Beruflich Dokumente
Kultur Dokumente
Discretization Schemes for Fractional-Order in a proper range of frequencies of practical interest [18], [16]. More-
Differentiators and Integrators over, the fractional order can be a complex number as discussed in [18].
In this paper, we focus on the case when the fractional order is a real
Yang Quan Chen and Kevin L. Moore number.
The key step in digital implementation of a FOC is the numerical
evaluation or discretization of the fractional-order differentiator. In
Abstract—For fractional-order differentiator where is a real
number, its discretization is a key step in digital implementation. Two general, there are two discretization methods: direct discretization
discretization methods are presented. The first scheme is a direct re- and indirect discretization. In indirect discretization methods [18],
cursive discretization of the Tustin operator. The second one is a direct two steps are required, i.e., frequency domain fitting in continuous
discretization method using the Al-Alaoui operator via continued fraction time domain first and then discretizing the fit s-transfer function.
expansion (CFE). The approximate discretization is minimum phase and Other frequency-domain fitting methods can also be used but without
stable. Detailed discretization procedures and short MATLAB scripts are
given. Examples are included for illustration. guaranteeing the stable minimum-phase discretization. Existing direct
discretization methods, e.g., [19], [20], applied the direct power
Index Terms—Al-Alaoui operator, discretization, fractional differen-
tiator, fractional-order differentiator, fractional-order dynamic systems,
series expansion (PSE) of the Euler operator, continuous fractional
recursive, Tustin operator. expansion (CFE) of the Tustin operator and numerical integration
based method.
In this brief, we first focus on the direct discretization method using
I. INTRODUCTION the well known Tustin operator which is a straightforward scheme to
Fractional calculus is a 300-year-old topic. The theory of frac- discretize the fractional-order derivative [19], [20]. The major contribu-
tional-order derivative was developed mainly in the 19th century. tion of this brief is to introduce a recursive formula for discretization
Recent books [1]–[4] provide a good source of references on fractional with different order of approximation which simplifies the program-
calculus. However, applying fractional-order calculus to dynamic ming efforts. Moreover, the discretized transfer function is stable and
systems control is just a recent focus of interest [5]–[9]. For pioneering minimum phase. However, as pointed out in [21], [22], the Tustin oper-
work on this regard, we cite [10]–[13]. ator based discretization scheme exhibits large errors in high frequency
In theory, the control systems can include both the fractional order range. A new mixed scheme of Euler and Tustin operators is proposed
dynamic system or plant to be controlled and the fractional-order con- which yields the Al-Alaoui operator [21]. Using the continued fraction
troller. However, in control practice it is more common to consider the expansion of the Al-Alaoui operator, this paper contributes a new di-
fractional-order controller. This is due to the fact that the plant model rect discretization scheme with a very good magnitude fit to that of the
may have already been obtained as an integer order model in the clas- original continuous fractional differentiator.
sical sense. In most cases, our objective is to apply fractional-order This brief is organized as follows. In Section II, we present a new re-
control (FOC) to enhance the system control performance. For ex- cursive formula for Tustin discretization of the fractional-order deriva-
ample, as in the CRONE1 control [14], [7], [8], fractal robustness is tive with a symbolic validation and an illustrative example. Section III
pursued. The desired frequency template leads to fractional transmit- presents another new direct discretization scheme by continued frac-
tance [15], [16] on which the CRONE controller synthesis is based. tion expansion of the Al-Alaoui operator with an illustrative example,
In the CRONE controller, the major ingredient is the fractional-order too. Section IV concludes this paper.
derivative sr , where r is a real number and s is the Laplacian operator.
Another example is the PI D controller [6], [17], an extension of PID II. RECURSIVE TUSTIN DISCRETIZATION OF FRACTIONAL-ORDER
controller. In general form, the transfer function of PI D is given by DERIVATIVE
Kp + Ti s0 + Td s , where and are positive real numbers; Kp is A. Generating Function
the proportional gain, Ti the integration constant and Td the differenti-
ation constant. Clearly, taking = 1 and = 1, we obtain a classical In general, the discretization of the fractional-order differentiator sr
PID controller. If = 0 (Ti = 0) we obtain a PD controller, etc. All (r is a real number) can be expressed by the so-called generating func-
these types of controllers are particular cases of the PI D controller. tion s = ! (z 01 ). This generating function and its expansion determine
It can be expected that the PI D controller may enhance the systems both the form of the approximation and the coefficients [23]. For ex-
ample, when a backward difference rule is used, i.e., ! (z 01 ) = (1 0
z 01 )=T , performing the power series expansion (PSE) of (1 0 z 01 )6r
control performance due to more tuning knobs introduced. Actually, in
theory, PI D itself is an infinite dimensional linear filter due to the
fractional order in the differentiator or integrator. It should be pointed gives the discretization formula.
