See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/228392562

Some approximations of fractional order operators used in control theory

Article in Fractional Calculus and Applied Analysis · January 2000

CITATIONS READS

209 4,127

4 authors, including:

Blas M. Vinagre Vicente Feliu

Universidad de Extremadura University of Castilla-La Mancha
216 PUBLICATIONS 9,807 CITATIONS 436 PUBLICATIONS 10,520 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Control del proceso productivo de fabricación del acero View project

Control of dominant time delay processes View project

All content following this page was uploaded by Vicente Feliu on 08 November 2014.

The user has requested enhancement of the downloaded file.

 SOME APPROXIMATIONS
OF FRACTIONAL ORDER OPERATORS
USED IN CONTROL THEORY AND APPLICATIONS

B. M. Vinagre1 , I. Podlubny2 , A. Hernández1 , V. Feliu3

Abstract

In this paper, starting from the formulation of some possible models of fraction-
al-order systems, several approximations are discussed. For continuous models,
some methods for obtaining an approximated rational function using evaluation,
interpolation and curve fitting techniques are studied. For discrete models, ap-
proximations using Lubich’s formula, the trapezoidal rule, and the application of
continued fractions expansion technique to integro-differential operators formu-
lated in the Z domain are studied. The methods are compared, in both the time
and the frequency domains, using an illustrative example.
Mathematics Subject Classification: 26A33 (main), 93C15, 93C55, 93C80
Key Words and Phrases: fractional calculus, fractional-order systems, frac-
tional-order controllers, integer-order approximations

1. Introduction

At least since the sixties some researchers have been interested in obtaining ap-
proximated integer order models for fractional order systems, or finite dimensional
models for infinite dimensional systems. Most of these researchers worked in elec-
trochemistry where the problem is (roughly speaking) to build an equivalent elec-
trical circuit for processes in which diffusion is present. Even earlier, other authors
Partially supported by FEDER Research Grant IFD97-0755-C02-01
and by VEGA Reserach Grant 1/7098/20.
2 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

used fractional-order models for modeling physical systems, in which the presence
of viscoelasticity was important, or in the characterization of certain class of real
materials. However, applications of fractional-order models in control theory are
relatively new comparing to mentioned applications. As far as the authors of this
paper know, except in old papers of Carlson and Halijak [3, 4], only in the last
two decades the possibility of using fractional order controller has been considered
(see, for example, [1, 2, 12, 14, 17, 18, 20, 19, 21, 23, 24, 28]). In all the cases, the
important point for the purposes of this paper, is that the equivalent circuit or
the fractional-order controller in its practically realizable form is a finite dimen-
sional integer-order system resulting from the approximation of a fractional-order
or infinite-dimensional system.
In this paper the Riemann-Liouville definition of fractional integration and
differentiation is used [13, 16, 24, 27]. For the case of 0 < α < 1 and f (t) being a
causal function of t, that is, f (t) = 0 for t < 0, the fractional integral is defined
as:
Z t
1
D−α f (t) = (t − τ )α−1 f (τ ) dτ, (1)
Γ(α) 0

and the expression for the fractional order derivative is:

Z t
1 d
Dα f (t) = (t − τ )−α f (τ ) dτ (2)
Γ(1 − α) dt 0

Of particular interest for the purposes of this paper are the expressions of the
former operators in the Laplace domain. The Laplace transform of the Riemann-
Liouville fractional integral is:

£ D −α f (t) = s −α F (s), (3)

and for the fractional derivative of order 0 < α < 1 it is:

 
£ {D α f (t)} = s α F (s) − D α−1 f (t) t=0 (4)

A fractional-order control system can be described by a fractional differential

equation of the form [24, 23]:

an Dαn y(t) + an−1 Dαn−1 y(t) + . . . + a0 Dα0 y(t) =

bm Dβm u(t) + bm−1 Dβm−1 u(t) + . . . + b0 Dβ0 u(t)

or by a continuous transfer function of the form [24, 23]:

bm sβm + bm−1 sβm−1 + . . . + b0 sβ0

G(s) = (5)
an sαn + an−1 sαn−1 + . . . + a0 sα0
 APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 3

For obtaining discrete models of fractional order systems, it is necessary to

use discrete approximations of the fractional integrals and derivatives. By doing
so in the expression given above for G(s) a general expression for the discrete
transfer function of the fractional system, G(z), can be obtained, of the form:


