zero offset voltage is a difficult enough task that one should always
assume some nonzero offset in all digitized sensor data.
The maximum M-bit two’s complement positive number is 2M−1 − 1 and the minimum negative
number is −2M−1. This is because one of the bits is used to represent the sign of the number and one
number is reserved to correspond to zero. We want zero to be “digital zero” and we could just leave
it at that but it would make addition and subtraction logically more complicated. That is why two’s
complement format is used for signed integers. Even though the ADC and analog circuitry offset is
small, it is good practice in any signal processing system to numerically remove it. This is simply
done by recursively computing the mean of the A/D samples and subtracting this time-averaged
mean from each ADC sample.

TABLE 1.1
Effect of Subtracting 2 from the Range of Numbers from a 3-bit Two’s Complement A/D
Voltage N Binary N Binary N2 Voltage N2
+3.5 011 001 +1.5
+2.5 010 000 +0.5
+1.5 001 111 −0.5
+0.5 000 110 −1.5
−0.5 111 101 −2.5
−1.5 110 100 −3.5
−2.5 101 011* +1.5*
−3.5 100 010* +0.5*
Random documents with unrelated
content Scribd suggests to you:
 “Then have at me, Robert—proceed!”
“Im-mediate, sir!”
And Sir John’s toilet commenced forthwith; during which nice
business they conversed as follows:

Sir John: Any news, Bob?

Robert: Nothing to mention, sir ... though I did ’appen to hear
that five soldiers and two o’ the coastguard are reported
wounded, sir.
Sir John: Nothing serious, I hope?
Robert: We hope not, sir.
Sir John: An ugly business, Bob.
Robert: On-commonly, sir!
Sir John: Have you seen or heard anything of Mr. Potter?
Robert: No, sir. It seems he’s vanished away again, being
badly wanted by the preventive authorities. For I did
’appen to hear as ’twas him as is judged responsible for
most o’ the casualties, sir.
Sir John: To be sure, he was wearing his old frieze coat! Ha’
you been far abroad this morning, Bob?
Robert: I did ’appen to step across the fields, sir.
Sir John: Very right, Bob. Health! Sunshine! Dew!
Robert: It was a little doo-ey, sir.
Sir John: And you carried the basket, Bob, of course?
Robert: Basket, sir...?
Sir John: Her basket, Bob ... and pray keep the shaving-
brush out o’ my mouth!
Robert: Your pardon, sir!
Sir John: Her basket, Bob!
Robert: I judged it over heavy for a young fe——
Sir John: Damsel, Bob.
Robert: Yes, sir.
Sir John: To be sure ’twas too heavy—and I fancy you ha’
lathered me enough.
Robert: I think so too, sir.
Sir John: She hath a remarkably neat foot, Bob!
Robert: I have ob-served same, sir.
Sir John: And her voice grows upon one.... A voice
suggestive of a nature sweet and——
Robert: One moment, sir—your upper lip!
(A moment’s silence while the Corporal plies deft razor.)
Sir John: I chanced to see her in converse with a young ...
creature, Robert—a tall young woman in a laced cap?
Robert: I re-marked same young person myself, sir.
Sir John: Is she a friend of Mistress Ann’s?
Robert: Not knowing, can’t say, sir.
Sir John: Do you chance to know anything about this—er—
young person?
Robert: Nothing, sir, except as she seems to run very much
to legs——
Sir John: Legs—begad!
Robert: Pre-cisely, sir ... leggy, your honour.
Sir John: Ha, leggy! Didn’t you think her a young goddess?
Robert: She didn’t strike me as such, sir.
Sir John: But you must ha’ remarked her beauty?
Robert: Nothing to mention, sir.
 Sir John: But damme—her shape! Her form! Her air! Her
carriage! Her grace!
Robert: Too much of ’em all, sir.
Sir John: ’S death, man—you must be blind!
Robert: Very good, sir.
Sir John: No, Bob, not blind—thou’rt merely in love and that
is infinitely worse.
Robert: It is, sir!
Sir John: Why, then, go a-wooing, man, go a-wooing and put
thyself out o’ thy misery one way or t’other.
Robert: Can’t be done, sir. Misery must be endoored.
Sir John: Because thou’rt forty-five, Bob?
Robert: And she’s scarce twenty turned, sir.