out that a band-limit implementation of FOC is important in practice, In this brief, we first consider the trapezoidal (Tustin) rule as a gen-
i.e., the finite-dimensional approximation of the FOC should be done erating function
!(z 01 )
6r = 2 1 0 z01 6r
1 + z 01
Manuscript received January 26, 2001; revised September 23, 2001 and (1)
November 12, 2001. This work was supported in part by U.S. Army Automo- T
tive and Armaments Command (TACOM) Intelligent Mobility Program under
Agreement DAAE07-95-3-0023. This paper was recommended by Associate for obtaining the coefficients and the form of the approximation.
Editor P. K. Rajan.
The authors are with the Center for Self-Organizing and Intelligent Systems
B. Recursive Tustin Discretization
(CSOIS), Department of Electrical and Computer Engineering, College of
Engineering, Utah State University, Logan, UT 84322-4160 USA (e-mail: The key point of the Tustin discretization of fractional-order
yqchen@ieee.org). differentiator is how to get a recursive formula. Here we introduce
Publisher Item Identifier S 1057-7122(02)02269-9.
the so-called Muir-recursion originally used in geophysical data
1CRONE is a French abbreviation for “Contrôle Robuste d’Ordre Non Entier” processing with applications to petroleum prospecting [24]. The
(which means noninteger order robust control). Muir-recursion motivated in computing the vertical plane wave
TABLE I
TABLE OF FORMULAS A (z ; r ) FOR n = 1; . . . ; 9
+ r + 2 r4
4 2
z 04 0 2
r + 4 r3 + 4
r5 z 05
3 3 5 3 15
+ r + 8 r4 + 4 r6
46 2
z 06
45 9 45
0 2
r + 56 r3 + 4 r5 + 8
r 7 z 07
7 45 9 315
+
88
r2 + 44 r4 + 8
r6 + 2
r 8 z 08
105 45 45 315
0 2
r + 3272 r3 + 76
r5 + 8
r7 + 4
r 9 z 09 : (5)
9 2835 135 135 2835
0 z 01 G1 (z ) = 44:72z 0 22:36
r
!(z01 )
r
r 2 1
1 + z 01
=
T z + 0:5
2
r
A (z 01 ; r)n G3 (z ) = 44:72 z 3 0 22:36z 2 + 3:727z 0 7:454
=
T !1 A (z 01 ; 0r)
lim
n n
(2) z + 0:5z 2 + 0:083 33z + 0:1667
3
A (z 01 ; r) = A 01 (z 01 ; r) 0 c z A 01 (z; r);
n n n
n
n (3) 44:72z 0 22:36z + 4:969z 0 8:075z
9 8 7 6
sr
T2
r
A (z 01 ; r) :
n We present four plots as shown in Fig. 1 to show the effectiveness of
A (z 01 ; 0r)
n the approximate discretization.
It should be pointed out that the direct discretization method intro-
Symbolically, Table I lists the expressions of A (z 01 ; r)
n for n = duced above always gives a Z -transfer function with stable minimum
1; . . . ; 9. phase characteristics. The convergence of the recursive scheme [24] is
clearly demonstrated by Fig. 1.