β

β

β
bm w z −1 m + bm−1 w z −1 m−1 + . . . + b0 w z −1 0
G(z) = (6)
an (w (z −1 ))αn + an−1 (w (z −1 ))αn−1 + . . . + a0 (w (z −1 ))α0



where w z −1 denotes the discrete equivalent of the Laplace operator s, ex-
pressed as a function of the complex variable z or the shift operator z −1 .
As can be seen in the former expressions, a fractional-order system has an
irrational continuous transfer function in the Laplace domain or, in the Z domain,
an infinite dimensional discrete transfer function. In other words, a fractional-
order system has an unlimited memory, being the integer order systems particular
cases of this general case in which the memory is limited. It is obvious that only
in the case of integer order it is possible to realize a transfer function exactly by
using conventional lumped elements (resistances, inductances, and capacitors, in
the case of analog realizations), or procedures (finite order difference equations or
digital filters in the case of discrete realizations).
Because of this, and taking into account that the final step for applying a
fractional controller demands a realizable form of it, in this paper some contin-
uous and discrete integer-order approximations of fractional-order operators are
considered.

2. Integer-Order Continuous Models of Fractional Order Systems

The problem of obtaining a continuous realizable model for a fractional order

controller can be viewed as a problem of obtaining a rational approximation of
the irrational transfer function, modeling the fractional controller. Among other
mathematical methods, two of them are particularly interesting for this purpose,
from a control theory point of view: the continued fraction expansion method
used for evaluation of functions, and the rational approximation method used in
interpolation of functions.
On the other hand, the use of frequency identification or curve fitting meth-
ods for obtaining rational approximations to the irrational frequency responses,
characterizing fractional-order systems, is proposed.
In this section, the general form of the methods and some specially interesting
applications of them are described.
4 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

Approximations using continued fraction expansions

and interpolation techniques
It is well known that the continued fraction expansions (CFE) is a method for
evaluation of functions, that frequently converges much more rapidly than power
series expansions, and converges in a much larger domain in the complex plane
[25]. The result of such approximation for an irrational function, G(s), can be
expressed in the form:
b1 (s)
G(s) ' a0 (s) + b2 (s)
a1 (s) + b3 (s)
a2 (s)+ a
3 (s)+...

b1 (s) b2 (s) b3 (s)

= a0 (s) + ... (7)
a1 (s)+ a2 (s)+ a3 (s)+
where a0i s and b0i s are rational functions of the variable s, or are constant. The ap-
b
plication of the method yields a rational function, G(s), which is an approximation
of the irrational function G(s).
On the other hand, for interpolation purposes, rational functions are some-
times superior to polynomials. This is, roughly speaking, due to their ability
to model functions with poles. (As it can be seen later, branch points can be
considered as accumulations of interlaced poles and zeros). These techniques are
based on the approximations of an irrational function, G(s), by a rational function
defined by the quotient of two polynomials in the variable s:
Pµ (s)
G(s) ' Ri(i+1)...(i+m) =
Qν (s)
p0 + p 1 s + . . . + p µ sµ
= (8)
q0 + q 1 s + . . . + q ν sν
m+1 = µ+ν+1
passing through the points (si , G(si )), . . . , (si+m , G(si+m )).

General CFE method for approximation of fractional integro-differ-

ential operators. In general [26], a rational approximation of the function
G(s) = s−α , 0 < α < 1 (the fractional integral operator in the Laplace domain)
can be obtained by performing the CFE of the functions:
1
Gh (s) = (9)
(1 + sT )α
 
1 α
Gl (s) = 1+ (10)
s
where Gh (s) is the approximation for high frequencies (ωT >> 1), and Gl (s) the
approximation for low frequencies (ω << 1).
 APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 5

Carlson’s method. The method proposed by Carlson in [4], derived from a

regular Newton process used for iterative approximation of the α-th root, can be
considered as belonging to this group. The starting point of the method is the
statement of the following relationships:

(H(s))1/α − (G(s)) = 0; H(s) = (G(s))α (11)

Defining α = 1/q, m = q/2, in each iteration, starting from the initial value
H0 (s) = 1, an approximated rational function is obtained in the form:

(q − m) (Hi−1 (s))2 + (q + m)G(s)

Hi (s) = Hi−1 (s) (12)
(q + m) (Hi−1 (s))2 + (q − m)G(s)

