PHYSIOLOGICAL

CONTROL SYSTEMS
IEEE PRESS SERIES IN BIOMEDICAL ENGINEERING
The focus of our series is to introduce current and emerging technologies to biomedical and electrical
engineering practitioners, researchers, and students. This series seeks to foster interdisciplinary
biomedical engineering education to satisfy the needs of the industrial and academic areas. This requires
an innovative approach that overcomes the difficulties associated with the traditional textbook and edited
collections.
Series Editor
Metin Akay
Dartmouth College
Advisory Board
Ingrid Daubechies
Murray Eden
James Greenleaf
Simon Haykin
Murat Kunt
Paul Lauterbur
Larry Mclntire
Robert Plonsey
Richard Satava
Malvin Teich
Herbert Voigt
Lotfi Zadeh
Editorial Board
Eric W. Abel
Peter Adlassing
Akram Aldroubi
Erol Basar
Katarzyna Blinowska
Bernadette Bouchon-Meunier
Tom Brotherton
Sergio Cerutti
Jean-Louis Coatrieux
Maurice Cohen
John Collier
Steve Cowin
Jerry Daniels
Andrew Daubenspeck
Jaques Duchene
Patrick Flandrin
Walter Greenleaf
Daniel Hammer
Dennis Healy
Gabor Herman
Helene Hoffman
Donna Hudson
Yasemin Kahya
Michael Khoo
Yongmin Kim
Andrew Laine
Rosa Lancini
Swamy Laxminarayan
Richard Leahy
Zhi-Pei Liang
Jennifer Linderman
Richard Magin
Jaakko Malmivuo
Yves Meyer
Michael Neuman
Tim Olson
Banu Onaral
Keith Paulsen
Peter Richardson
Richard Robb
Joseph Rosen
Christian Roux
Janet Rutledge
Wim L. C. Rutten
Alan Sahakian
Paul S. Schenker
G. W. Schmid-Sch nbein
Ernest Stokely
Ahmed Tewfik
Nitish Thakor
Michael Unser
Eugene Veklerov
AI Wald
Bruce Wheeler
William Williams
Andy Yagle
Yuan-Ting Zhang
Books in the IEEE Press Series in Biomedical Engineering
TIME FREQUENCY AND WAVELETS IN BIOMEDICAL SIGNAL PROCESSING
Edited by Metin Akay
1998 Hardcover 768 pp ISBN 0-7803-1147-7
PRINCIPLES OF MAGNETIC RESONANCE IMAGING: A SIGNAL PROCESSING PERSPECTIVE
Zhi-Pei Liang and Paul C. Lauterbur
2000 Hardcover 416 pp ISBN 0-7803-4723-4
NEURAL NETWORKS AND ARTIFICIAL INTELLIGENCE FOR BIOMEDICAL ENGINEERING
Donna L. Hudson and Maurice E. Cohen
2000 Hardcover 256 pp ISBN 0-7803-3404-3
PHYSIOLOGICAL
CONTROL SYSTEMS
Analysis, Simulation, and Estimation
Michael C. K. Khoo
Biomedical Engineering Department
University of Southern California
IEEE Engineering in Medicine and Biology Society, Sponsor

IEEE Press Series on Biomedical Engineering
Metin Akay, Series Editor
4HEEE
The Institute of Electrical and Electronics Engineers, Inc., New York
^WILEY-
/INTERSCIENCE
A JOHN WILEY & SONS, INC., PUBLICATION
IEEE Press
445 Hoes Lane, P.O. Box 1331
Piscataway, NJ 08855-1331
J. B. Anderson
P. M. Anderson
M. Eden
M. E. El-Hawary
IEEE Press Editorial Board
Robert J. Herrick, Editor in Chief
S. Furui P. Laplante
A. H. Haddad M. Padgett
S. Kartalopoulos W. D. Reeve
D. Kirk G. Zobrist
Kenneth Moore, Director of IEEE Press
Linda Matarazzo, Assistant Editor
Anthony VenGraitis, Project Editor
IEEE Engineering in Medicine and Biology Society, Sponsor
EMB-S Liaison to IEEE Press, Metin Akay
Cover design: William T. Donnelly, WT Design
Technical Reviewers
Professor Giuseppe Baselli, Polytechnic University Milano, Italy
Dr. Susan M. Blanchard, North Carolina State University
Professor Sergio Cerutti, Polytechnic University Milano, Italy
Professor J. Andrew Daubenspeck, Dartmouth Medical School
Professor Alan V. Sahakian, Northwestern University
Dr. Hun Sun, Drexel University
To
Pam, Bryant, and Mason
and in memory of
John H. K. Khoo
2000 THE INSTITUTE OF ELECTRICAL AND ELECTRONICS
ENGINEERS, INC. 3 Park Avenue, 17
,h
Floor, New York, NY 10016-
Published by John Wiley & Sons, Inc., Hoboken, New Jersey.