“Ha!” exclaimed Sir John portentously. “Hum!” And, his toilet at last
accomplished, he ran lightly down the stair to find awaiting him a
most inviting breakfast, of which he made short work, despite Mr.
Bunkle’s shocked remonstrances and reproachful looks.
“This here b’iledam, sir,” quoth Mr. Bunkle, caressing the edible in
question with the fork of an expert—“this here b’iledam desarves to
be ate respectful an’ dooly slow, wi’ thought to every chew an’ a
pause betwixt each swaller!”
“Forgive me, Mr. Bunkle,” smiled Sir John as he rose from the
table, “but, like the chameleon, I could feed on air—for a time at
least! Robert, my holly-stick! I think I will call on our Ancient Mr.
Dumbrell. Have ye any message, Bob?”
“None, sir.”
“Why, then, I must invent some. You might step over to Dering
later in the day, Robert. Adieu, Mr. Bunkle.”
“Dinner at ’arf-past twelve, sir!” sighed Mr. Bunkle, laying down the
carving-fork, “roast Sir Loin—’ot!”
 CHAPTER XXVII
TELLETH HOW MR. DERWENT BEGAN HIS
WOOING

Away strode Sir John across sunny fields, light of foot, treading a
springy turf, breathing a fragrant air, swinging his holly-stick and
vaulting stiles for the pure joy of it all. Birds piped and chirped from
hedge and thicket, larks carolled in the blue, rills bubbled and
laughed, and scabious flowers danced and swayed in the gentle
wind in tune with the universal gladness.
And so in good time came Sir John to High Dering. For there,
perched upon his accustomed stile in well-brushed hat and snowy
smock-frock, sat the Ancient Person in animated converse with one
who leaned gracefully against the gnarled post of the old stile,
listening to the Aged One’s talk, but watching Sir John from the
shadow of her hat, with eyes quick to heed all the careless, easy
grace of him as he came light-treading across the sun-dappled ling.
“Rose!” said he, and bared his head; now, beholding her startled,
upward glance, how should he know of the eyes that had taken such
note of his altered appearance, his plain attire? “Rose,” said he,
“thou rose of love!” And stood bare-headed, glad-eyed, to await her
greeting.
“La, Mr. Derwent,” said she, “you wear strange, small hat, sir, yet
methinks it do become you better than your night-cap!”
“And yet ’tis a very excellent night-cap!” he retorted.
“Eh—eh?” piped the Aged One. “Be ye man an’ woife, then?”
“Not yet, Mr. Dumbrell, but——”
“Then wot’s she know about your noight-cap, young man, eh—eh?
Tell oi that!”
 “I—I saw it this morning,” explained my lady, rather hastily—“this
morning as he leaned out of his chamber window——”
“Then, young man, ’ow dare ’ee stick y’r noight-cap out o’ winder
in a purty maid’s face? Shamed at ’ee, oi be!”
“But I drew it in again, Mr. Dumbrell!”
“No matter, young man, oi be shamed at ’ee! Wi’ y’r noight-cap an’
arl!”
“It shall not happen again, Mr. Dumbrell.”
“Oi be a ol’, ancient man, aye—a aged soul, oi be, an’ oi knaws
wot oi knaws an’ oi knaws as us doan’t want ’ee, young man, wi’
your noight-cap, an’ arl!” Here the Aged One glared at the intruder
with truculent eye, but Sir John was looking at my lady, of course.
“So I have found thee at last, my Rose!” said he softly.
“Ha’ you looked for me, sir?”
“These very many weary days, child.”
“Your honour expected me, then?”
“Hourly.”
“And now that you behold me?”
“Now, Rose, the sun shines, the birds sing, the scabious flowers
are a-dance in their myriad hosts, and here standeth John Derwent
to woo thee——”
“Well, go ’way!” snarled the Aged One fiercely. “Go ’way; us doan’t
want ’ee no’ow, young man! Us be a-’arking to each other an’ doan’t
want nobody—du us, my pretty? Lord, ’e du ha’ put me out! Wot was
oi a-tellin’ ye, my dainty dear?”
“Of the day you and Sir Hector saved old Penelope the witch from
being drowned ... but the sun is very hot, pray put your hat on again,
Mr. Dumbrell! Nay, suffer me!” So saying, my lady took the well-
brushed hat and set it upon the old, white head so gently and with
such pretty grace that the Aged One leered at Sir John in chuckling
triumph.
