Академический Документы
Профессиональный Документы
Культура Документы
net/publication/260586232
CITATIONS READS
31 925
3 authors:
A. Galip Ulsoy
University of Michigan
351 PUBLICATIONS 9,969 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by A. Galip Ulsoy on 13 March 2015.
A new design method is presented for PI controllers of first-order plants in the presence of time-delays. In general,
time-delays can limit and degrade the achievable performance of the controlled system, and even induce instability.
Thus, the PI gains should be selected carefully considering such effects of time-delays. Unlike existing methods,
the design method presented is based on solutions to delay differential equations, which are derived in terms of the
Lambert W function. PI controllers for first-order plants with time-delays are designed by obtaining the rightmost (i.e.,
dominant) eigenvalues in the infinite eigenspectrum of time-delay systems, and assigning them to desired positions
in the complex plane. The process is possible due to a novel property of the Lambert W function. The controllers
designed using the presented method can improve system performance and successfully stabilize an unstable plant.
Also, sensitivity analysis of the rightmost eigenvalues is conducted and compares favorably to a prediction-based
method for eigenvalue assignment. Extension to design of PID controllers is discussed with examples.
Index Terms
I. INTRODUCTION
A variety of methods have been developed to design PI/PID controllers to meet specifications for time-delay
systems, such as time-domain specifications and/or disturbance rejection, and are available in the literature [30].
Such methods use experimental tuning, pole-placement in the complex plane, plant response in the frequency
domains, and/or graphical methods. However, when time-delays exist between the application of inputs and their
resulting effects, stability analysis is not as straightforward as for non-delayed systems and. Thus, design of feedback
control becomes difficult [27]. The principal difficulty is due to the fact that the time-delay terms always lead to
an infinite eigenspectrum.
The stability analysis for time-delay systems can be conducted using bifurcation methods (see, e.g., [29] and
the references therein). Based on the analysis, the gains in controllers are chosen based on the stability regions in
Sun Yi (corresponding author) is with the Department of Mechanical Engineering, NC A&T State University, Greensboro, NC 27411
syi@ncat.edu
Patrick W. Nelson is with the Center for Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor, MI 48109.
A. Galip Ulsoy is with the Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI 48109.
2
the proportional and integral control gain space [30], often combined with the Nyquist method [14]. Alternatively,
stabilization problems have been addressed by using an observer based controller with a discrete model [26],
or H∞ feedback control with a discrete event based model to improve robustness to disturbances and modeling
uncertainties [18], the Smith predictor [21] and its adaptive version [46], nonlinear adaptive controllers [35], and Padé
approximation approaches [10]. PI controllers with the Smith predictor have been shown to successfully improve
performance in simulations and experiments [47]. Pole-placement for time-delay systems, which is not feasible using
traditional control methods, can be achieved by employing the Smith predictor. Such a prediction-based method
was compared to an experimental tuning method, the classical Ziegler-Nichols method in [21]. Alternatively, the
act-and-wait control strategy can be used for pole-placement [7], [11], [13]. The method uses a periodic controller
where, for example, its period, T , is double that of the time-delay, h (i.e, T = 2 × h). Then, during the first delay
interval h, the control input is zero, and during the second h, the control gain is D (i.e., u(t) = Dx(t)). Then, for
the second half of the period, the governing equation can be written as an ODE, and the pole-placement problem can
be formulated as a finite dimensional one. The control gain is chosen such that the transition matrix has eigenvalues
of which absolute values are smaller than one (or as small as possible). For a thorough discussion on the method,
refer to [12]. Also, a numerical stabilization method was developed using sensitivities of eigenvalues with respect
to changes in the feedback gain in [19] (also, refer to [43] for a numerical comparison by the authors).
In this paper, PI feedback controllers for linear time-invariant (LTI) plants having a time-delay are designed
through pole-placement. The pole-placement is achieved by using the Lambert W function-based approach without
the use of a predictor. The rightmost (i.e., dominant) eigenvalues in the infinite eigenspectrum can be identified and
assigned using the Lambert W function [45]. This approach is very similar to pole placement design of PI controllers
for systems without delay, and does not require the use of a predictor. Through such assignment, the PI gains are
calculated. The designed controllers can improve performance as well as successfully stabilize unstable systems.