No part of this publication may be reproduced, stored in a retrieval system, or
transmitted in any form or by any means, electronic, mechanical, photocopying,
recording, scanning, or otherwise, except as permitted under Section 107 or 108
of the 1976 United States Copyright Act, without either the prior written
permission of the Publisher, or authorization through payment of the appropriate
per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive,
Danvers, MA 01923, 978-750-8400, fax 978-750-4470, or on the web at
www.copyright.com. Requests to the Publisher for permission should be
addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River
Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, e-mail:
[email protected]. For general information on our other products
and services please contact our Customer Care Department within the U.S. at
877-762-2974, outside the U.S. at 317-572-3993 or fax 317-572-4002.
10 9 8 7 6 5 4
For MATLAB product information, please contact:
The MathWorks, Inc., 3 Apple Hill Drive, Natick, MA 01760-2098 USA
Tel: 508-647-7000, Fax: 508-647-7101, E-mail: [email protected]
Web: www.mathworks.com
Library of Congress Cataloging-in-Publication Data
Khoo, Michael C. K.
Physiological control systems : analysis, simulation, and
estimation / Michael C. K. Khoo
p. cm. (IEEE Press series on biomedical engineering)
Includes bibliographical references and index.
ISBN 0-7803-3408-6
1. Physiology-Mathematical models. 2. Biological control
systemsMathematical models, I. Title. II. Series.
QP33.6.M36K48 1999 99-27107
571.7015118dc 21 CIP
Contents
PREFACE xiii
ACKNOWLEDGMENTS xvii
CHAPTER 1 Introduction 1
1.1 Preliminary Considerations 1
1.2 Historical Background 2
1.3 Systems Analysis: Fundamental Concepts 3
1.4 Physiological Control Systems Analysis: A Simple
Example 5
1.5 Differences between Engineering and Physiological
Control Systems 7
1.6 The Science (and Art) of Modeling 9
Bibliography 11
Problems 11
CHAPTER 2 Mathematical Modeling 13
2.1 Generalized System Properties 13
2.2 Models with Combinations of System Elements 16
2.3 Linear Models of Physiological Systems: Two
Examples 19
2.4 Distributed-Parameter versus Lumped-Parameter
Models 21
2.5 Linear Systems and the Superposition Principle 23
2.6 Laplace Transforms and Transfer Functions 24
2.7 The Impulse Response and Linear Convolution 26
vii
v i
Contents
2.8 State-Space Analysis 28
2.9 Computer Analysis and SimulationMATLAB and
SIMULINK 30
Bibliography 35
Problems 35
CHAPTER 3 Static Analysis of Physiological Systems 39
3.1 Introduction 39
3.2 Open-Loop versus Closed-Loop Systems 40
3.3 Determination of the Steady-State Operating Point 42
3.4 Steady-State Analysis Using SIMULINK 45
3.5 Regulation of Cardiac Output 48
3.5.1 The Cardiac Output Curve 49
3.5.2 The Venous Return Curve 51
3.5.3 Closed-Loop Analysis: Heart and Systemic Circulation
Combined 54
3.6 Regulation of Glucose 55
3.7 Chemical Regulation of Ventilation 59
3.7.1 The Gas Exchanger 59
3.7.2 The Respiratory Controller 61
3.7.3 Closed-Loop Analysis: Lungs and Controller Combined 62
Bibliography 65
Problems 66
CHAPTER 4 Time-Domain Analysis of Linear Control Systems 69
4.1 Linearized Respiratory Mechanics: Open-Loop versus
Closed-Loop 69
4.2 Open-Loop and Closed-Loop Transient Responses:
First-Order Model 71
4.2.1 Impulse Response 71
4.2.2 Step Response 73
4.3 Open-Loop versus Closed-Loop Transient Responses:
Second-Order Model 73
4.3.1 Impulse Responses 73
4.3.2 Step Responses 78
4.4 Descriptors of Impulse and Step Responses 81
4.4.1 Generalized Second-Order Dynamics 81
4.4.2 Transient Response Descriptors 84
4.5 Open-Loop versus Closed-Loop Dynamics: Other
Considerations 86
4.5.1 Reduction of the Effects of External Disturbances 86
4.5.2 Reduction of the Effects of Parameter Variations 87
4.5.3 Integral Control 88
4.5.4 Derivative Feedback 90
4.6 Transient Response Analysis Using MATLAB 91
Contents
ix
4.7 SIMULINK Application: Dynamics of Neuromuscular
Reflex Motion 91
4.7.1 A Model of Neuromuscular Reflex Motion 92
4.7.2 SIMULINK Implementation 94
Bibliography 96
Problems 97
CHAPTER 5 Frequency-Domain Analysis of Linear Control Systems 101
5.1 Steady-State Responses to Sinusoidal Inputs 101
5.1.1 Open-Loop Frequency Response 101
5.1.2 Closed-Loop Frequency Response 105
5.1.3 Relationship between Transient and Frequency Responses 105
5.2 Graphical Representations of Frequency Response 108
5.2.1 Bode Plot Representation 108
5.2.2 Nichols Charts 111
5.2.3 Nyquist Plots 113
5.3 Frequency-Domain Analysis Using MATLAB and
SIMULINK 114
5.3.1 Using MATLAB 114
5.3.2 Using SIMULINK 116
5.4 Frequency Response of a Model of Circulatory
Control 119
5.4.1 The Model 119
5.4.2 Simulations with the Model 121
5.4.3 Frequency Response of the Model 121
5.5 Frequency Response of Glucose-Insulin Regulation 124
5.5.1 The Model 124
5.5.2 Simulations with the Model 125
5.5.3 Frequency Responses of the Model 125
Bibliography 128
Problems 129
CHAPTER 6 Stability Analysis: Linear Approaches 131
6.1 Stability and Transient Response 131
6.2 Root Locus Plots 134
6.3 Routh-Hurwitz Stability Criterion 137
6.4 Nyquist Criterion for Stability 139
6.5 Relative Stability 143
6.6 Stability Analysis of the Pupillary Light Reflex 146
6.6.1 Routh-Hurwitz Analysis 147
6.6.2 Nyquist Analysis 149
6.7 Model of Cheyne-Stokes Breathing 151
6.7.1 C0
2
Exchange in the Lungs 151
6.7.2 Transport Delays 153
X
Contents
6.7.3 Controller Responses 153
6.7.4 Loop Transfer Functions 154
6.7.5 Nyquist Stability Analysis Using MATLAB 154
Bibliography 156
Problems 156
CHAPTER 7 Identification of Physiological Control Systems 159
7.1 Basic Problems in Physiological System Analysis 159
7.2 Nonparametric and Parametric Identification
Methods 161
7.2.1 Numerical Deconvolution 162
7.2.2 Least Squares Estimation 163
7.2.3 Estimation Using Correlation Functions 166
7.2.4 Estimation in the Frequency Domain 167
7.2.5 Optimization Techniques 169
7.3 Problems in Parameter Estimation: Identifiability and Input
Design 175
7.3.1 Structural Identifiability 175
7.3.2 Sensitivity Analysis 176
7.3.3 Input Design 179
7.4 Identification of Closed-Loop Systems: "Opening the
Loop" 182
7.4.1 The Starling Heart-Lung Preparation 183
7.4.2 Kao's Cross-Circulation Experiments 184
7.4.3 Artificial Brain Perfusion for Partitioning Central and Peripheral
Chemoreflexes 185
7.4.4 The Voltage Clamp 186
7.4.5 Opening the Pupillary Reflex Loop 186
7.4.6 Read Rebreathing Technique 187
7.5 Identification Under Closed-Loop Conditions: Case Studies 189
7.5.1 Minimal Model of Blood Glucose Regulation 190