C. Symbolic Validation
III. DIRECT DISCRETIZATION USING CONTINUED FRACTION
We now examine the correctness of the recursive discretization of
EXPANSION (CFE) OF AL-ALAOUI OPERATOR
the fractional-order derivative operator by using MATLAB Symbolic
Toolbox. We can get the symbolic Taylor expansion for (1) for the first We see from Fig. 1 that the Bode plot of the discretized model based
9 terms. The result is summarized as follows: on the recursive expansion of the Tustin operator or trapezoidal rule
is with large error in the high frequency range. To have a better ap-
1 0 z 01 r proximation in the high frequency range, here we use the Al-Alaoui
1 + z 01
operator [21] which is a weighted sum of the rectangular rule or Euler
operator and the trapezoidal rule. This will give a new straightforward
1 0 (2r)z01 + (2r2 )z02 0 2
3
r + 4 r 3 z 03
3
scheme using continued fraction expansion (CFE) to discretize the frac-
tional-order derivative with the aid of a symbolic computation tool.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—I: FUNDAMENTAL THEORY AND APPLICATIONS, VOL. 49, NO. 3, MARCH 2002 365
A. Al-Alaoui Operator Based Discretization where CFEfug denotes the continued fraction expansion of u; p and
The Al-Alaoui operator is a mixed scheme of Euler and Tustin op- q are the orders of the approximation and P and Q are polynomials of
erators. Based on the Al-Alaoui operator, the generating function for degrees p and q . Normally, we can set p = q = n. In MATLAB Sym-
discretization is bolic Toolbox, we can easily get the approximate direct discretization
of the fractional order derivative by the following script, for a given n
6r (replace 14 by 2n):
8 1 0 z 01
!(z01 ) 6 = r
:
7T 1 + z 01 =7
(6)
clear all; close all; syms x z r
%Al 0 Alouoi0 s scheme
Clearly, (6) is an infinite order of rational discrete-time transfer func-
tion. To approximate it with a finite order rational one, continued frac- x = ((1 0 z)=(1 + z=7))^r;
tion expansion (CFE) is an efficient way. In general, any function G(z ) [RESULT; STATUS] = maple (\with (numtheory)")
can be represented by continued fractions in the form of %7-th order; put 2 3 7 here:
h7 = maple (\cfrac"; x; z; 14);
G(z ) ' a0 (z ) + b1 (z )
h7n = maple (\nthnumer (%%; 14)");
a1 (z ) + b2 (z )
h7d = maple (\nthdenom (%%%; 14)");
a2 (z ) + b3 (z )
a3 (z ) + 1 1 1 h7ns = sym (h7n); h7ds = (h7d);
num7 = collect (h7ns; z); den7 = collect (h7ds; z);
where the coefficients ai and bi are either rational functions of the vari- fn7 = subs (num7; z; 1=z); fd7 = subs (den7; z; 1=z):
able z or constants. By truncation, an approximate rational function,
G^ (z ), can be obtained. Remark III.1: The CFE scheme presented in the above contains
The resulting discrete transfer function, approximating fractional- two tuning parameters, namely p and q , while in the recursive Tustin
order operators, can be expressed as: method introduced in Section II, there is only one tuning parameter, n.
The optimal choice of these parameters is possible based on a quanti-
6r 1 0 z 01
6r
D6 (z )
r 8
CFE
tative measure. One possibility is the use of the least squares (LS) error
7T 1 + z 01 =7 between the continuous-frequency response and discretized frequency
response. Note that in practice, p and q can usually be set equal. So,
p; q
8
6r Pp (z 01 ) in the discretization methods presented in this paper, there is only one
=
Qq (z 01 )
(7)
7T tuning knob which will be more attractive in practice. Compared to
366 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—I: FUNDAMENTAL THEORY AND APPLICATIONS, VOL. 49, NO. 3, MARCH 2002
the frequency-banded fitting methods, e.g., [18], [16], the direct dis- linear phase drop, a half sample phase advance is used which means
cretization methods presented in this paper is straightforward and no that we should cascade z 0:5r to the obtained approximately discretized
band-pass weighting filter is assumed. transfer function G(z ). In this example, the phase compensator is z 0:25
which is noncausal. In implementation, we can simply use z 00:75 =z 01
B. An Illustrative Example instead. Note that the linear phase drop compensation does not affect
the magnitude fit. When multi-rate sampling is possible, the phase-drop
The discretization of the half-differentiator s0:5 sampled at 0.001 s. compensation can be implemented with a smaller error which deserves
is studied numerically, and the approximate models are further investigation. It should also be pointed out that, by checking
the pole-zero map, the direct discretization method introduced above
0 169
236:6z always gives a Z -transfer function with stable minimum phase charac-
G1 (z ) =
017z teristics.
49 0 49 2 + 7 + 1
G3 (z )= 3 z z z IV. CONCLUDING REMARKS
2 47 04 0 5 999 04 +4 941 04
5 4 3
: e z : e z : e z This brief presents two discretization methods for the fractional-
01 512 04 +956 9 +98 48
2
: e z : z : order differentiator sr where r is a real number. The discretization is
730 7 5 0 1357 4 +745 7 3 0 89 48 2 0 15 52 +1
G5 (z )=
: z z : z : z : z a key step in digital implementation of the fractional-order controller
3 128 05 0 1 028 06 +1 283 06 0 7 433 05
7 6 5 4 containing sr . The first scheme is a direct recursive discretization of
: e z : e z : e z : e z
Manuscript received March 31, 2001; revised October 1, 2001. This paper
was recommended by Associate Editor G. Palumbo.
The authors are with the Electronic Department of the Politecnico di Torino,
Turin 10129, Italy (e-mail: fiori@eln.polito.it).
Publisher Item Identifier S 1057-7122(02)02276-6.