 “Us doan’t want ’ee, young man, du us, my flower?”
“Indeed,” she laughed, “but you find wondrous pretty names for
me——”
“Because ’tis purty you be ... no, ’andsome’s the word—a foine
’andsome wench.”
“But over-large for a flower, I fear,” she sighed.
“Sizeable!” nodded the Aged One. “But oi loikes ’em big—allus
did. So doan’t ’ee worrit naun ’count o’ y’r size. An’ as fur ol’
Penelope, ’er desarved arl ’er got, bein’ a witch.... An’ when it come
to savin’ of ’er, I dunno as Sir ’Ector done so tur’ble much! Oi be an
ol’ ancient man, but oi bean’t nowise doddlish, an’ can save a witch
as well as some young ’uns an’ better’n most—ah, that oi can!”
“I’m sure of it! And is she still alive?”
“That she be. Witches bean’t easy to kill an’ doan’t aften doi—not
in Sussex, they doan’t. Oi been buryin’ folk arl my days an’ oi only
buried one witch, an’ ’er only doied because she ’appened to drown,
not being able to swim wi’ a stone round ’er neck, d’ye see——”
“A—a stone?” exclaimed my lady in tones of horror.
“Aye, a stone fur sure, my pretty. Toied ’un round ’er neck, they
did, an’ ’ove ’er into the river, they did, an’ so ’er doied. But this were
years an’ ages ago, when oi were younger. And ol’ Penelope be a
tur’ble powerful witch—give me a spell agin the axey as done me arl
manner o’ good.”
“Did she cure you by magic?”
“Lord bless y’r pretty eyes—no! There bean’t nobody nor nothink
can cure oi, what wi’ that theer ol’ musket-ball o’ mine. But oi were
moighty bad, an’ ’long come a man one day in a p’inted ’at an’ a
gownd wi’ silver stars on to it an’ sold me a charm wrote on a three-
carnered piece o’ paper wi’ these words as oi were to say three
toimes over, marnin’, noon an’ noight:

Axey, axey oi defoi thee,

Three days shiver, three days shake,
 Mak’ me well fur Marcy’s sake.

Well, oi sez ’em over an’ over ’till oi were black i’ the faace, but it
didn’t seem tu du me no good at arl, ’till one day ’long comes ol’
Penelope, tears up my charm an’ gi’es me some stuff in a liddle
bottle as oi must rub arl over myself ... which oi done. An’ Lord—
arter a bit oi got that skittish—used t’ kick up my ’ind legs loike any
colt ... an’ me a married man an’ arl. Oi dunno as if oi——”
“Grandfeäther!”
“Dannle it! That be my rum-an’-milk!” exclaimed the Aged One,
scowling.
“Grandfeäther, be ye comin’?”
“Arl roight, lass, arl roight!” piped the old man pettishly, getting
from his perch with surprising nimbleness. “Oi’ll ’ave to go, my pretty
bird, oi’ll ’ave to leave ’ee or ’twill be milk an’ no rum! Ann be that
’ard-’earted an’ ... Arl roight, Nan, ’ere oi be!” This as his
granddaughter appeared, who, beholding Sir John, blushed and
curtsied. Quoth she:
“’Tis tur’ble kind o’ you to bide an’ keep ’im comp’ny, Rose—mam,
for ’e du be that mischievious——”
“Never tak’ no ’eed o’ my Nan, ’er’s a babe!” retorted the Aged
One. “An’ oi du ’ope as you’ll come an’ talk tu oi again, my Beauty
Broight, fur oi doan’t tak’ naun account o’ little ’uns, an’ you be a
foine up-standin’, down-sittin’ wench, sure-ly! An’ the young ’un ’ere
thinks the same, doan’t ’ee, young man?”
“I do!” answered Sir John fervently. “Indeed, I have never seen a
more up-standing, down-sitting wench in all my life!”
“Well, then, whoy doan’t ’ee up an’ tell ’er so, wi’out me a-doin’ it
fur ’ee. You be sweet on ’er, oi s’pose?”
“Monstrous so!”