Matsuda’s method. The method suggested in [12] is based on the approxi-

mation of an irrational function by a rational one, obtained by CFE and fitting
the original function in a set of logarithmically spaced points. Assuming that the
selected points are sk , k = 0, 1, 2, . . . , the approximation takes on the form:
s − s0 s − s1 s − s2
H(s) = a0 + ··· , (13)
a1 + a2 + a3 +
where
s − si
ai = vi (si ), v0 (s) = H(s), vi+1 (s) = (14)
vi (s) − ai

Approximations using curve fitting or identification techniques

In general, any available method for frequency domain identification can be ap-
plied in order to obtain a rational function, whose frequency response fits the
frequency response of the original irrational transfer function. For example, this
may be minimization of the cost function of the ISE form, that is
Z
2
J = W (ω) G(ω) − G(ω) b dω

where W (ω) is a weighting function, G(ω) is the original frequency response, and
b
G(ω) is the frequency response of the approximated rational function.

Oustaloup’s method. The method [17, 18, 20] is based on the approximation
of a function of the form:

H(s) = sµ , µ ∈ R+ (15)
6 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

by a rational function:
N
Y
b 1 + s/ωk
H(s) =C (16)
1 + s/ωk0
k=−N

using the following set of synthesis formulas:

0
ωk+1 ωk+1
ω00 = α−0.5 ωu ; ω0 = α0.5 ωu ; = = αη > 1; (17)
ωk0 ωk
0
ωk+1 ωk log (ωN /ω0 ) log α
= η > 0; = α > 0; N= ; µ=
ωk ωk0 log (αη) log (αη)
with ωu being the unit gain frequency and the central frequency of a band of
√
frequencies geometrically distributed around it. That is, ωu = ωh ωb , ωh , ωb are
the high and low transitional frequencies.

Chareff ’s method. This method, proposed in [5], which is very close to Ous-
taloup’s method, is based on approximation of a function of the form
1
H(s) = (18)
(1 + psT )α

by a quotient of polynomials in s in a factorized form:

Q 
n−1 
1 + zsi
b
H(s) = i=0
n   (19)
Q
1 + psi
i=0

where the coefficients are computed for obtaining a maximum deviation from the
original magnitude response in the frequency domain of y dB. Defining
a = 10y/10(1−α), b = 10y/10α , ab = 10y/10α(1−α) (20)
the poles and zeros of the approximated rational function are obtained by applying
the following formulae:
√
p0 = pT b, pi = p0 (ab)i , zi = ap0 (ab)i (21)
The number of poles and zeros is related to the desired bandwidth and the
error criteria used by the expression:
  
log ωmáx
p0
N = +1 (22)
log (ab)
 APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 7

3. Discrete Models

In general, if a function f (t) is approximated by a grid function, f (nh), where

h the grid size, the approximation for its fractional derivative of order α can be
expressed as [6]:


α
yh (nh) = h−α ω ζ −1 fh (nh) (23)


where ζ −1 is the shift operator, and ω ζ −1 is a generating function. This gen-
erating function and its expansion determine both the form of the approximation
and its coefficients.
It is worth mentioning that, in general, the case of controller realization is not
equivalent to the cases of simulation or numerical evaluation of the fractional inte-
gral and differential operators. In the case of controller realization it is necessary
to take into account some important considerations. First of all, the value of h,
the step when dealing with numerical evaluation, is the value of the sample period
T , and it is limited by the characteristics of the microprocessor-based system, used
for the controller implementation, in two ways: (i) each microprocessor-based sys-
tem has its own minimum value for the sample period, and (ii) it is necessary to
perform all the computations required by the control law between two samples.
Due to this last reason, it is very important to obtain good approximations with
a minimal set of parameters. On the other hand, when the number of parameters
in the approximation increases, it increases the amount of the required memory
too.
It is also important to have discrete equivalents or approximations with poles
and zeros, that is, in a rational form.
In the following, the notation normally used in control theory is adopted, that
is: T , the sample period, is used instead of h, and z, the complex variable resulting
from the application of the Z transform to the functions y(nT ), f (nT ) considered
as sequences, is used instead of ζ.