The proposed PI design method is compared with other existing methods available in the literature for illustrative
examples. Also, sensitivity of the rightmost eigenvalue with respect to system parameters (including delays) is
studied analytically, and via this analysis the robustness of the controller is shown to compare favorably to Smith
predictor-based controllers. Although the Smith predictor is a long standing control technique, it is still one of the
most widely used and studied control schemes for time-delay systems [47]. Because it, and other similar predictor
based methods, are the most effective tool for pole-placement for time-delay systems, the Smith predictor is used
here as a benchmark for performance and sensitivity comparisons. However, the proposed method is also compared
to other PI control design approaches in subsequent sections. Unlike systems of ODEs, PID design through ‘direct’
eigenvalue assignment has not been feasible for systems of DDEs. Using the method presented, one can extend the
eigenvalue assignment approach for PI/PID tuning to time-delay systems. Because many properties of a closed-loop
system depend on the locations of its poles, pole-placement is one of the mainstream methods in control system
design [2]. Thus, the method presented here provides a direct way to extend well-developed control techniques for
ODEs to DDEs.
+ + KI u
r GS ( s ) KP e-sh GP(s) y
y - v - s
1 e Gˆ (s)
shˆ
P
Smith Predictor
Gc
+ KI u
r GS ( s ) KP GP(s) e-sh y
y - s
(b) PI controller with the Smith predictor when the model is EXACTLY same as the plant
(i.e.,Gˆ P ( s ) GP ( s ) ).
u
+ K
r GL ( s ) KP I e-sh GP(s) y
y - s
(c) PI controller using the Lambert W function-based approach without the Smith predictor
Fig. 1. Block diagrams for a closed-loop system with the Smith Predictor-based control (a), its equivalent system when the response of the
plant is predicted perfectly (b) [32], and the closed-loop system with the Lambert W function-based control (c).
The Smith predictor in Fig. 1-(a) results in a delayed response of a delay-free system by moving the time-delay
outside the feedback loop only when the model in the Smith predictor, ĜP (s) and ĥ, is assumed to be exactly the
same as the plant, GP (s) and h. Then, the controller GS (s) in (2) can be designed considering only the delay-free
plant, GP (s) (see Fig. 1-(b)). This is the main advantage of the Smith predictor control.
For example, in order to meet given time domain specifications, the desired eigenvalues can be chosen from the
desired natural frequency, ωn , and the desired damping ratio, ζ as
λd = −ωn ζ ± ωn 1 − ζ 2 i = −σ ± ωd i (3)
In this subsection, as an alternative to the Smith predictor, a design approach for the PI controller is developed
via rightmost eigenvalue assignment using the Lambert W function. The open-loop transfer function with a PI
controller is
KM KI y
Gopen = e−sh KP + = (6)
τM s + 1 s e
where e = −y + r. Then, the closed-loop system as in Fig. 1-(c) in the time-domain becomes
1 KM KP KM KI
ÿ = − ẏ − ẏ(t − h) − y(t − h)
τM τM τM (7)
+ {KM KP s + KM KI } r(t − h)
By defining x1 ≡ y, x2 ≡ ẏ, the equation can be re-written as
ẋ1 = x2
1 KM KP KM K I
ẋ2 = − x2 − x2 (t − h) − x1 (t − h) (8)
τM τM τM
+ {KM KP s + KM KI } r(t − h)
Then, we obtain the closed-loop system in state-space form, where we ignore the reference input to focus on
stability, as follows:
ẋ(t) =
⎡ ⎤ ⎡ ⎤
0 1 0 0
⎣ 1 ⎦ x(t) − ⎣ KM KI KM KP ⎦ x(t − h)
(9)
0 −
τ τM τM
M
≡A ≡Ad
From the roots of the characteristic equation of the system (9), the eigenvalues of the system are obtained.
However, due to the time-delay, e−sh , the system is infinite-dimensional and, thus, there exists an infinite number
of eigenvalues. The principal difficulty in analyzing and controlling systems with time-delays arises from this
transcendental character, and the determination of this eigenspectrum typically requires numerical, approximate,
or other approaches [27]. Obtaining and controlling the entire infinite eigenspectrum is not as straightforward as
for systems of ODEs. Instead, for DDEs, like Eq. (9), it is desired to locate the dominant eigenvalues, which are
rightmost in the complex plane, and to assign them to desired positions. The Lambert W function-based approach
is an efficient tool for doing this and in the subsequent section the approach is applied to the system (9) to design
the PI controller and compared with the Smith predictor. With the coefficient matrices, A and Ad defined in Eq.