“Well, then, whoy caan’t ’ee tell ’er summat about it? Ye caan’t
expect oi tu du it fur ’ee arl the toime. ’Ere you’ve stood a-lookin’ an’
a-starin’ an’ so silent as a turmut! That bean’t no waay tu win a
wench—no! Lord, oi were different in my young days; oi knawed the
waay tu go a-wooin’! An’ oi ain’t forgot yet, though I be such a ol’,
aged soul!”
“Then perhaps you will help me, now and then?” Sir John
suggested.
“Whoy, sence you ax me so sensible an’ modest-loike, oi dunno as
oi wun’t. For, if you bean’t much to look at, you be batter’n some, an’
she moight du worse.”
“It is possible!” sighed Sir John.
“So oi dunno as oi wun’t put in a word for ’ee noo an’ then wi’ the
lass. But moind ye if oi win ’er for ’ee an’ she doan’t turn out arl as
you expect, an’ woives never do no’ow, doan’t ’ee go fur to blame
oi!”
“Grandfer, your rum-an’——”
“Hesh a minute, Nan, hesh an’ lemme finish, will ’ee? Marriage,
young man, be arl roight whiles ye be single, but when you be
married ’tis generally-mostly-arlways arl wrong—oi’ve troid it twoice,
an’ oi knaw! So jest so soon as she begins to feel weddin’ish, oi
leaves the matter to you. An’ now, Nan, gimme y’r arm!”
“Boide a minute, Grandfeäther——”
“Whaffor, Nan? Ain’t ye kep’ me a-waiting long enough?”
“I’ve a message for the gen’elman——”
“Gen’elman, lass? ’Oo? Wheer? D’ye mean—’im?” And the Aged
One pointed at Sir John with wavering stick. “’E bean’t no gen’elman
—look at ’is ’at! Gen’elman’s ’ats ’as goold lace onto ’em loike Sir
’Ector’s of a Sunday an’ Lord Sayle’s of a week-day. Look at ’is coat
—so plain! An’ ’e aren’t got no sword neether! Gen’elman—’im? ’E
be jest a respectable young man——”
“You hear that, Rose?” cried Sir John, ecstatic. “You hear? There
speaketh hoary Wisdom!”
“’Oo’s ’oary—me?” demanded the Aged Soul, scowling.
 “Yourself, Mr. Dumbrell, and are therefore to be revered. Your
hand, Sir Reverence, your hand, I beg!”
“Whoy, oi dunno as oi loike the sound o’ that ’ere word——”
“Mr. Dumbrell, you in your nescience saw ’neath the hollow shams
and know me for what I truly am, a respectable young man. O most
excellent Aged Soul, I thank thee for that word! Mr. Dumbrell, your
hand, pray.”
So, after some little hesitation, the sharp-tongued, little old man
reached tremulous hand to Sir John’s warm clasp, and, looking up
into Sir John’s smiling eyes, the Aged Soul smiled also; quoth he:
“Young man, oi dunno as you bean’t better-lookin’ than what oi
thought—leastways your eyes is worth any lass a-lookin’ at, oi
rackon, an’—whoy, what be this ’ere?” And the old man stared down
at his open palm. “By the pize—a guinea! Dannel it, young man,
what be this fur? What do ’ee mean by it?”
“Do not be angry, Mr. Dumbrell; pray accept it as a small mark of
esteem and gratitude from one respectable man to another.”
“Whoy, since you puts it that ways, young man, we woan’t arg’
about it, an’ oi dunno as oi bean’t almoighty glad of’t.... A guinea,
Nan, a goolden guinea! ’Ere be baccy for oi an’ that ’ere cherry
ribband for you, an’ sugar for oi, an’ a noo ’at for oi.... Young man, oi
thank ’ee, an’ so du Nan.... Thank ’un, Nan; mak’ y’r reverence an’
show y’r manners, lass!”
“Not forgetting your message, Ann,” prompted Sir John.
“Yes, sir,” she answered, curtsying repeatedly, “though ’twere only
Gammer Haryott as bid me say if I see you, sir, as she would like a
word wi’ you, sir.”
“What about my rum-an’-milk?” demanded the Aged Soul pettishly.
