Vous êtes sur la page 1sur 1

AuWmatica,Vol.23, No. 3, pp.

413-417, 1987

PergamonJournalsLtd. Printedin Great Britain.


InternationalFederationof AutomaticControl

Book Reviews

Feedback Control of Dynamic Systems*


Gone F. Franklin, J. David Powell and Abbas Emami-Naeini
Reviewer: STEPHEN KAHNE

margin or transient response to standard input signals. The


problem many practising control engineers have had with state
feedback design is lack of intuition about the effects of changes
in feedback gains. Our authors have attempted to help with the
intuition hurdle by emphasizing the role of state feedback gains
on separate placement of the estimator and controller poles. Of
course, this requires a discussion of estimators, poles and zeros
in the state variable description of the system, appropriate
canonical forms and other such matters which are not part of
the "usual" transfer-function system description. These are
introduced in the context of understanding system behaviour
rather than as mathematical structures of some independent,
intrinsic interest. A number of design examples are presented
which are interpreted in terms of the now familiar PID controller
family. This is not an easy chapter but, well taught, the concepts
should become part of the student's control design repertoire
with practice.
Is there anything wrong with the book? Each of us has a list
of items we would prefer to see in a course at this level and
each prefers different emphasis. I personally find Mason's signal
flow graph a useful analysis tool. The authors hardly mention
it except in an appendix. The preparation students have for
later studies of multivariable systems would be enhanced by a
more detailed treatment of the inverse Nyquist plot. Discussions
of non-linear systems include five pages on Lyapunov stability
and one page on describing functions. There could have been
more emphasis on system type and more use of it in the design
examples, but these are minor points. There are quite a number
of typographical errors, most of which were not corrected in
the second or third printing. These are only minor annoyances
but can be disruptive to the student.
The brief chapter on digital control serves well to convince
students that they should study this important aspect of fecdback
control systems but is not a particularly useful teaching chapter
since so little is said about so much. Readers are encouraged to
continue their studies of this topic, possible with the authors'
older text (Franklin and Powell, 1981).
The book is a good pedagogical contribution to the control
systems design field. The integration of state feedback with
the classical approaches is largely successful and should be
particularly useful in the next decade as computational tools
become more widely available for control engineers to use in
the work-place. For those inclined to specialize in CACSD
(computer-aided control system design), this book could well
serve as an introduction to the field. Following this would be
digital (Franklin and Powell, 198 I) and multivariable (Patel and
Munro, 1982) texts and such conferences as the IEEE Control
Systems Society series on CACSD.

Oregon Graduate Center, Beaverton, Oregon, U.S.A.


DESIGN IS partly art and partly science. In most books on control
system design the science is stressed to the exclusion of the art.
Analysis is a key feature of the designer's work but the intuition
and subjective insight into design problems are often overlooked.
In the Franklin book the art of design is evident. There are few
control books which emphasize the design of these systems to
the extent that this one does and for that the authors deserve
credit. It is consistent with their intent as described in the preface
of this volume (in its third printing in its first year of publication)
and is pretty well carried off in the body of the text. Certainly
the authors are not the first to have attempted to highlight the
intuitive aspects of design - - the 1963 Horowitz book comes to
mind - - but this is perhaps the most balanced comprehensive
treatment of design fundamentals for the college senior.
The book is divided into three main segments. The first three
chapters review analysis techniques. The next three describe
three classes of synthesis tools-- root locus, frequency response
and state feedback. The next chapter contains four rather
detailed ease studies of feedback control designs. Finally, there
is a glimpse of digital computer realization of analogue controllers. A number of mathematical appendices end the book. This
organization is quite effective and consistent with the authors'
stated objective of producing a balanced, design oriented treatment of material, much of which is now accepted as part of the
control engineer's tool kit. Thus, there is no pretense to
originality of content but rather to organization and emphasis
of the subject matter. The use of computers for carrying out
calculations and plotting curves is taken for granted and
there are references to the source of appropriate programs.
Unfortunately, there is still not much standardization of these
programs so the reader is on his own to set up his design tool
collection.
Some books are quite culture-specific. Not this one. The
selection of problems at the end of each chapter is good and
not particularly oriented toward situations characteristic of one
country or another. There are also enough problems that several
years of homework could be assigned before one would have
to reuse a problem. Apparently there is a solution manual for
the homework problems but it was not available from the
publisher at the time this review was prepared.
One particularly attractive feature of the book is the inclusion
of numerous explanatory comments which show why a certain
result should or should not conform to the reader's intuition.
As noted earlier, the use of intuition as well as appropriate
analytical tools is the key to good design. Incisive comments in
the classroom by a knowledgeable teacher are indicative of good
teaching. The authors have gone farther in that direction than
is common in many textbooks, at least in our field.
Let's pursue this issue of intuition. A key value of either root
locus or frequency response approaches to feedback control
design is that the interaction of the designer with the graphical
presentation of the current results encourages experimentation
with feedback parameters and even with feedback structures.
Classically this intuition was obtained by looking at the s-plane
location of closed-loop poles or the magnitude and phase vs
frequency plots of the open-loop components and translating
this into closed loop performance through gain and phase

References
Franklin, G. and J. Powell (1981). Digital Control of Dynamic
Systems. Addison-Wesley, Reading, MA.
Horowitz, I. (1963). Synthesis of Feedback Systems. Academic
Press, New York.
Patel, R. and N. Munro (1982). Multivariable System Theory
and Design. Pergamon Press, Oxford.

About the reviewer


Professor S. Kahne is with the Oregon Graduate Center
outside Portland, Oregon, U.S.A. Previously he has been at the
Polytechnic University in New York, Case Western Reserve
University in Cleveland and the University of Minnesota in
Minneapolis. In IFAC he has served in numerous capacities
and presently is Chairman of the IFAC Policy Committee.

*Feedback Control of Dynamic Systems by G. F. Franklin,


J. D. Poweil and A. Emami-Naeini. Addison-Wesley, Reading,
MA (1986). U.S. $41.95.
413

Vous aimerez peut-être aussi