(9), the solution matrix, S0 is computed as
1
S0 = W0 (Ad hQ0 ) + A (10)
h
where the unknown matrix Q0 is obtained by solving
The equation in Eq. (11) can be solved numerically to obtain the matrix Q0 for the principal branch (k = 0)
using nonlinear solvers (e.g., fsolve in MATLAB). Then, substitute the matrix Q0 into Eq. (10) to obtain S0 and its
eigenvalues. Then one can set the desired locations for eigenvalues equal to those of S0 , i.e., λi (S0 ) = λi,desired
for i = 1, · · · , n, where, λi (S0 ) is ith eigenvalue of the matrix S0 . The gains, KP and KI are then obtained by
solving the equation numerically. For a detailed explanation of the Lambert W function-based approach, including
a method for eigenvalue assignment, refer to [45]. Also, open-source software for the Lambert W function method
based on MATLAB is available at [5]. As mentioned in [43], depending on the structure or parameters of a given
system, there exists a limitation on the rightmost eigenvalues and they cannot be arbitrarily assigned. If non-feasible
values of λi,desired are selected, then the above approach does not yield a solution. To resolve the problem, one
may try again with fewer desired eigenvalues, or different values of the desired rightmost eigenvalues. Then, the
solution is obtained numerically for a variety of initial conditions by an iterative trial and error procedure.
Desired positions for the rightmost eigenvalues, λi,desired , can be chosen from the desired natural frequency and
damping ratio using the relation in Eq. (3). In assignment of eigenvalues, the principal branch of the Lambert W
function is used to find the rightmost eigenvalues. For 1st order scalar DDEs, it has been proven that the rightmost
eigenvalues are always obtained by using the principal branch (e.g., see [28]). For general systems of DDEs, such
a proof is not available. It has been observed that if the coefficient, Ad , does not have repeated zero eigenvalues,
the rightmost eigenvalues are obtained with the principal branch [40]. If Ad has repeated zero eigenvalues, they are
obtained by using the branches k = 0 and k = −1 [42].
In this section, the Lambert W function-based approach is applied to two different cases: an open-loop unstable
plant and an open-loop stable plant.
Response, y
3
Smith predictor
Lambert W function
0
0 5 10 15 20
Time, t
Fig. 2. Simulated responses of the unstable system (12) controlled with the Smith predictor (dashed line) and the Lambert W function-based
approach (solid line) using Simulink. Because the Smith predictor is based on unstable pole-zero cancelation, it fails to stabilize the system.
Thus, one can attempt to place a few dominant poles [37]. The Smith predictor is a well-known method for pole-
placement. The gains of the PI controller, using the approach in Subsection II-A, are found to be K̄I = 3.1250
and K̄P = 1.4500 when the desired natural frequency is ωn = 2.5 and the desired damping ratio is ζ = 0.5. With
those gains one can assign the eigenvalues of the system to be −1.2500 ± 2.1651i, which are stable, theoretically.
However, due to initial conditions, disturbance, and errors in simulation, this control leads to instability as seen in
Fig. 2 (dashed). If the system (1) is unstable, the characteristic equation of the closed-loop system with the Smith
predictor in Fig. 1-(a) retains the unstable pole of the open-loop system. Therefore, the Smith predictor cannot
stabilize the system [6]. Figure 2 shows responses simulated using Simulink. Even though there is no disturbance
or no initial condition mismatch, due to errors in numerical integration, the Smith predictor cannot successfully
stabilize the unstable plant (12). Using the act-and-wait control method, pole-placement can also be achieved [12].
For example, for the time constant (τ = 1/1.25), the corresponding PI gains are KI = 0.8953 and KP = 2.0527.
Those two methods (the Smith predictor and the act-and-wait control) make the number of eigenvalues finite and
assign them. On the other hand, the Lambert W function-based approach in Subsection II-B does not reduce the
number of eigenvalues. Instead, it assigns the rightmost eigenvalues of the infinite spectrum to the desired positions.
For example, the resulting gains for the rightmost eigenvalues, −1.2500 ± 2.1651i (ωn = 2.5 and ζ = 0.5) are
KI = 1.4309 and KP = 1.2232. Table I shows the PI gains corresponding to other values of ωn . In designing
PI control for systems of ODEs, an increase in ωn induces a decrease in rise time while not changing overshoot.
As seen in Fig. 3, by adjusting the values of ωn , one can tune the gains to meet time-domain specifications in a
similar way to ODEs.
Because the Lambert W function-based approach does not require pole-zero cancellation, the designed controller
safely stabilizes the system (see Fig. 2). Another prediction-based approach, finite spectrum assignment (FSA), also
has unsolved problems regarding integral approximation and, thus, can fail to stabilize unstable systems [27]. It has
1.8
1.6
1.4
1.2
Response, y
1
0.8
increasing ω
n
0.6 (1.0, 1.5 and 2.0)
0.4
0.2
0
0 1 2 3 4 5 6 7 8
Time, t
Fig. 3. Simulated responses of the unstable system (12) controlled with PI controllers designed by using the Lambert W function-based
approach. Increase in ωn from 1 to 1.5 to 2.0, induces a decrease in the rise time of the transient response.