“’Ere be oi a-vadin’ an’ famishin’ an’ perishin’ awaay, an’ you a-
maggin’ an’ me a-waitin’ an’ nobody to ’tend to oi no’ow, nowhen nor
nothin’! Come an’ gimme my rum-an’-milk or no ribbands, moind
that! G’marnin’, young man, an’ doan’t ’ee go a-throwin’ your money
away so woild-loike an’ rackless! Marnin’, my purty dear! You’ll foind
oi settin’ a-top o’ stoile every marnin’ when it be sunny.” So saying,
the Aged Soul bared his white head gallantly, nodded, and suffered
his dutiful granddaughter to lead him away.
My lady was silent awhile, watching them as they went, the girl so
young and strong and motherly, the old man so bowed and feeble;
and Sir John, regarding his companion keen-eyed, saw in her look
an unwonted tenderness and, when at last she spoke, heard her
voice strangely tender also.
“O Sussex!” she murmured. And then: “They are worth caring for,
these unspoiled folk o’ the Down Country.”
“They are, Herminia!” he answered. At this she turned and looked
at him, frowning a little.
“Have you done so, Sir John?” she questioned. “Have you cared
for their comfort and welfare?”
“Alas, no!” he answered. “I, like you, my lady, have preferred the
town hitherto, and, heaven help me, was therewith fairly content!
Which is matter for some wonder, for here were the Downs and here
the Dumbrell——”
“That Aged Soul!” she added, smiling suddenly. “As gallant as any
town beau, more dignified, and infinitely more sincere.”
“Rose child, I perceive thou hast also found eyes to see withal!”
“Is this so amazing, your honour?”
“Not so much as to behold a fine lady who honours Rusticity and
finds joy in simple, homely things.”
“Indeed, sir, I do love the country, especially Sussex, for, as your
honour may ha’ forgot, I was born here.”
“Then, if you will, I can show you other wonders. First, there is
Dame Penelope Haryott, whom fools call a witch and rogues have
sought to murder, ere now.”
“Murder!” exclaimed my lady, wrinkling her brow. “Oh! And yet
surely witches be horrid creatures! Ha’n’t you read of ’em?...
Leagued with all manner of evil spirits for the working of evil.... Ha’n’t
you read what learned philosophers ha’ writ concerning’ em, sir?”
“Aye, I have.”
“Well, if this woman be truly a witch——”
“But was there truly ever a witch, child?”
“Your honour may have heard of the Witch of Endor?”
“Hum!” quoth Sir John. “Can it be that you believe in witchcraft,
black magic and the like fooleries?”
“Don’t you, sir?”
“No more than I do in ghosts, child.”
“The girl Ann tells me that ghosts often walk in these parts.”
“Aye, so they do,” laughed Sir John, “and to some purpose.”
“Then, despite the Bible and philosophers, your superior wisdom
doth not believe in witches?”
“No, indeed.”
“Nor ghosts?”
“No, child.”
“Because you chance never to ha’ seen one, sir!”
“Because I have, rather. Indeed, Rose, a most effective ghost——”
“You have positively seen a ghost? When? Where?” she
demanded. But, turning a bend in the road they came upon a
horseman, a cadaverous person in threadbare clerical garb, who
bestrode a very plump steed.
“A fair prospect to the eye!” he exclaimed, nodding gloomily
towards Dering village, where it nestled under the sheltering Down.
“Aye, a fair prospect, and yet, in very truth, a ‘whited sepulchre’ ...
not a thatch that doesn’t leak, scarce a cottage that is truly habitable
——”
“Shameful!” exclaimed my lady.
 “And wicked!” added the parson in his gentle voice, his haggard
face very woeful. “For how shall folk take heed to their soul’s welfare
until their bodies be comfortable? Alas, you behold yonder the evils
of a bad landlord. Sir John Dering hath much to answer for. Better he
were dead and the land in better keeping.”
“Dead, sir!” exclaimed my lady, aghast.
“And wherefore not?” continued the parson in his gentle accents,
while his eyes smouldered. “A merciless, grinding bailiff and a
profligate landlord make for a suffering tenantry.”
“You are the Reverend Mr. Hartop, I think, sir?” questioned Sir
John, bowing.
“The same, sir,” answered the parson, returning the salute. “And I,
who know and love these rustic folk, say again that for the general
good, an evil landlord is better dead.... And consider Sir John
Dering’s reputation, his scandalous life!”
“True!” sighed Sir John; “his reputation doth show him a very
monster of iniquity.”
