1066-033X/21©2021IEEE
» BOOKSHELF
80 IEEE CONTROL SYSTEMS » FEBRUARY 2021
ANALYTICAL DESIGN OF
PID CONTROLLERS
by IVÁN D. DÍAZ-RODRÍGUEZ,
SANGJIN HAN, and SHANKAR P.
BHATTACHARYYA
Reviewed by Lee H. Keel
T
he proportional-integral-
derivative (PID) controller
dominates the control in-
dustry and, by some estimates,
accounts for more than 95% of
the controllers in use world-
wide [1], [2]. Its working is
based on the following simple
fact: if a stable dynamic system is subject to constant refer-
ence and disturbance inputs, all signals in the system tend
asymptotically to constant values and the input to each
integrator tends to zero. This applies to continuous-time;
discrete-time; linear; nonlinear; single-input, single-output
(SISO); and multivariable, multiple-input, multiple-output
systems with and without time delays. This immediately
suggests that tracking and disturbance rejection of con-
stant inputs can be accomplished by generating the track-
ing error and using it to drive integrators that generate
the control signal. This architecture is inherently robust
relative to state feedback, observer-based, high-order op-
timal systems [3].
This monograph is focused exclusively on the PID con-
troller. It is timely and contributes substantially to the art
and science of control system engineering.
» Timely, because PID controllers are increasingly
being used in dynamically changing environments,
such as those encountered in driverless cars, unmanned
aerial vehicles, and distributed robotics. In these
applications, there is an urgent need for design meth-
ods and algorithms that can quickly and accurately
update controller gains.
» Art, because the book proposes a multiobjective
approach to design, allowing the designer to cre-
atively blend various performance measures using
classical and modern approaches.
» Science, because rigorous analytical approaches to
achieve such designs are also presented.
This monograph is the third in a series (the others are
[4] and [5]) focused on the development of PID controller
design theory based on the computation of the stabilizing set.
CONTENTS
The book is organized into three parts. Part I lays the foun-
dation for the rest of the book. It describes methods for the
computation of the complete set of PID controllers stabi-
lizing a given SISO linear time-invariant plant. The plant
description can be in the form of a transfer function, or it may
be given as frequency-response data for a continuous-time
[6] or discrete-time system [7]. Clearly, every design must
reside in this stabilizing set. It is shown that the stabilizing
set can be computed by solving a set of linear programming
problems with a scalar sweeping parameter varied over a
specifiable range. The linear programming formulation
results from the novel application of a root counting for-
mula that generalizes the classical Hermite–Bieler theorem.
This applies to continuous-time systems with time delay
and discrete-time systems. In the latter case, the frequency
response of the system is represented using Tchebyshev
polynomials, which simplify the computations. An interest-
ing by-product of these results is the computation of all sta-
bilizing PID controllers for the Ziegler–Nichols plant (that
is, first order with time delay), generalizing this classical
1942 result [8]. Another interesting result is the computation
of all PID controllers that shift all the closed-loop poles to
the left of a line at , s v =- resulting in the determination
of the maximum achievable v for the given plant [9]. This
number is inversely proportional to the settling time of the
step response.
Using the stabilizing set, Part II proceeds to search for
achievable performance for the closed-loop system under
I
EEE Control Systems welcomes suggestions for books to
be reviewed in this column. Please contact either Scott R.
Ploen, Hong Yue, or Thomas Schön, associate editors for
book reviews.
Springer, 2019,
ISBN, 978-3-030—18227-4,
314 pages, US$149.99.
Digital Object Identifier 10.1109/MCS.2020.3032803
Date of current version: 18 January 2021