TABLE I
G AINS , KI AND KP , OF PI CONTROLLER : OBTAINED BY USING THE L AMBERT W FUNCTION APPROACH VIA RIGHTMOST EIGENVALUE
ASSIGNMENT FOR THE UNSTABLE PLANT IN E Q . (12)
TABLE II
G AINS , KI AND KP , OF PI CONTROLLER : OBTAINED BY USING THE L AMBERT W FUNCTION APPROACH VIA RIGHTMOST EIGENVALUE
ASSIGNMENT
ωn ζ K̄I K̄P KI KP
also been shown in [43] that the Padé approximation does not guarantee stability of the controlled system due to
inherent inaccuracy.
150
100
50
−1.2500 ± 2.1651i
Imag(λ)
0
−50
−100
−150
Fig. 4. Eigenspectrum of the closed-loop system in Fig. 1-(c). By using the Lambert W function-based approach in the subsection II-B, the
rightmost eigenvalues of the infinite eigenspectrum are assigned to the desired positions, −1.2500 ± 2.1651i.
When the time constant, τM , is 0.5, and the time delay, h, is 0.2 with KM = 1 for the plant in Eq. (1), the
gains for PI control, for several desired natural frequencies and damping ratios, are obtained by using the Smith
predictor-based approach and the Lambert W function-based approach, and given in Table II. The Smith predictor
moves the time-delay outside the feedback loop as seen in Fig. 1-(b) and makes the number of poles finite (in this
case 2) by canceling the other (infinite number of) eigenvalues. On the other hand, the Lambert W function-based
approach, without the cancellation, obtains the rightmost eigenvalues and assigns them to the desired positions.
For example, when the desired natural frequency ωn = 2.5 and the desired damping ratio is ζ = 0.5 (thus, the
desired positions for the rightmost eigenvalues are −1.2500±2.1651i), the eigenspectrum of the closed-loop system
is depicted in Fig. 4. The rightmost eigenvalues are placed exactly on the desired positions, which are obtained
from Eq. (3), and others are to the left of them. The response of the closed systems by using the Smith predictor
(Fig. 1-(a)) and the Lambert W function-based approach are shown in Fig. 5. Although the systems in Fig. 1-(c)
have an infinite number of eigenvalues, by assigning the rightmost (thus, dominant) eigenvalues to the desired
positions, one can meet time-domain specifications [44] assuming that the other eigenvalues that are not controlled
are sufficiently far from the rightmost eigenvalues (e.g., Fig. 4). However, if such subdominant eigenvalues get
closer to the dominant ones (e.g., as time-delays increase [16]), it is more difficult to improve transient responses
via the presented eigenvalue assignment method.
As for robustness, if the system (1) is open-loop stable, the Smith predictor works as desired and does not have
apparent mismatch problems unlike the example in Subsection III-A. However, as shown in Fig. 1, the Lambert W
function-based controller has a form simpler than the Smith predictor-based one. Also, for the Smith predictor-based
control in Fig. 1-(a), the feedback signal for the controller, GS (s), uses y · esh , which is the output, y, predicted one
time-delay ahead. Therefore, it is reasonable to consider sensitivity with respect to mismatch of parameters, and
1.2
Response, y
0.8
0.6
Smith Predictor
0.4
Lambert W Function
0.2
Desired (Non−delay)
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Time, t
Fig. 5. Responses of systems controlled by using the Smith predictor (Fig. 1-(a)) and the Lambert W function (Fig. 1-(c)) when the desired
rightmost eigenvalues are −1.2500 ± 2.1651i for τM =0.5, h=0.2, and KM = 1 in Eq. (1).
compare the robustness of the two different control methods. Sensitivity is discussed in the next two subsequent
subsections. In the literature, it has been shown that estimation of time-delay in continuous LTI systems is more
difficult than other system parameters [34]. Also, it is well known that control using the Smith predictor is sensitive
with respect to delay mismatches [20]. Thus, the sensitivity analysis is focused on the time-delay. However, the
analysis for other parameters can be conducted in a similar way and those results are also discussed.
In this subsection the sensitivity of the closed-loop eigenvalues to small variation in the time-delay, h, is considered
for the Smith predictor-based controller.