“God forgive him!” sighed the parson. “Duellist and man of blood,
desperate gambler and of wild, unholy life.... A few poor hundreds of
the guineas he throws away at the gaming-table or wastes on
nameless evil would mean all the difference ’twixt misery and
happiness, sickness and health to the folk of High Dering. Heaven
forgive the Wicked Dering the evil he hath wrought.”
“Amen!” added Sir John. “How potent and far-reaching is a man’s
reputation, Rose!”
“How different the son from his honoured sire!” sighed Mr. Hartop.
“Alas, yes, sir!” answered Sir John. “And yet, sir, I have it on
excellent authority that this most iniquitous gentleman hath lately
become a ‘respectable young man.’”
“Sir,” exclaimed the parson, opening his mild eyes a little wider
than usual, “sir, you amaze me! Heaven send it be indeed so, for his
own sake and the future welfare of his neglected people.” Saying
which, Mr. Hartop lifted shabby hat and rode gloomily away.
 “‘For the general good,’” repeated Sir John wistfully, “‘for the
general good an evil landlord were better dead.’ Here is an arresting
thought, child ... and how bitterly true!”
“But you are alive!” said she, staring towards the quiet village
beneath wrinkled brows. “Live, then, to better purpose.”
“Ah, Rose,” he sighed, “thy pretty moralities fall so trippingly from
thy rosy, innocent lip; thou art in thy simple wisdom such an angel of
inspiration that I would we had met ... five weary years ago!”
“Five years ago?” she repeated, turning upon him. “Have you
forgot——?” Here, beholding his grim-smiling mouth, the mockery of
his eyes, she caught her breath and was silent.
“Five long years ago, child, I killed a man—by accident. Ah, sweet
Rose, gentle maid, if only thou hadst come to me then ... to soothe
my bitter grief! Dear, lovely Rose, that little ‘if’ held, then as now, a
world of possibilities even for such an abandoned wretch as ‘the
Wicked Dering.’ But we are still alive, and to live is to hope.... And
Dame Haryott desires speech with me. And thou would’st behold a
witch, so come thy ways with thy loving, gentle John.”
“Gentle?” cried she angrily. “Aye, with the eyes of a mocking
fiend!”
“But the heart of a respectable young man, Rose!”
“Your crime brought its own consequences, sir.”
“It did!” he sighed. “And not the least of ’em, thyself! When wilt
marry me?”
“Never!”
“Then the matter being settled—for the present let us to the witch,
hand in hand like good friends.”
“No!”
“Yes!”
“Leave me, sir!”
“Give me thy hand.”
 “Oh—I hate you!” she cried passionately.
“Good!” he nodded placidly. “’Tis better than indifference. Thy
hand, Rose.”
For answer she turned away, silently contemptuous, and began to
retrace her steps; but he caught her wrist and checked her suddenly,
whereupon she struck viciously at him, knocking off his hat, then her
other hand was ’prisoned also in so tense a grip that, knowing it vain
to struggle, she disdained further effort and faced him, coldly defiant.
“Coward, you hurt me!”
“Madam, you behave like a peevish hoyden! Such tricks may pass
with your hysterical fine ladies but, while in Sussex, I suggest you
ape the dignified calm o’ Rusticity.”
“Will you loose me?”
“Are you done with your fishwifely tantrums?”
My lady held herself pridefully, glared furiously, then suddenly bit
her lip, bowed her head, and something bright and sparkling fell
upon his hand; at this he loosed her suddenly and she as suddenly
turned her back upon him.
Sir John picked up his hat, knocked the dust from it, put it on, and
stood regarding her pensively.
“Rose,” said he at last, “dear child, suffer me to take thy hand.”
Then he reached and clasped her unresisting fingers; and thus, hand
in hand, they went on down the lane together.
 CHAPTER XXVIII
TELLETH HOW MY LADY ADOPTED A FAIRY
GODMOTHER

High Dering, drowsing in the sun, opened a door here and there to
stare in idle wonderment as Sir John handed his companion in at
Dame Haryott’s garden wicket, for visitors were rare, more especially
such visitors as these who bowed and curtsied to each other with
such courtly, albeit frigid, ceremony; so High Dering opened its doors
a little wider and became a trifle more awake as Sir John knocked.
And, after some while, chains rattled, bolts creaked, the heavy
door opened, and old Penelope stood peering at them from the dim
interior.