Discrete Approximations using Numerical Integration

and Power Series Expansion
Using the generating

function

corresponding to the backward fractional difference
rule, ω z
−1 = 1 − z −1 , and performing the power series expansion (PSE) of
α
1 − z −1 , the Grünwald–Letnikov formula for the fractional derivative of order
α is obtained:
∞
X  
α
∇α
T f (nT ) =T −α
(−1)k
f ((n − k)T ) (24)
k
k=0
8 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

−α
Performing the PSE of the function 1 − z −1 leads to the formula given
by Lubich for the fractional integral of order α [11]:
∞
X  
α k −α
∇−α
T f (nT ) =T (−1) f ((n − k)T ) (25)
k
k=0

In any case, the resulting transfer function, approximating the fractional-order

operators, can be obtained by applying the relationship:
n
±α o
Y (z) = T ∓α PSE 1 − z −1 F (z) (26)

where T is the sample period, Y (z) is the Z transform of the output sequence
y(nT ), F (z) is the Z transform of the input sequence f (nT ), and PSE{u} denotes
the expression, which results from the power series expansion of the function u.
Doing so gives:

Y (z) n
±α o
D±α (z) = = T ∓α PSE 1 − z −1 (27)
F (z)

where D ±α (z) denotes the discrete equivalent of the fractional-order operator,

considered as processes.
Another possibility for the approximation is the use of the trapezoidal rule,
that is, the use of the generating function

1 − z −1
ω(z −1 ) = 2 (28)
1 + z −1
It is known that the forward difference rule is not suitable for applications to
causal problems [6, 11].
It should be mentioned that, at least for control purposes, it is not very
important to have a closed-form formula for the coefficients, because they are
usually pre-computed and stored in the memory of the microprocessor. In such
a case, the most important is to have a limited number of coefficients because of
the limited available memory of the microprocessor system.

Discrete approximations using numerical integration

and continued fraction expansion
The approximations, considered in the previous section, lead to discrete transfer
functions in the form of polynomials, and this is not convenient, at least from
the control point of view. On the other hand, it can be recalled that the CFE
leads to approximations in rational form, and often converges much more rapidly
than PSE and has a wider domain of convergence in the complex plane, and,
 APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 9

consequently, a smaller set of coefficients will be necessary for obtaining a good

approximation.
In view of these reasons, a method for obtaining discrete equivalents of the
fractional-order operators, which combines the well known advantages of the
trapezoidal rule in the control theory and the advantages of the CFE, is proposed
here. The method implies:

• the use of the generating function

1 − z −1
ω(z −1 ) = 2
1 + z −1

where z is the complex variable, and z −1 is the shifting operator,

• and the continued fraction expansion (CFE) of

 ±α
−1

±α 1 − z −1
ω(z ) = 2
1 + z −1

for obtaining the coefficients and the form of the approximation.

The resulting discrete transfer function, approximating fractional-order oper-

ators, can be expressed as:
( ±α )
±α Y (z) ∓α 1 − z −1
D (z) = =T CFE 2
F (z) 1 + z −1
p,q
Pp (z −1 )
= T ∓α (29)
Qq (z −1 )

where T is the sample period, CFE{u} denotes the function resulting from apply-
ing the continued fraction expansion to the function u, Y (z) is the Z transform of
the output sequence y(nT ), F (z) is the Z transform of the input sequence f (nT ),
p and q are the orders of the approximation, and P and Q are polynomials of
degrees p and q, correspondingly, in the variable z −1 .

Other approximations
In the paper [9] some others continuous approximations have been studied, which,
in fact, are particular cases of the methods considered here. Some methods are
not considered in the present paper, because they can be used only for particular
values of fractional order (e.g., [3] for α = 0.5).
On the other hand, it is necessary to mention the diffusive realizations pro-
posed in [15] and [8]. While the starting point in these works is quite different
10 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

from the starting point of the methods considered in the present paper, the re-
sulting approximations can be viewed, in the Laplace domain, as rational approx-
imations of the fractional-order operators. Furthermore, these approximations
exhibit a common feature, which we observe in all good rational approximations:
they have poles and zeros interlaced on the negative real axis of the s plane, and
the distance between successive poles and zeros decreases as the approximation
is improved by increasing the degree of the numerator and denominator polyno-
mials. Probably, this fact was noted for the first time in [10], where the following
idea appeared: a dense interlacing of simple poles and zeros along a line in the s
plane is, in some way, equivalent to a branch cut; and sα , 0 < α < 1, viewed as
an operator, has a branch cut along the negative real axis for arguments of s on
(−π, π) but is otherwise free of poles and zeros.