The transfer function of the closed-loop system with the Smith predictor in Fig. 1-(a) can be represented as
y GS GP e−sh
=
(13)
r 1 + 1 − e−sĥ GS ĜP + GS GP e−sh
Assuming all parameters except the time delay are well-matched, i.e., ĜP = GP , the denominator will be
DE(s) ≡ 1 + 1 − e−sĥ GS ĜP + GS GP e−sh
= 1 + GS GP − GS GP e−sĥ + GS GP e−sh (14)
= 1 + GS GP − GS GP e−sh (e−sδ − 1)
TABLE III
S ENSITIVITY COMPARISON WITH RESPECT TO h: THE L AMBERT W FUNCTION - BASED APPROACH SHOWS SMALLER VALUES OF THE
SENSITIVITY AND IMPROVEMENT IN ROBUSTNESS
∂λ
Rightmost Sensitivity, ∂h
Improvement (%)
ωn ζ
Re(S1)−Re(S2)
eigenvalue, λ Smith predictor (S1) Lambert W function (S2) Re(S1)
× 100
From the state equation (9) of the closed-loop system in 1-(c), the sensitivity of the eigenvalue with respect to
the time-delay, h, is obtained in a way similar to the previous derivation. The characteristic equation of the system
(9) is
1 KM KP −sh KM KI −sh
s2 + s+ se + e =0 (20)
τM τM τM
In a similar way to the previous section, by differentiating both sides the sensitivity of the eigenvalues with respect
to change in the time-delay, h, is given by
∂s
=
∂h (21)
s2 KM KP e−sh + sKM KI e−sh
2sτM + 1 + KM KP e−sh − (KM KP s + KM KI )e−sh h
The numerical values of the sensitivity with respect to h for the Lambert W function-based approach and the Smith
predictor are compared in Table III. As seen in Table III, for several rightmost eigenvalues, which are arbitrarily
chosen, the Lambert W function-based approach shows smaller values of the sensitivity and, thus, improvement in
robustness. For comparison, ‘improvement’ is calculated as the ratio of decrease in real parts of the sensitivities
to the real part of the sensitivity of the Smith predictor (i.e., (Re(S1) − Re(S2))/Re(S1) × 100). Because in the
Smith predictor control the feedback signal for the controller (GS (s) in Fig. 1-(a)) uses the “predicted” output, y,
(e.g., yesh ) by canceling signals [47], it is sensitive with respect to infinitesimal delay mismatches [20]. On the
other hand, for the Lambert W function-based approach, the feedback signal for the controller (GL (s) in Fig. 1-(c))
is not predicted and does not require any cancelation. This may result in improvement in robustness as seen in
Table III.
For other parameters in Eq. (1), sensitivity analysis can be conducted in a way similar to the time-delay, h. For
example, for KM and τM , the obtained sensitivities are summarized in Tables IV. The sensitivity with respect to
KM for the Lambert W function also has smaller values than the Smith predictor. Thus, the Lambert W function-
based approach enhances robustness. However, this does not mean the presented approach always leads to less
sensitive controllers. In Table IV, for the sensitivity with respect to τM , the results are mixed. Thus, use of the
Lambert W function does not always reduce sensitivity of the rightmost eigenvalues. Therefore, for stable first-order
plants, after comparing sensitivity with respect to parameters (especially, ones having larger variance, for example,
due to difficulty in estimation) one can choose more suitable method to design PI controllers that is more robust
against variance in system parameters.
The matrix Lambert W function can be used for general systems of retarded DDEs. Thus, the presented method
for PI control can be extended to more general cases, such as higher order plants and/or PID design.
Consider the PID controller
KI
GL (s) = Kp + + KD s (22)
s
and the second-order plant with a pure time-delay:
KM
G(s) = e−sh (23)
(τm s + 1)(s + a)
TABLE IV
C OMPARISON OF THE SENSITIVITY: W ITH RESPECT TO KM , THE L AMBERT W FUNCTION - BASED APPROACH SHOWS SMALLER VALUES OF
THE SENSITIVITY AND AN IMPROVEMENT IN ROBUSTNESS . O N THE OTHER HAND , WITH RESPECT TO τM : UNLIKE h AND KM , THE
SENSITIVITY SHOWS MIXED RESULTS .