“Good-day, Mrs. Penelope,” said Sir John, removing his hat and
saluting her in his easy, unaffected manner. “You desired to see me,
I think?”
“Aye, I did,” she answered ungraciously, “but not along of a tattlin’
wench.”
My lady stared and flushed angrily.
“I will go!” said she, and drawing herself to her noble height, turned
away, supremely disdainful as an outraged goddess; but old
Penelope, who knew little of goddesses and cared less, was no whit
abashed.
“Hoity-toity!” quoth she; “bide a bit, wench!” and my Lady Herminia
found her stately progress checked by the crook of old Penelope’s
stick that had hooked itself suddenly about her arm.
My lady turned and, amazed beyond speech, viewed the
audacious old creature from head to foot until, meeting the fierce old
eyes, her gaze paused there and thus, for a long moment, they
stared at each other, the old woman and the young, while Sir John
wisely held his peace.
“Ha!” exclaimed Dame Haryott at last, looking more malevolent
and witch-like than usual, “an’ who be you, young mistress, wi’ y’r
white ’ands, an’ dressed out like a country-lass, as do carry y’rself so
proud-like? Hush and I’ll tell ’ee. You be one as long loved Love, an’
sought it vainly till, one day, ye found it—in your own heart ... the
love for a man——”
“I—I love no man!” cried my lady, with a strange vehemence.
“Bah!” quoth Penelope harshly, “’tis peepin’ at me from y’r eyes,
flushin’ in y’r cheek. First, ’twas love o’ y’rself, which was a bad love,
but now ... aha, now it be love for a man! A love as shall grow an’
grow till it be a pain ... some love be a pain, I know ... and ’tis the
only love worth ’aving!”
“I love no man!” repeated my lady.
“Shall I speak his name, mistress?”
“No—no!” answered my lady, a little breathlessly.
“Oho!” chuckled old Penelope in most witch-like manner. “Oho! ...
‘no, no!’ quo’ she!... An’ ’er so proud an’ arl! But I know, aye, ol’ Pen
knows! For I loved once when the world was younger an’ kinder.... I
were tall then, and nigh prideful as you, afore age an’ sorrow bent
me an’ love humbled me. Love? Aye, but ’twas worth the pain, for
’twas a love hath sweetened the bitter o’ the long, weary years, an’
cheered my loneliness ... a love as I shall tak’ wi’ me to a better
place an’ find Happiness at last, maybe—Happiness ... after s’much
bitter solitude!”
Suddenly the old eyes were upturned to the radiant heaven, their
fierceness was softened by the glitter of slow-gathering, painful
tears; and then, upon that bowed and aged shoulder came a hand, a
gentle hand yet strong, for all its white delicacy; and my lady spoke
in voice Sir John had never heard from her before:
“Art so very lonely?”
 “Lonely?” The word was a groan, and the drooping shoulders sank
lower. “I’ve been a lone soul all my days—wi’ none to care for me
since HE died, an’ none to tak’ my part except Jarge and Sir Hector
... the liddle children mock me ... the women be worse! An’ I du be
gettin’ that old and weary!... Sometimes I can scarce brave it any
more!”...
“Wilt take me for thy friend, old Penelope?”
The old woman lifted white head proudly as any person of quality
might have done and stared at my lady keenly, then reached up and
patted the hand upon her shoulder.
“’Tis come too late!” sighed she. “You be too young an’ I be too old
for friendship ... but I thank ye kindly.”
“Then you’ll suffer me to come and talk with you sometimes,
Penelope?”
“Why, ye see, the roof leaks, an’ the chimbley smokes——”
“The more shame to Sir John Dering!” exclaimed my lady fiercely.
“Aye,’twere different in the ol’ squire’s time—the other Sir John as
marched away wi’ his sojers an’ never came back ... the world was
better then ... ’specially High Dering. But to-day they name me witch,
an’ a witch’s cottage bean’t no place for young maids—’specially
your sort! But since you be here, come in an’ sit ye down—both on
ye! An’ if ye’ll wait ’till my kittle b’iles I’ll brew ye a dish o’ tea——”
“Tea?” exclaimed my lady.
“Aye, I generally tak’s a drop towards noon; it do warm my old
bones!” So saying, she led them into the cottage and very carefully
locked, bolted and chained the door.