7.5.2 Closed-Loop Identification of the Respiratory Control
System 193
Bibliography 200
Problems 201
CHAPTER 8 Optimization in Physiological Control 203
8.1 Optimization in Systems with Negative Feedback 203
8.2 Single-Parameter Optimization: Control of Respiratory
Frequency 206
8.3 Constrained Optimization: Airflow Pattern
Regulation 208
8.3.1 Lagrange Multiplier Method 208
8.3.2 Optimal Control of Airflow Pattern 209
8.4 Constrained Optimization: Control of Aortic Flow
Pulse 212
Contents xi
8.4.1 Calculus of Variations 212
8.4.2 Optimal Left Ventricular Ejection Pattern 213
8.5 Adaptive Control of Physiological Variables 218
8.5.1 General Considerations 218
8.5.2 Adaptive Buffering of Fluctuations in Arterial P
co
219
Bibliography 226
Problems 227
CHAPTER 9 Nonlinear Analysis of Physiological Control Systems 229
9.1 Nonlinear versus Linear Closed-Loop Systems 229
9.2 Phase-Plane Analysis 232
9.2.1 Local Stability: Singular Points 234
9.2.2 Method of Isoclines 236
9.3 Nonlinear Oscillators 240
9.3.1 Limit Cycles 240
9.3.2 The van der Pol Oscillator 240
9.3.3 Modeling Cardiac Dysrhythmias 245
9.4 The Describing Function Method 252
9.4.1 Methodology 252
9.4.2 Application: Periodic Breathing with Apnea 254
9.5 Models of Neuronal Dynamics 257
9.5.1 The Hodgkin-Huxley Model 257
9.5.2 The Bonhoeffer-van der Pol Model 260
Bibliography 267
Problems 267
CHAPTER 10 Complex Dynamics in Physiological Control Systems 271
10.1 Spontaneous Variability 271
10.2 Nonlinear Control Systems with Delayed
Feedback 274
10.2.1 The Logistic Equation 274
10.2.2 Regulation of Neutrophil Density 278
10.2.3 Model of Cardiovascular Variability 281
10.3 Coupled Nonlinear Oscillators: Model of Circadian
Rhythms 290
10.4 Time-Varying Physiological Closed-Loop Systems:
Sleep Apnea Model 295
10.5 Propagation of System Noise in Feedback Loops 301
Bibliography 306
Problems 307
xii
Contents
APPENDIX I Commonly Used Laplace Transform Pairs 309
APPENDIX II List of MATLAB and SIMULINK Programs/
Functions 311
INDEX 315
ABOUT THE AUTHOR 319
Preface
Control mechanisms provide the basis for the maintenance of homeostasis at all levels of
organization in the hierarchy of living systems. As such, one's knowledge of the workings of a
given biological system is incomplete unless one can arrive at some understanding of the
regulatory processes that contribute to its natural operating characteristics. In order to attain
this understanding, a conceptual model of the various interacting processes involved is
necessarybut not sufficient. To determine whether one's model reflects the underlying
reality, one has to make predictions with the model. However, more often than not, the factors
in play are complex and dynamic, and the behavior of the model may depend strongly on the
numerical values of certain key parameters. Under such circumstances, the rigorous frame-
work provided by a quantitative approach becomes indispensable. Indeed, some of the most
notable advances in the physiological sciences over the past several decades have been made
dirough the application of quantitative models. Physiological control modeling also has been
critical, directly or indirectly, for the development of many improved medical diagnostic
techniques and new technological therapeutic innovations in recent times.