∂λ ∂λ
Sensitivity, ∂KM
Improvement Sensitivity, ∂τM
Improvement
ωn ζ
Smith predictor Lambert W function (%) Smith predictor Lambert W function (%)
1.0 0.3 0.8797 ± 0.6350i 0.5495 ± 0.6061i 37.53 0.4525 ∓ 1.0170i 0.4930 ∓ 0.6794i -8.96
1.1 0.5 0.6824 ± 0.8857i 0.3845 ± 0.7864i 43.66 1.0714 ∓ 0.9286i 0.9675 ∓ 0.5443i 9.70
1.5 0.5 0.5823 ± 1.0599i 0.2354 ± 0.9663i 59.56 1.4258 ∓ 1.4201i 1.4364 ∓ 0.8225i -0.74
2.5 0.5 0.4085 ± 1.6487i −0.2070 ± 1.6364i 49.33 2.1361 ∓ 3.0292i 2.9950 ∓ 1.9965i -40.21
The system of DDEs representing the closed-loop system is given in matrix-vector form as:
⎡ ⎤ ⎡ ⎤
0 1 0 0 0 0
⎢ ⎥ ⎢ ⎥
⎢ ⎥ ⎢ ⎥
ẋ(t) = ⎢ 0 0 1 ⎥ x(t) + ⎢ 0 0 0 ⎥ x(t − h) (24)
⎣ ⎦ ⎣ ⎦
0 − τam − (τmτa+1)
m
− KM KI
τm − KM KP
τm − KM KD
τm
Assuming the desired positions of the two rightmost eigenvalues are s = −1 ± 1.5i, through the design method
using the Lambert W function, the resulting gains are KP = 2.0248, KI = 1.9061, and KD = 0.3769 for τm = 0.5,
KM = 1, a = 1, and h = 0.2. Fig. 6 shows eigenvalues of the system in Eq. (24). The eigenvalues obtained by using
the principal branch (k = 0) are rightmost. By adjusting the gains in Eq. (22), the eigenvalues are assigned to the
desired positions. For the 3-by-3 system of DDEs in Eq. (24), the solution matrix in Eq. (10) has three eigenvalues.
Thus, three rightmost eigenvalues can simultaneously be assigned. For example, the desired three eigenvalues are
s = −2 ± i, −1.5, the corresponding gains are KP = 2.8128, KI = 2.2009, and KD = 0.8353 (see Fig. 7). The
simulation results using Simulink are shown in Fig. 8 for the two PID gain sets obtained above. The solid line, of
which eigenvalues are in Fig. 7, shows a shorter rise time and a smaller overshoot than the dashed line, of which
eigenvalues are in Fig. 6.
In the examples, two and three of the rightmost eigenvalues are assigned and the gains can be obtained successfully.
However, as the dimension of systems and, thus, number of unknowns in Eq. (11) increase, computational load
also increases correspondingly, in general. For example, solving Eq. (11) for Qk takes 0.2 - 0.3 seconds for 2 by
2 examples. But it takes 3 - 4 seconds for 3 by 3 examples in this paper using a PC with 2.40GHz CPU and
4GB RAM. The nonlinear solver fsolve was used and computing time was measured using cputime in MATLAB.
Also, increased number of control gains adds more computing time. Thus, the presented method is not practically
feasible for arbitrary high dimensional systems. Besides, the relationship between the number of control gains and
the number of the eigenvalues that can be assigned should be studied further in connection with the controllability
of systems of DDEs [41].
300 3
200 2
Imag(λ)
0 0
−100 −1
−200 −2
−300 −3
−30 −25 −20 −15 −10 −5 0 −2.5 −2 −1.5 Real(λ) −1 −0.5 0
Real(λ)
Fig. 6. Eigenvalues of the system in Eq. (24): the eigenvalues obtained by using the principal branch (k = 0) are rightmost. By adjusting the
gains in Eq. (22), the eigenvalues are assigned to the desired position (s = −1 ± 1.5i).
300 2
1.5
200
The set of the rightmost eigenvalues 1
100 by the principal branch (k=0)
0.5
Imag(λ)
Imag(λ)
0 0
−0.5
−100
−1
−200
−1.5
−300 −2
−25 −20 −15 −10 −5 0 −2.5 −2 −1.5 −1 −0.5
Real(λ) Real(λ)
Fig. 7. Eigenvalues of the system in Eq. (24): the eigenvalues obtained by using the principal branch (k = 0) are rightmost. By adjusting the
gains in Eq. (22), the eigenvalues are assigned to the desired position (s = −2 ± i and s = −1.5).
PI controllers are the most commonly used controllers for first order time-delay systems. In this paper, we
have studied a new approach to the design of PI controllers for time-delay systems. Choosing the gains in the PI
controllers corresponding to desired system performance is essential to achieve the control goals. The presented
1.4
0.8
0.6
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Time, t
Fig. 8. Responses of the system in Eq. (23) with the PID controller in Eq. (22).
approach based on the Lambert W function enables one to choose the gains by assigning the rightmost (i.e.,
dominant) eigenvalues to the desired positions, which are obtained from the desired natural frequency and damping
ratio. Unlike prediction-based methods, the approach is not dependent on pole-zero cancellation and, thus, unstable
systems with time delays can be successfully stabilized as shown in Section III. Also, because prediction of the
plant is not required, the obtained controller has a simpler form as shown in Fig. 1 and is shown to be more robust
in the presence of delay mismatches.