4. Illustrative Example: Fractional Integrator of Order α = 0.5

For comparing the methods, the approximations of the fractional integrator of

order 0.5, and the corresponding step and frequency responses have been obtained.
For simplicity, the approximations obtained are (except for the one derived from
Lubich’s formula) of low order (4 or 5 depending on the particular requirements
of the methods), and necessary scale factors have been applied in order to have
the same central frequency and 0 dB at ω = 1 rad/s.

Approximated rational functions

General CFE method. Performing the CFE of the function (9), with T = 1,
α = 0.5, we obtain:

0.3513s4 + 1.405s3 + 0.8433s2 + 0.1574s + 0.008995

H1 (s) =
s4 + 1.333s3 + 0.478s2 + 0.064s + 0.002844

Performing the CFE of the function (10), with T = 1, α = 0.5, we obtain:

s4 + 4s3 + 2.4s2 + 0.448s + 0.0256

H2 (s) =
9s4 + 12s3 + 4.32s2 + 0.576s + 0.0256

1 1/2
Carlson’s method. Starting from: H(s) = s , H0 (s) = 1, after two itera-
tions we obtain:

s4 + 36s3 + 126s2 + 84s + 9

H3 (s) =
9s4 + 84s3 + 126s2 + 36s + 1
 APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 11

1/2
Matsuda‘s method. With G(s) = 1s , finitial = 1, ff inal = 100,
fk ={ 1, 1.7783, 3.1623, 5.6234, 10, 17.783, 31.623, 56.234, 100 }, we obtain:

0.08549s4 + 4.877s3 + 20.84s2 + 12.995s + 1

H4 (s) =
s4 + 13s3 + 20.84s2 + 4.876s + 0.08551

Least-squares method. Using the MATLAB invfreqs function we obtain:

0.09593s4 + 5.909s3 + 40.7s2 + 44.29s + 3.859

H5 (s) =
s4 + 18.07s3 + 56.01s2 + 17.97s + 0.2685
Using a method similar to that proposed in [29], we obtain:

0.1002s4 + 4.011s3 + 11.26s2 + 5.076s + 0.3694

H6 (s) =
s4 + 8.654s3 + 9.364s2 + 1.771s + 0.03744
 
Chareff ’s method. For y = 2dB, pT = 1, ω ∈ 10−1 , 105 , the obtained ap-
proximation is:

6.3s4 + 74.84s3 + 121.1s2 + 29.79s + 0.9986

H7 (s) =
s5 + 29.85s4 + 121.8s3 + 76.85s2 + 7.497s + 0.1

Oustaloup’s method. Using the Oustaloup’s method with:

ωh = 102 , ωb = 10−2 ,

from which we have α = η = 2.5119, the obtained approximation is:

s5 + 74.97s4 + 768.5s3 + 1218s2 + 298.5s + 10

H8 (s) =
10s5 + 298.5s4 + 1218s3 + 768.5s2 + 74.97s + 1

Discrete approximation using backward rule and PSE. This approxima-

tion can be viewed as the Z version of Lubich’s formula with the necessary use of
the short-memory principle [22]. With T = 0.1 and 101 coefficients, the resulting
discrete transfer function is:
100
X  
0.5 −100 k −0.5
H9 (z) = 0.1 z (−1) z 100−k
k
k=0

Discrete approximation using backward rule and CFE. With T = 0.1

we obtain:
0.3162z 4 − 0.5534z 3 + 0.2965z 2 − 0.0494z + 0.0012
H10 (z) =
z 4 − 2.25z 3 + 1.6875z 2 − 0.4688z + 0.0352
12 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

Phase (deg); Magnitude (dB)

Phase (deg); Magnitude (dB)

10 10
5 5
H (s) H (s)
0 1 0 2
−5 −5

−20 −20

−40 −40
−2 −1 0 1 2 −2 −1 0 1 2
10 10 10 10 10 10 10 10 10 10
Frequency (rad/sec) Frequency (rad/sec)
Phase (deg); Magnitude (dB)

Phase (deg); Magnitude (dB)

20 20
H (s) H (s)
3 4
0 0

−20 −20

−20 −20

−40 −40
−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3
10 10 10 10 10 10 10 10 10 10 10 10 10 10
Frequency (rad/sec) Frequency (rad/sec)

Figure 1: Bode plots of H1 (s), H2 (s), H3 (s), and H4 (s).

