IEEE/CAA JOURNAL OF AUTOMATICA SINICA 1

Design and Implementation of Digital Fractional

Order PID Controller using Optimal Pole-Zero
Approximation Method for Magnetic
Levitation System
Amit S. Chopade, Swapnil W. Khubalkar, A. S. Junghare, M. V. Aware, and Shantanu Das

Abstract—The aim of this paper is to employ fractional order applications such as magnetically levitated train, magnetic
proportional integral derivative (FO-PID) controller and integer bearing, conveyor system, etc.[1] .
order PID controller to control the position of the levitated object In recent years, various methods have been proposed to
in a magnetic levitation system (MLS), which is inherently non-
linear and unstable system. The proposal is to deploy discrete improve control in MLS-based applications. In 2006, Chiang
optimal pole-zero approximation method for realization of digital et al. proposed the concept of integral variable-structure grey
fractional order controller. An approach of phase shaping by control[2] . Yang et al. introduced the concept of adaptive robust
slope cancellation of asymptotic phase plots for zeros and poles output-feedback control with K-filter approach in 2008[3] .
within given bandwidth is explored. The controller parameters In 2011, Chih-Min Lin et al. developed an adaptive PID
are tuned using dynamic particle swarm optimization (dPSO)
technique. Effectiveness of the proposed control scheme is verified controller and a fuzzy compensation controller for MLS[1] .
by simulation and experimental results. The performance of In the same year, Rafael Morales et al. proposed generalized
realized digital FO-PID controller has been compared with that proportional integral output feedback controller[4] . Recently in
of the integer order PID controllers. It is observed that effort 2014, Chih-Min Lin et al. proposed a function-link cerebellar
required in fractional order control is smaller as compared model articulation control system design based on the neural
with its integer counterpart for obtaining the same system
performance. network concept[5] . However, in spite of all these develop-
ments, there is scope for improving efficiency of the controller.
Index Terms—Digital control, Position control, Fractional The energy required to achieve and maintain the object’s
calculus, Particle swarm optimization (PSO), Approximation
methods, Magnetic levitation, Discretization, Fractional order position (in the face of disturbances) form an important part
PID controller (FOPID). of improving the control action. The aim of this paper is to
control and maintain the desired object’s position, with lesser
controller effort. The controller effort minimization is reported
I. I NTRODUCTION in literature[11−14] .
The conventional integer order controllers such as, PD and
I N 1914, American inventor Emile Bachelet presented his
idea of a magnetically levitated (maglev) vehicle with a dis-
play model. In magnetic levitation system (MLS), ferromag-
PID controller have been applied in industry for over half-a-
century to control linear and nonlinear systems[15] . Recently,
netic object levitate by the electromagnetic force induced due such control schemes have been extended to their generalized
to electric current flowing through coil around a solenoid[1−5] . form using fractional calculus[16−17] (differentiation and inte-
The MLS is inherently nonlinear and unstable[6−10] . However, gration of an arbitrary order). The FO-PID controller has frac-
the advantage is that, as the suspended object has no mechan- tional differ-integrator operations. This makes the controller
ical support, there is no friction and noise. This allows us to have memory (i.e. its action will memorize its past states)
position it accurately - a major advantage, explored in many and avoids instantaneous actions. Using the definition of con-
volution integral, the expression for the fractional integration
This article has been accepted for publication in a future issue of this (which also is embedded in the fractional differentiation) can
journal, but has not been fully edited. Content may change prior to final be written as the convolution of the function and the power
publication. function, which is elaborately explained in [17].
This work was supported by the Board of Research in Nuclear sciences
of the Department of Atomic Energy, India. Sanction no. 2012/36/69- In last few decades, the fractional order approach to rep-
BRNS/2012. Recommended by Associate Editor Antonio Visioli. resent the plant and its controllers are increasingly used to
Amit S. Chopade, Swapnil W. Khubalkar, Anjali S. Junghare, describe the dynamic process accurately[17] . The fractional or-
and Mohan V. Aware are with the Department of Electrical En-
gineering, Visvesvaraya National Institute of Technology, Nagpur, In- der transfer function is approximated by integer order transfer
dia 440010 e-mail: ([Link]@[Link], [Link]@[Link], function using various methods[16−20] . The proposed method
asjunghare@[Link], mvaware@[Link]). can achieve the desired accuracy over a much larger bandwidth
Shantanu Das is with Reactor Control Division, Bhabha Atomic Research
Centre, India. e-mail: (shantanu@[Link]). than has been achieved using earlier methods. In applications,
Digital Object Identifier 10.1109/JAS.2016.7510181 where non-integer order controllers are used for integer order
2 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