As shown with an example in Sec. IV, this research can be extended to more general cases. For example, it is
applicable to higher order systems with time delays and to design of more general controllers (e.g., PID control).
Further theoretical studies, as well as experimental implementation and evaluation studies, are currently being
conducted by the authors.
VI. ACKNOWLEDGMENTS
R EFERENCES
[1] K.H. Ang, G. Chong, and Y. Li. PID control system analysis, design, and technology. IEEE Transactions on Control Systems Technology,
13(4):559–576, 2005.
[2] K. J. Åström and T. Hägglund. Advanced PID Control. ISA-The Instrumentation, Systems, and Automation Society, 2006.
[3] M. Bozorg and F. Termeh. Domains of PID controller coefficients which guarantee stability and performance for LTI time-delay systems.
Automatica, 47(9):2122–2125, 2011.
[4] A.M. De Paor and M. O’malley. Controllers of Ziegler-Nichols type for unstable process with time delay. International Journal of Control,
49(4):1273–1284, 1989.
[5] S. Duan. http://www-personal.umich.edu/∼ulsoy/TDS Supplement.htm, 2010.
[6] T. Furukawa and E. Shimemura. Predictive control for systems with time-delay. International Journal of Control, 37(2):399–412, 1983.
[7] P. Gawthrop. Act-and-wait and intermittent control: some comments. IEEE Transactions on Control Systems Technology, 18(5):1195–1198,
2010.
[8] A.N. Gündes, H. Özbay, and A.B. Özgüler. PID controller synthesis for a class of unstable MIMO plants with i/o delays. Automatica,
43(1):135–142, 2007.
[9] L. Guzzella and A. Amstutz. Control of diesel engines. IEEE Control Systems Magazine, 18(5):53–71, 1998.
[10] L. Guzzella and C. H. Onder. Introduction to Modeling and Control of Internal Combustion Engine Systems. Springer, Berlin, 2004.
[11] T. Insperger. Act-and-wait concept for continuous-time control systems with feedback delay. IEEE Transactions on Control Systems
Technology, 14(5):974–977, 2006.
[12] T. Insperger and G. Stepan. Semi-Discretization for Time-Delay Systems: Stability and Engineering Applications. Springer, New York,
2011.
[13] T. Insperger, P. Wahi, A. Colombo, G. Stepan, M. Di Bernardo, and SJ Hogan. Full characterization of act-and-wait control for first-order
unstable lag processes. Journal of Vibration and Control, 16(7-8):1209, 2010.
[14] W. Krajewski, A. Lepschy, and U. Viaro. Designing PI controllers for robust stability and performance. IEEE Transactions on Control
Systems Technology, 12(6):973–983, 2004.
[15] T. Lee, J.M. Watkins, T. Emami, and S. Sujoldzic. A unified approach for stabilization of arbitrary order continuous-time and discrete-time
transfer functions with time delay. In 2007 46th IEEE Conference on Decision and Control, pages 2100–2105. IEEE, 2007.
[16] T. Li and E.K.W. Chu. Pole assignment for linear and quadratic systems with time-delay in control. Numerical Linear Algebra with
Applications, 2011.
[17] T. Liu, YZ Cai, DY Gu, and WD Zhang. New modified smith predictor scheme for integrating and unstable processes with time delay.
Control Theory and Applications, 152(2):238–246, 2005.
[18] L. Mianzo, H. Peng, and I. Haskara. Transient air-fuel ratio H∞ preview control of a drive-by-wire internal combustion engine. In Proc.
2001 American Control Conference, pages 2867–2871, 2001.
[19] W. Michiels, K. Engelborghs, P. Vansevenant, and D. Roose. Continuous pole placement for delay equations. Automatica, 38(5):747–761,
2002.
[20] W. Michiels and S. I. Niculescu. On the delay sensitivity of Smith predictors. International Journal of Systems Science, 34(8-9):543–551,
2003.
[21] S. Nakagawa, K. Katogi, and M. Oosuga. A new air-fuel ratio feed back control for ULEV/SULEV standard. SAE Paper 2002-01-0194,
SAE World Congress 2002.
[22] S. I. Niculescu and A. M. Annaswamy. An adaptive Smith-controller for time-delay systems with relative degree n∗ <= 2. Systems &
Control Letters, 49(5):347–358, 2003.
[23] L. Ou, W. Zhang, and L. Yu. Low-order stabilization of LTI systems with time delay. IEEE Transactions on Automatic Control, 54(4):774–
787, 2009.