Phase (deg); Magnitude (dB)

Phase (deg); Magnitude (dB)

20 20
H (s) H (s)
0 5 0 6

−20 −20

−10
−20 −20
−30
−40 −40
−50
−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3
10 10 10 10 10 10 10 10 10 10 10 10 10 10
Frequency (rad/sec) Frequency (rad/sec)
Phase (deg); Magnitude (dB)

Phase (deg); Magnitude (dB)

20 20
0 H7(s) H8(s)
0
−20
−40 −20
0
−20
−40 −20
−60
−80 −40
−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3
10 10 10 10 10 10 10 10 10 10 10 10 10 10
Frequency (rad/sec) Frequency (rad/sec)

Figure 2: Bode plots of H5 (s), H6 (s), H7 (s), and H8 (s).

APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 13

7 7

6 H (s) 6 H (s)
1 2

5 5

4 4

3 3

2 2

1 1

0 0
0 10 20 30 0 10 20 30
Time (sec.) Time (sec.)

7 7

6 H (s) 6 H (s)
3 4

5 5

4 4

3 3

2 2

1 1

0 0
0 10 20 30 0 10 20 30
Time (sec.) Time (sec.)

Figure 3: Step responses of H1 (s), H2 (s), H3 (s), and H4 (s).

7 7

6 H (s) 6 H (s)
5 6

5 5

4 4

3 3

2 2

1 1

0 0
0 10 20 30 0 10 20 30
Time (sec.) Time (sec.)

7 7

6 H7(s) 6 H8(s)

5 5

4 4

3 3

2 2

1 1

0 0
0 10 20 30 0 10 20 30
Time (sec.) Time (sec.)

Figure 4: Step responses of H5 (s), H6 (s), H7 (s), and H8 (s).

14 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

20 7

10 6

Phase(deg); Magnitude (dB)

H (z)
0 9
H (z)
5 9
−10

−20 4

20 3
0
2
−20

−40 1

−60 0
−1 0 1 2 0 5 10 15 20 25 30
10 10 10 10
Frequency (rad/sec) Time (sec)

10 7

H10(z) 6
Phase(deg); Magnitude (dB)

0
H (z)
5 10
−10

−20 4

0 3

−20 2

−40 1

−60 0
−2 −1 0 1 2 0 5 10 15 20 25 30
10 10 10 10 10
Frequency (rad/sec) Time (sec.)

20 7

10 6 H (z)
Phase(deg); Magnitude (dB)

H (z) 11
0 11
5
−10

−20 4

0 3

−20 2

−40 1

−60 0
−3 −2 −1 0 1 2 0 5 10 15 20 25 30
10 10 10 10 10 10
Frequency (rad/sec) Time (sec.)

Figure 5: Bode plots and step responses of H9 (z), H10 (z), H11 (z).

Discrete approximation using trapezoidal rule and CFE. With T = 0.1

we obtain:
0.2564z 4 − 0.6395z 3 + 0.4887z 2 − 0.0779z − 0.0277
H11 (z) =
z 4 − 3.2834z 3 + 3.9474z 2 − 2.0419z + 0.3779
The figures 1-5 shows the step responses, compared with the exact step re-
sponse, and the frequency responses (Bode plots) of the obtained approximations.

5. Controller Realization

In general, there are two possibilities for realizing a controller: a hardware real-
ization based on the use of a physical device, or a software (or digital) realiza-
tion based on a program, which will run on a computer or microprocessor. In
electronics, hardware realizations imply the use of electronic devices or circuits,
implementing the required function as an admittance or impedance function.

Analog realizations. For hardware electronic realizations, the starting point is

the admittance or impedance function. For realizing such functions, at least two
 APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 15

ways can be used: producing a microelectronic specific device that, for construc-
tion, has the required admittance or impedance (see, for example, [7]), or realizing
an approximate rational function by using a finite lumped-element network, in a
ladder, tree, cascaded, or lattice topology. Since this paper deals with rational
approximations of the fractional-order operators, only the last way is discussed
below.
In order to be realizable by a finite lumped-element network, the rational
approximation must be a positive real rational function. We see that the contin-
uous approximations given in Section 2, fulfill this requirement for being rational
functions with poles and zeros interlaced on the negative real axis of the s plane.
In view of this, they can be realized as the driving point impedance or admit-
tance of RL or RC networks, that is, as passive RL or RC filters. Furthermore,
these realizations can be complemented by using active electronic devices, such
as operational amplifiers. That is, a more flexible way for realizing the required
functions can be obtained by considering the possibility of using active filters. In
laboratory conditions, we have successfully tested an active filter, which realizes
the fractional integrator of order 0.5 by using a lattice RC network, obtained from
a former version of the Carlson’s method [3].