Because of its importance, the study of physiological control systems is generally
incorporated, in one form or another, into the typical undergraduate biomedical engineering
curriculum. Some programs offer courses that deal explicitly with physiological control
systems, whereas in others, basic control theory may be incorporated into a course on
quantitative physiology. Numerous high-quality research volumes on this subject have been
published over the years, but there exist only a few books that most instructors would consider
suitable for use as a comprehensive text in an upper-level undergraduate or first-year graduate
course. Milsum's Biological Control Systems Analysis and Milhorn's The Application of
Control Theory to Physiological Systems are two classic examples of possible texts, but these
and the handful of other alternatives were published in the 1960s or early 1970s. The present
book is aimed at filling this void. In addition to the classical methods that were covered in
previous texts, this book also includes more contemporary topics and methodologies that
continue to be employed in bioengineering research today.
xiii
XIV
Preface
The primary goals of this book are to highlight the basic techniques employed in
control theory, systems analysis, and model identification, and to give the biomedical
engineering student an appreciation of how these principles can be applied to better
understand the processes involved in physiological regulation. The assumption made here
is that the book would be used in a one-semester course on physiological control systems or
physiological systems analysis taken by undergraduates in the junior or senior year. The book
and its accompanying programs may also prove to be a useful resource for first-year
biomedical engineering graduate students, as well as interested life science or clinical
researchers who have had little formal training in systems or control theory. Throughout
this book, I have emphasized the physiological applications of control engineering, focusing
in particular on the analysis of feedback regulation. In contrast, the basic concepts and
methods of control theory are introduced with little attention paid to mathematical derivations
or proofs. For this reason, I would recommend the inclusion of a more traditional,
engineering-oriented control theory course as a supplement to the material covered in this
volume.
The book begins with a presentation of some historical perspectives, a discussion of the
differences between technological and physiological control systems, and an introduction to
the basic concepts of systems analysis and mathematical modeling. The subsequent five
chapters cover classical control theory and its application to physiological systems. These
begin in Chapter 2 with a tutorial on linear modeling. Here, we discuss generalized system
properties, model analogs, lumped-parameter versus distributed-parameter models, and the
utility of employing time-domain and frequency-domain descriptions of linear systems. In
Chapter 3, we explore the techniques for steady-state analysis of physiological closed-loop
systems. These problems traditionally have relied on graphical solution, as exemplified by the
classic cardiac output-venous return analyses of Guyton and coworkers. Here, we also
explore a decidedly more "modern" approachthat of employing computer analysis to solve
the problems. Chapter 4 covers the transient response analyses of simple linear open-loop and
closed-loop systems. We discuss the effect on system dynamics of "closing the loop," as well
as changing the type of feedback from proportional to integral or derivative. In Chapter 5, we
present the major methods for representing the frequency response of linear models, and also
discuss the relationship between time-domain and frequency-domain approaches. Chapter 6
deals with the topic of stability, an issue of critical importance to physiological regulation. We
discuss a range of techniques for assessing stability under conditions in which the assumption
of linearity can be made. Chapter 7 addresses the problem of system identification,
particularly in systems that operate under closed-loop conditions. Previous texts on physio-
logical control have paid little attention to this important topic in spite of the fact that every
bioengineering researcher has had to confront this problem at some point or other. In this
chapter, we also discuss the related issues of parameter identifiability, sensitivity to noise, and
input design. In Chapter 8, we move on to the application of "modern" control theory to
physiological systems: these methods are based on the principle of optimization. We end this
chapter with a brief exposition of how adaptive control theory may be applied in practice to
regulate spontaneous fluctuations in a physiological signal. Chapter 9 presents a survey of
some of the more common nonlinear analysis methodologies employed for investigating
physiological systems. We recognize that this limited coverage, due to space constraints, does
not do justice to the many other important nonlinear techniques and applications that have
appeared in the research literature in the past two decades. Nevertheless, we believe we have
included sufficient material to give the student a good "feel" of this area of study. Finally, we
conclude the book in Chapter 10 with an examination of the potential mechanisms that could
Preface
xv
give rise to complex dynamic behavior in physiological control systems. These include
spontaneous variability arising from structural nonlinearity in the systemthe phenomenon
of "chaos," interactions between different control systems, and nonstationarity in the system
parameters. Throughout the book, I have attempted to include models of a wide range of
physiological systems, although I cannot deny that there is somewhat of a bias towards my
own favorite area of interest: cardiopulmonary control.