plant, there is more flexibility in adjusting the gain and phase

characteristics as compared to integer order controllers. This i = k1 u (2)
flexibility makes fractional order control a more versatile tool
where, k is a constant depending on coil (electromagnet)
in designing robust and precise control systems.
parameters, m is the mass of sphere, g is gravitational force,
This paper presents the control of magnetic levitation sys-
k1 is an input conductance, u is control voltage, and x is
tem using FO-PID controller based on optimal pole-zero ap-
a ball position. The values of these parameters are given in
proximation method. An algorithm is developed to realize digi-
Appendix-A. A relation between control voltage x and coil
tal FO-differentiators and FO-integrators. The proposed design
current i is given in (2). The control signal ranges between
procedure aims to ensure that the performance is within re-
[−5V, +5V ].
quired tolerance bandwidth. Five parameters (kp , ki , kd , α, β)
of FO-PID need to be tuned for designing the controller. This
paper utilizes dynamic PSO optimization (dPSO) method to A. Linearization of MLS Model
achieve the required values. Finally, a comparative study of The nonlinear form of maglev model is linearized for
the performance parameters of the controller is presented to analysis of the system[21] . The linear form of the model is
evaluate the advantages of deployment of FO-PID against obtained from (1) as follows:
the conventionally used integer-order controllers. The control
effort minimization by FO-PID controller is quantified and ẍ = g − f (x, i) (3)
demonstrated. i
where, f (x, i) = k mx
2
2
This work is organized as follows: section II presents the Equilibrium point is calculated by setting ẍ = 0,
system description. Design procedure of proposed digital FO-
PID controller using discrete optimal pole-zero approximation g = f (x, i)|io ,xo (4)
method and dPSO technique is discussed in section III. In Linearization is carried out about the equilibrium point of
section IV simulation and experimental results on MLS are xo = −1.5V (the position is expressed in volts), io = 0.8A[8] .
provided to validate effectiveness of the proposed controller. Using series expansion method, (5) is obtained.
Paper concludes with a summary of the results obtained in
section V. ³ ∂f (i, x) ¯
¯ ∂f (i, x) ¯¯ ´
ẍ = − ¯ 4i + ¯ 4x (5)
∂i io ,xo ∂x io ,xo
II. S YSTEM I DENTIFICATION OF MLS M ODEL Application of Laplace Transform on (5) simplifies it to (6).
A laboratory scale magnetic levitation system is used to 4X(s) −Ki
evaluate the performance of proposed controller in a controlled = 2 (6)
4I(s) s + Kx
environment. MLS levitates an object (metallic ball with mass
2mg
m) in a desired position by controlling the electromagnetic where, Ki = io and Kx = − 2mg
xo
field counteracting the gravitational force. The applied control
input is voltage, which is converted into current via embedded Linearized model transfer function p (6) has two poles, one of
driver[21] . Fig.1 shows the schematic diagram of MLS. The which is in the right half plane at (2mg/x0 ), which makes
system model is nonlinear, that means at least one of the two the MLS open-loop unstable. Transfer function, obtained by
states (i–current, x–ball position) is an argument of a non- the linearization, is verified using system identification proce-
linear function. The nonlinear model of MLS relating the ball dure.
position x and coil current i is given as (1):
B. Integer Order System Identification of MLS Model
System identification is a process for obtaining mathe-
matical model using input and output system response. The
identified model response should fit with measured response
for input applied to the system model[21] . Usually there are two
methods for system identification, least mean square (LMS)
method and instrumental variable method. The identification
of MLS is generally accomplished via traditional least squares
method, and is implemented in MATLAB[21−22] .
As MLS is unstable, it has to be identified with a run-
ning, stabilizing controller i.e. closed loop identification. Fig.2
shows the scheme of unstable system identification. LMS
method minimizes error between the model and plant output.
Fig. 1. Schematic diagram of MLS.
The optimal model parameters, for which the square of the
error is minimal is the result of identification. In order to
i2 carry out identification experiment, a discrete controller has
mẍ = mg − k (1) to be applied, in the absence of which, the ball falls down,
x2
CHOPADE et al.: DESIGN AND IMPLEMENTATION OF DIGITAL FRACTIONAL ORDER PID CONTROLLER · · · 3

rendering dentification impossible. The reference signal r(t) 1) : Riemann-Liouville (RL) definition is given in (9).
i.e. random binary sequence signal is given to excite the MLS
and output y(t) is monitored. 2500 samples of the input, output à !n Z
t
α 1 d f (τ )
signals are collected from the system with sampling period of a Dt f (t) = dτ (9)
0.01s. Γ(n − α) dt a (t − τ )α−n+1

for (n − 1) ≤ α < n
where, n is an integer, α is a real number, and Γ is Euler
gamma function. Laplace transform of the RL fractional
derivative/integral (9), under zero initial conditions, is given
in (10).