Digital realizations. For digital realizations, a finite difference equation is

needed. This equation can be obtained by numerical approximation of by perform-
ing the inverse Z transform of a discrete transfer function. For this, the discrete
approximations given in Section 3 or discrete equivalents of the continuous ap-
proximations can be used. Furthermore, by considering the system described by a
discrete transfer function as a IIR digital filter, useful considerations well known
in digital filter realization can be taken into account in order to choose a best
structure for realizing a discrete approximation.

6. Conclusion

In this paper some approximations of fractional-order operators have been con-

sidered, in order to use them for implementing fractional-order controllers. These
approximations have been compared, both in frequency domain and in time do-
main, for a fractional integrator of order α = 0.5.
From the obtained results a general conclusion can be made: both in con-
tinuous and discrete domains it is possible to use rational approximations for
these operators in order to have suitable realizable forms in control applications
(with some limitation regarding the frequency bandwidth or the time interval,
correspondingly).
Furthermore, it has been shown that very good continuous approximations can
be obtained by using the methods proposed by Carlson, Oustaloup, Matsuda, and
Chareff, or least-square methods for fitting the original frequency response. For
16 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

discrete models, our results seem to indicate that there are two suitable techniques
for obtaining discrete approximations of the fractional operators: the discretiza-
tion of an appropriate continuous approximation, or the use of the CFE of the
generating function defined by the trapezoidal rule.
We would like to mention that there is a demand and motivation for further
work towards practical implementation of fractional-order controllers.

References

[1] R. L. B a g l e y and R. A. C a l i c o. Fractional order state equations

for the control of viscoelastically damped structures. Journal of Guidance,
14, No. 2 (1991), 304–311.

[2] P. B i d a n. Commande diffusive d’une machine électrique: une introduction.

In D. Matignon and G. Montseny, editors, Proceedings of the Colloquium
FDS’98: Fractional Differential Systems: Models, Methods and Applications,
No. 5 (1998), 55–68.

[3] G. E. C a r l s o n and C. A. H a l i j a k. Simulation of the fractional

√ √
derivative operator s and the fractional integral operator 1/ s. Proc. of
the CSSCM, Kansas State U. B., 45, No. 7 (1961), 1–22.

[4] G. E. C a r l s o n and C. A. H a l i j a k. Approximation of fractional

capacitors (1/s)1/n by a regular Newton process. IRE Transactions on Circuit
Theory, CT-11, No. 2 (1964), 210–213.

[5] A. C h a r e f f, H. H. S u n, Y. Y. T s a o, and B. O n a r a l. Fractal sys-

tem as represented by singularity function. IEEE Transactions on Automatic
Control, 37, No. 9 (1992), 1465–1470.

[6] R. G o r e n f l o. Fractional calculus: Some numerical methods. In:

A. Carpinteri and F. Mainardi (eds.). Fractals and Fractional Calculus in
Continuum Mechanics, Springer Verlag, Vienna – New York (1997).

[7] C. H a b a, M. M a r t o s, G. A b l a r t, and P. B i d a n. Composants

électroniques á impédance fractionnaire. In D. Matignon and G. Montseny,
editors, Proceedings of the Colloquium FDS’98: Fractional Differential Sys-
tems: Models, Methods and Applications, No. 5 (1998), 99–109.

[8] D. H e l e s c h e w i t z and D. M a t i g n o n. Diffusive realisations

of fractional integrodifferential operators: structural analysis under approxi-
mation. IFAC Conference on System, Structure and Control, Vol. 2, Nantes,
France, (1998), 243-248.
 APPROXIMATIONS OF FRACTIONAL ORDER OPERATORS 17

[9] A. H e r n á n d e z, B. V i n a g r e, and V. F e l i u. Realización de reg-

uladores fraccionarios. Actas de las XX Jornadas de Automática, Salamanca
(1999), 235–240.

[10] H. E. J o n e s and B. A. S h e n o i. Maximally flat lumped-element

approximation to fractional operator inmitance function. IEEE Transactions
on Circuit Theory (1970), 17, No. 1 (1970), 125-128.