The ubiquity of personal computers among today's college students and the widespread
use of MATLAB and SIMULINK (The Mathworks, Natick, MA) for systems analysis
and simulation in the vast majority of engineering curricula have presented us with the
opportunity to add a more "hands-on" flavor to the teaching of physiological control systems.
As such, almost all chapters of the book include physiological applications that have
accompanying MATLAB/SIMULINK simulation models. These, along with the computer
exercises that accompany the end of each chapter, should aid the learning process by allowing
the student the opportunity to explore "first-hand" the dynamics underlying the biological
mechanisms being studied. This feature of the book is nonexistent in previous texts on
physiological control systems. However, in incorporating this feature, we make the implicit
assumption that the reader has some basic familiarity with MATLAB and/or SIMULINK. For
the reader who has not used MATLAB or SIMULINK, it is fortunate that there are currently
many "primers" on the subject that can be easily found in any academic bookstore. Appendix
II lists and explains the MATLAB/SIMULINK functions used in the examples presented in
this book, along with the names of the files that contain the model simulation/analysis
programs discussed in the text. Details of how one can obtain these program files through the
Internet are also given in the appendix.
Michael C. K. Khoo
Biomedical Engineering Department
University of Southern California
Acknowledgments
The "birth" of this book took place some two and one-half years ago when Dr. Metin Akay,
the editor in chief of this series of monographs, first encouraged me to submit a proposal to
the IEEE Press. It has been a long, arduous journey since that momentindeed, there were
times when it felt as if the end would never be in sight. I thank Metin for his faith in me and
for his continual support throughout the entire process. I am also thankful to John Griffin and
Linda Matarazzo at IEEE Press for their assistance and cordiality. The anonymous reviewers
who had to suffer through the drafts of various chapters in this book have my deepest
gratitude for pointing out errors and for giving me valuable feedback. The collegial
environment provided by my faculty colleagues, the staff, and students of the Biomedical
Engineering Department at the University of Southern California (USC) has been an essential
ingredient in the development of this book. I would be remiss if I did not mention my eternal
indebtedness to the late Professor Fred Grodins, who was always my role model of what a true
bioengineer should be like. Much of the intellectual stimulation that has led to the writing of
this book is drawn from my own research work in the modeling of neurocardiorespiratory
control, an activity sponsored by the National Institutes of Health through the Biomedical
Simulations Resource (BMSR) at USC. I thank my colleagues and codirectors of the BMSR,
Professors Vasilis Marmarelis and David D'Argenio, for their encouragement and support
over the years. Finally, this endeavor certainly would not have been possible if I had not been
blessed with the environment of love, joy, and understanding created so generously by those
closest to me. My deepest feelings of appreciation go deservedly to my wife, Pam, and my
sons, Bryant and Mason. It is to them that I dedicate this book.
Michael C. K. Khoo
Biomedical Engineering Department
University of Southern California
xvii