L{a Dt±α f (t)} = s±α F (s) (10)

Fig. 2. Block diagram of MLS control and close loop system
2) : Another definition is based on the concept of fractional
identification.
differentiation i.e. Grunewald-Letnikov (GL) definition. It is
Fig.3 presents the comparison between measured and iden- given in (11).
tified model output. Input and output data is taken from MLS £ ¤ µ ¶
system for real-time identification. The best fit obtained is α −α
t−a
α
90.78% for integer order identification, which gives close loop a Dt f (t) = lim h Σj=0 (−1)j
h
f (t − jh) (11)
h→0 j
discrete transfer function as in (7): £ ¤
where, t−a h −→ Integer
3) : One more option for computing fractional derivatives
is Caputo fractional derivative, its definition is as follows (12):
Z t
C α 1 f n (τ )
a Dt f (t) = dτ (12)
Γ(n − α) a (t − τ )α+1−n
where, (n − 1) ≤ α < n, n is an integer, and α is a real
number.
Initial conditions for Caputo’s derivatives are expressed in
terms of initial values of integer order derivatives. It is noted
Fig. 3. Measured and simulated model output.
that for zero initial conditions RL, GL, and Caputo fractional
derivatives coincide. Hence, any of the mentioned methods
G(z −1 ) may be used, using the case of zero initial conditions. That
Y (z −1 ) = (7) would then eliminate the differences arising due to different
1 + C(z −1 )G(z −1 )
initial conditions (amongst the three methods).
where, Y (z −1 ) is complete system transfer function, C(z −1 )
is controller transfer function, and G(z −1 ) is MLS model
B. Digital Realization of Fractional Order Differintegrals with
transfer function in discrete domain.
Optimal Pole-Zero for Phase Shaping
The aim behind the choice of frequency domain rational
III. DESIGN OF DIGITAL FRACTIONAL ORDER PID approximation of FO-PID controller is to realize the controller
CONTROLLER in real time using existing analog/digital filters[16−20, 23−25] .
A. Fractional Calculus Precise hardware implementation of multi-dimensional na-
tured of fractional order operator is difficult. However, re-
Fractional calculus is a branch of mathematics that studies
cent research work revealed that band-limited implementation
the possibility of taking real or complex number powers of
of FO-PID controllers using higher order integer transfer
differential and integral operator. Basic definitions of fractional
function approximation of the differintegrals give satisfactory
calculus and approximation of fractional integrator and frac-
performance[26] . This paper, hence utilizes optimal pole-zero
tional differentiator are described in the literature[16−17] . The
algorithm to realize fractional differintegrals in the frequency
real order operator is generalized as follows in (8):
domain.
 dα

 α>0  1) Optimal pole-zero approximation for phase shaping:
dt
Dα = 1 α=0 (8) Any rational transfer function is characterized by its poles and
 Rt −α  zeros. The Bode magnitude plot of non-integer order transfer
a
(dτ ) α < 0
function has a slope of ±α20 dB/dec and the Bode phase
where, α ∈ R plot lies in the range of ±α90o (α is a real number). This is
achieved by the interlacing of real poles and zeros alternately
Some popular definitions used for general fractional deriva- on the negative real axis[19−20, 27−28] . Thus, depending on
tives/integrals in fractional calculus are : the error band ² around required phase angle αreq = α90o
4 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

and the frequency band of interest (ωL , ωH ), the nth order number of pole-zero pairs. Fig.6 shows the basic idea of
approximation is obtained[28] . The proposed algorithm is de- frequency band tightening. The problem regarding frequency
veloped to obtain the number of optimal pole-zero pairs to band shrinking is tackled by designing the rational approx-
ωl
maintain the phase value within the tolerance, of around 1o . imation on wider frequency band ( 10 δ
γ , 10 ωh ) followed by

In this algorithm, poles and zeros given by (13) are obtained curtailing the frequency overhang on either side, such that the
as follows: phase remains within φreq ± ε in range of (ωl , ωh ). Nominal
φreq +45logωl values to expand frequency band are γ = 3, δ = 2.
F irst pole, p1 = 10[ 45 +1]