[24] H. Özbay and A.N. Gündes. Resilient PI and PD controller designs for a class of unstable plants with I/O delays. Appl. Comput. Math,
6(1):18–26, 2007.
[25] E. Poulin and A. Pomerleau. PID tuning for integrating and unstable processes. volume 143, pages 429–435. IET, 1996.
[26] J. D. Powell, N. P. Fekete, and C. F. Chang. Observer-based air-fuel ratio control. IEEE Control Systems Magazine, 18(5):72–83, 1998.
[27] J. P. Richard. Time-delay systems: an overview of some recent advances and open problems. Automatica, 39(10):1667–1694, 2003.
[28] H. Shinozaki and T. Mori. Robust stability analysis of linear time-delay systems by Lambert W function: Some extreme point results.
Automatica, 42(10):1791–1799, 2006.
[29] G. J. Silva, A. Datta, and S. P. Bhattacharyya. PI stabilization of first-order systems with time delay. Automatica, 37(12):2025–2031,
2001.
[30] G. J. Silva, A. Datta, and S. P. Bhattacharyya. PID Controllers for Time-Delay Systems. Control Engineering. Birkhäuser, Boston, 2005.
[31] G.J. Silva, A. Datta, and SP Bhattacharyya. New results on the synthesis of PID controllers. IEEE Transactions on Automatic Control,
47(2):241–252, 2002.
[32] O.J.M. Smith. Closer control of loops with dead time. Chemical Engineering Progress, 53:217–219, 1957.
[33] S. Sujoldzic and J.M. Watkins. Stabilization of an arbitrary order transfer function with time delay using a PID controller. In 2006 45th
IEEE Conference on Decision and Control, pages 846–851. IEEE, 2006.
[34] J. Tuch, A. Feuer, and Z. J. Palmor. Time-delay estimation in continuous linear time-invariant systems. IEEE Transactions on Automatic
Control, 39(4):823–827, 1994.
[35] R. Turin and H. Geering. Model-reference adaptive A/F-ratio control in an SI engine based on Kalman-filtering techniques. In Proc. 1995
American Control Conference, pages 4082–4090, 1995.
[36] V. Venkatashankar and M. Chidambaram. Design of P and PI controllers for unstable first-order plus time delay systems. International
Journal of Control, 60(1):137–144, 1994.
[37] Q.G. Wang, Z. Zhang, K.J. Astrom, and L.S. Chek. Guaranteed dominant pole placement with PID controllers. Journal of Process Control,
19(2):349–352, 2009.
[38] Ya-Gang Wang and Hui-He Shao. Optimal tuning for PI controller. Automatica, 36(1):147–152, 2000.
[39] K. Yamanaka and E. Shimemura. Effects of mismatched smith controller on stability in systems with time-delay. Automatica, 23(6):787–
791, 1987.
[40] S. Yi, P. W. Nelson, and A. G. Ulsoy. Delay differential equations via the matrix Lambert W function and bifurcation analysis: Application
to machine tool chatter. Mathematical Biosciences and Engineering, 4(2):355–368, 2007.
[41] S. Yi, P. W. Nelson, and A. G. Ulsoy. Controllability and observability of systems of linear delay differential equations via the matrix
Lambert W function. IEEE Transactions on Automatic Control, 53(3):854–860, 2008.
[42] S. Yi, P. W. Nelson, and A. G. Ulsoy. Eigenvalues and sensitivity analysis for a model of HIV-1 pathogenesis with an intracellular delay.
In Proceedings of 2008 ASME Dynamic Systems and Control Conference, Ann Arbor, MI, Oct. 2008, 2008. DSCC2008-2408.
[43] S. Yi, P. W. Nelson, and A. G. Ulsoy. Eigenvalue assignment via the Lambert W function for control for time-delay systems. Journal of
Vibration and Control, 16(7-8):961–982, 2010.
[44] S. Yi, P. W. Nelson, and A. G. Ulsoy. Robust control and time-domain specifications for systems for delay differential equations via
eigenvalue assignment. Journal of Dynamic Systems, Measurement and Control, 132(7), 2010. 7 pages.
[45] S. Yi, P. W. Nelson, and A. G. Ulsoy. Time-Delay Systems: Analysis and Control Using the Lambert W Function. World Scientific, 2010.
[46] Y. Yildiz, A. Annaswamy, D. Yanakiev, and I. Kolmanovsky. Adaptive air fuel ratio control for internal combustion engines. In Proc.
2008 American Control Conference, Seattle, WA, 2008. 2058-2063.
[47] Q. C. Zhong. Robust Control of Time-Delay Systems. Springer, London, 2006.