“I do this,” she explained, “because happen they may come an’
mak’ trouble for me—sudden-like!”
“Who, pray?” demanded my lady indignantly.
“Any fule as finds ’is cow gone dry, or ’is crop blighted, or ’is horse
off its feed, or his child in a fit.... Lord bless ’ee, child, doan’t stare so!
Ye see folks thinks I’ve ‘the evil eye’ an’ can blast ’em with a look ...
aye, but I wish I could, that I du!”
“And so,” continued Sir John, “they have stoned her, set dogs on
her, and threatened her with death by water and the fire, ere now
——”
“Aye, but the dogs be worst!” cried old Penelope, giving the fire a
savage poke. “I can’t abide dogs!”
“By heaven!” exclaimed my lady in sudden ferocity, “would I were
a man!”
“By heaven!” retorted Sir John, “I rejoice that you are not!”
“Tush!” she cried angrily, “’tis time there came a man to High
Dering!”
“I have thought so too!” he answered gravely.
“Nay, I mean a strong man—a man of action!”
So saying, my lady rose, contemptuous, seeming to fill the small
place with the majesty of her presence.
“Dear Penelope,” said she gently, “suffer me to do that for you—I’ll
lay the cloth and——”
“No, no!”
“But I say yes!”
“O do ye an’ arl!” exclaimed the old woman fiercely. “This be my
own cottage till they turn me out an’ then——”
“Turn you out?”
“Aye, in two or three wiks!”
“You hear, sir; you hear?”
“I do!” answered Sir John.
“And when you are homeless, Penelope, what shall you do?”
“Walk an’ tramp ’till I caan’t go no further, an’ then find a quiet
corner to die in——”
 “Nay, that you shall not!” cried my lady passionately. “I will take ye
—you shall come to me, I will adopt you——”
“Eh—eh!” gasped old Penelope, and very nearly dropped her
cherished Chinese teapot.
“You shall come to me, Penelope,” repeated my lady, taking the
teapot from her tremulous fingers. “I shall adopt you—nay, my dear
soul, never doubt me, I mean it every word!”
“But ... but,” stammered old Penelope, “they call me a witch! They
... they——”
“Devil take ’em!” exclaimed my lady. “I will care for thee, Penelope!
Shalt find peace and comfort at last, thou brave soul!” And here,
seeing the old creature’s pitiful amaze, my lady stooped suddenly
and pressed warm lips on her wrinkled brow.
“Lord God!” exclaimed old Penelope, and sinking into the elbow-
chair, hid her face in her toil-worn hands. And presently she spoke in
voice harsh and broken, “There be nobody ... has kissed me ... since
my dyin’ mother, long an’ long ago!”
“My dear soul!” said my lady, and Sir John saw her eyes suddenly
brim with tears. “My dear soul, there is a woman shall kiss away thy
sorrows if she may.... For to-day, Penelope, thou hast found a friend
and I a—a fairy godmother! Let me kiss thee again, godmother!”
Slowly old Penelope raised her head to look into the face bowed
above her.
“Happen I be dreamin’,” she sighed, “an’ shall wake by an’ by—
but, O child, it be good to dream—sometimes.”
 CHAPTER XXIX
GIVETH SOME DESCRIPTIONS OF A TEA-
DRINKING

“’Tis most excellent tea!” quoth my lady. “I vow I have never drank
better!”
“Arl the way from Chaney, mam.”
“And these beautiful dishes!”
“Chaney, too!” nodded old Penelope proudly. “An’ look at my
teapot! I means to tak’ it along wi’ me when they do turn me out,
though ’twill be a bit ’ard to carry, I rackon. But ye see, mam, I——”
“Nay, godmother, call me Rose.”
“No, mam, it doan’t come easy to my tongue.”
“I may call you Penelope, mayn’t I?”
“For sure!”
“And fairy godmother?”
“Aye, though I be more witch than fairy, I rackon.”
“Then, godmother Penelope, pray call me Rose.”
“Rose, then!” she snapped.
“I think,” said Sir John in his pleasant voice, “you have some
message for me, Mrs. Penelope?”
“Gimme time, young man, gimme time! I bean’t kissed an’ called a
fairy every day, so gimme——” She paused suddenly and seemed to
listen intently, “I rackon you’d best be goin’—both on ye!”
“But why, pray?” demanded my lady.