F irst zero, z1 = 10ωl

Second pole, p2 = 10[log(p1 )+2−µ]
(13)
Second zero, z2 = 10[log(z1 )+2−µ]
..
.
till pn ≥ ωh
As a particular case, asymptotic phase plot for fractional
order integrator circuit having α = −0.4, φreq = −36o ,
ωL = 0.1rad/s and ωH = 100rad/s is given in Fig.4 - Fig.6.
The selection of three pairs of poles and zeros with α = −0.4
fraction is shown in Fig.4. The asymptotic phase plot is a
straight line at φreq , but the actual phase plot is oscillating Fig. 5. Asymptotic phase plot with seven pole-zero pairs.
about asymptotic phase plot, apart from that the average value
of phase angle −37.31o is also different from φreq . In Fig.4 the
required correction of phase is achieved over three decades by
three pole-zero pairs only, which is however less in pursuit
of more accuracy. This problem is rectified by increasing
the pole-zero density, i.e. having more pole-zero pairs in the
desired frequency band. Number of pole-zero pairs depend on
the permissible error and the desired band of frequency.

Fig. 6. The basic idea of frequency band tightening.

2) Design of Digital Fractional Order Integrator: The key

point in digital implementation of fractional order controller
is discretization of fractional order differintegral[24, 27−29] .
Contributions related to the discretization have been reported
in literature[30−33] . The pole-zero pairs obtained by algo-
rithm in the above case are discretized using first order hold
Fig. 4. Asymptotic phase plot with three pole-zero pairs for α = (foh), zero order hold (zoh), Tustin operator, impulse invari-
−0.4(−36o ). ant, matched pole-zero, and Tustin with pre-warp frequency
methods. In Fig.7, Bode plot for s−0.4 digital integrator
Generally, three pole-zero pairs per decade give the phase
is shown and it depicts that digital integrator with Tustin
plot within ε = ±1o error, but it depends on the value of α
approximation method matches closely with continuous time
as well. For the same parameters, i.e. α = −0.4, φreq = 36o ,
integrator. Tustin approximation method with a sample time
ωL = 0.1rad/s and ωH = 100 rad/s with seven pole-zero
of 0.01s is used for discretization. To relate s-domain and z-
pairs, phase plot is shown in Fig.5. The actual phase plot is
domain transfer functions, Tustin and bilinear methods use the
oscillating with rms error of 0.6471o (< 1o ). Apart from that,
following approximation as (14).
average value of phase angle (−35.9999o ≈ −36o ) is same as
φreq . Moreover, this is achieved over 3 decades of cycle by 1 + sTs /2
seven pole-zero pairs. In order to maintain the phase margin z = esTs ≈ (14)
1 − sTs /2
tolerance within the lower limits, more pole-zero pairs in the
desired frequency band are required. This can be done by The optimal pole-zero algorithm for digital fractional in-
adjusting z1 , p2 , z3 .... closer towards left. To achieve this shift, tegrator of s−0.4 within desired band of frequency ωL =
design parameter µ is introduced. Frequency band of constant 0.1rad/s and ωH = 100rad/s gives pole-zero pairs which are
phase shrinks on both the ends with increasing µ for constant listed in Table 1 with gain value 0.010127.
CHOPADE et al.: DESIGN AND IMPLEMENTATION OF DIGITAL FRACTIONAL ORDER PID CONTROLLER · · · 5

a particle and its individual best or the global best. Here,

the change in position of a particle is directly proportional
to iteration, which further depends on individual best, global
best, and a random velocity[49] . dPSO searches the workspace
similar to a simple PSO and velocity of a particle is obtained
by (15):
vid = (f (pid ) − f (xid )) × (pid − xid ) × sf1
+ (f (pgd ) − f (xid )) × (pgd − xid ) × sf2
+ rand() × signis() × sf3 (15)
where, vid : velocity of a particle, pid : individual best, xid :
current position of a particle, pgd : global best, rand: random
function, sf1 , sf2 , sf3 : to scale the calculated value in the
Fig. 7. Bode plot of s−0.4 digital integrator. range of the control variable, signis: function which generates
random positive or negative value.
Table I Population size is taken as 100, maximum iteration is set
THE POLE-ZERO PAIRS OF THE RATIONAL as 50, lower and higher translation frequencies are taken as
APPROXIMATION OF s−0.4 ON (10−1 , 102 )rad/s ωL = 0.1rad/s and ωH = 100rad/s. ITAE (Integral Time Ab-
solute Error) is chosen as performance criterion. The values of
i 1 2 3 4 5 6 7
controller parameters, obtained from dPSO, are implemented
zi −0.9253 −0.4842 0.9307 0.992 0.9991 1 0.5136
in PD, PID, and FO-PID controller in simulation as well
pi −0.8286 0.7651 0.9707 0.9967 1 1 −0.0875
as in real time mode on MLS. The optimized values of the
Digital fractional differentiator is designed along the lines controllers are presented in Table II.
of approach similar to that of digital fractional integrator. The Table II
architecture of digital FO-PID with digital fractional integrator dPSO OPTIMIZED GAIN AND FRACTIONAL ORDER VALUES
and digital fractional differentiator is shown in Fig.8. USED FOR DIFFERENT CONTROLLERS ( α :ORDER OF
INTEGRATOR, β :ORDER OF DIFFERENTIATOR)