[11] C h. L u b i c h. Discretized fractional calculus. SIAM Journal of Mathe-

matical Analysis, 17, No. 3 (1986), 704–719.

[12] K. M a t s u d a and H. F u j i i. H∞ optimized wave-absorbing

control: Analytical and experimental results. Journal of Guidance, Control,
and Dynamics, 16, No. 6 (1993), 1146–1153.

[13] K. S. M i l l e r and B. R o s s. An Introduction to the Fractional Calculus

and Fractional Differential Equations. John Wiley & Sons. Inc., New York
(1993).

[14] G. M o n t s e n y, J. A u d o u n e t, and D. M a t i g n o n. Fractional

integrodifferential boundary control of the Euler–Bernoulli beam. 36th IEEE
Conference on Decision and Control, San Diego (December 1997), 4973–4978.

[15] G. M o n t s e n y, J. A u d o u n e t, and B. M b o d j e. Opti-

mal models of fractional integrators and applications to systems with fading
memory. Proceedings of the Conference IEEE Systems, Man and Cybernetics,
Le Touquet, France, (1993), 65-70.

[16] K. B. O l d h a m and J. S p a n i e r. The Fractional Calculus. Academic

Press, New York - London (1974).

[17] A. O u s t a l o u p. Systèmes Asservis Linéaires d’Ordre Fractionnaire:

Théorie et Pratique. Editions Masson, Paris (1983).

[18] A. O u s t a l o u p. La commande CRONE. Editions Hermès, Paris, (1991).

[19] A. O u s t a l o u p. La Robustesse. Editions Hermès, Paris (1994).

[20] A. O u s t a l o u p, F. L e v r o n, B. M a t h i e u, and F. M. N a n o t.
Frequency-band complex noninteger differenciator: Characterization and sys-
thesis. IEEE Trans. on Circuits & Systems I: Fundamental Theory and Ap-
plications, 47, No. 1 (2000), 25–39.

[21] I. P e t r a s and L. D o r c a k. Some possibilities of realization of

fractional-order controllers. Envirautom, 4, No. 1 (1999), 83–90.
 18 B. M. Vinagre, I. Podlubny, A. Hernández, V. Feliu

[22] I. P o d l u b n y. Numerical solution of ordinary fractional differential

equations by the fractional difference method. In S. Elaydi, I. Gyori, and
G. Ladas, (eds.), Advances in Difference Equations, Gordon and Breach,
Amsterdam (1997), 507–516.
[23] I. P o d l u b n y. Fractional-order systems and PIλ Dµ -controllers. IEEE
Transactions on Automatic Control, 44, No. 1 (1999), 208–214.
[24] I. P o d l u b n y. Fractional Differential Equations. Academic Press, San
Diego (1999).
[25] W. H. P r e s s, S. A. T e u k o l s k y, W. T. V e t t e r l i n g, and
B. P. F l a n n e r y. Numerical Recipes in C. The Art of Scientific
Computing, 2nd edition, Cambridge University Press (1992).
[26] S. C. D. R o y. On the realization of a constant-argument immitance or
fractional operator. IEEE Transactions on Circuit Theory, 14, No. 3 (1967),
264–274.
[27] S. G. S a m k o, A. A. K i l b a s, and O. I. M a r i t c h e v. Integrals and
Derivatives of the Fractional Order and Some of Their Applications. Nauka
i Tekhnika, Minsk, 1987.
[28] S. B. S k a a r, A. N. M i c h e l, and R. K. M i l l e r. Stability of
viscoelastic control systems. IEEE Transactions on Automatic Control, 33,
No. 4 (1988), 348–357.
[29] B. V i n a g r e, V. F e l i u, and J. F e l i u. Frequency domain identification
of a flexible structure with piezoelectric actuators using irrational transfer
function models. In Proceedings of the 37th IEEE Conference on Decision
and Control, Volume 2, Tampa, Florida, USA (1998), 1278–1280.

1)Industrial Engineering School Received: Month, Date, Year

University of Extremadura
06071 Badajoz – SPAIN
e-mail: bvinagre@unex.es
2)Dept. of Informatics and Control Engineering
Technical University of Kosice
04200 Kosice – SLOVAK REPUBLIC
e-mail: podlbn@ccsun.tuke.sk
3)Industrial Engineering School
University of Castilla-La Mancha
13071 Ciudad Real – SPAIN
e-mail: vfeliu@ind-cr.uclm.es

View publication stats