Gain and Fractional Order Value

Sr. No. Controllers
Kp Ki Kd α β
1. PD 4 – 2 – 1
2. PID 5.5 2 0.2 1 1
3. FOPID 7 12 1 0.4 0.8

IV. MLS CONTROL: SIMULATION AND

HARDWARE
A. Closed-Loop Control System Simulation
Fig. 8. Digital FO-PID controller. Control of MLS using optimized PD, PID, and FO-PID con-
troller is studied by MATLAB simulation. A sinusoidal excita-
3) dynamic Particle Swarm Optimization: Recently, many tion signal is used to study the effects. The controller generates
researchers have focused on fractional order controllers tuning, a compensating control signal (based on the positional error) to
and have obtained meaningful results[34−47] . In this work, achieve desired ball position. Controller parameters are tuned
dPSO method is used to tune the gains and orders of the using dPSO method as discussed in section III-B-3. Fig.9 -
controller. PSO is a method for optimizing hard numerical Fig.11 present simulation results of the controlled output of
functions, analogous to social behavior of flocks of birds, MLS using PD, PID, and FO-PID controller respectively. Here,
schools of fish, etc. Here, each particle in swarm represents encircled part pointed by an arrow shows deviation between
a solution to the problem defined by its instantaneous po- desired and actual ball position.
sition and velocity[48] . The position vector of each particle
is represented by unknown parameters to be ascertained. In
present case, five control parameters (kp , ki , kd , α, β) of FO-
PID controller need to be ascertained. The desired number of
particles is known as population. The population is varied to
carry out a search in multidimensional space. Each particle in
population will travel with the updated velocity and direction
to converge as early as possible to the optimal solution point.
Dynamic PSO is an improvement in PSO by adding the
product of differences in objective function value between Fig. 9. Controlled output result of MLS using PD.
6 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

are used to implement control algorithm. It includes MATLAB

control toolbox, real time windows workshop (RTW), real
time windows target (RTWT), and visual C as programming
environment. The flowchart required to obtain executable file
is shown in Fig.13. RTW builds a C++ source code from
the Simulink Model. C code compiler compiles and links the
code to produce executable program. RTWT communicates
with executable program acting as the control program and
Fig. 10. Controlled output result of MLS using PID. interfaces with hardware through input/output board. The
block diagram of MLS close loop control is shown in Fig.14.
The measured and desired ball positions with PD, PID,
and FO-PID controllers are quantitatively presented in Table
III. The simulation results indicate that deviation between
measured and desired ball positions by using dPSO tuned FO-
PID controller, is less as compared to PD or PID controllers.

Fig. 12. Experimental setup.

Fig. 11. Controlled output result of MLS using FO-PID.

Error values presented in Table III are calculated using (16): 1) Experimental Results using a PD Controller: The mea-
sured and desired ball positions using real time PD controller
is shown in Fig.15(a) and control signal c(t) before digital to
P ercent error = analog (D/A) conversion is given in Fig.15(b). This control
desired position − actual position signal is used to levitate the object at desired position. The
× 100%
actual position plant input signal m(t) after D/A conversion and output signal
(16) y(t), captured on the digital storage oscilloscope (DSO), is
From the data presented in Table III, it is observed that FO- presented in Fig.16.
PID controller tracks the desired position more efficiently than The control effort required by controller to maintain object’s
PD or PID controllers. position can be observed from the control signal c(t). The
ball position is tracked by infrared sensor and is fed back to
B. Real Time Implementation of Closed-Loop System Simulink environment via analog to digital (A/D) converter. It
The MLS used for experimentation is shown in Fig.12. Due is observed from Fig.15 - Fig.16 that there is more deviation
to high nonlinearity and open-loop instability, MLS system in ball position and control effort required by the controller,
is a very challenging plant. Assembly of MLS consists of a and is higher in case of PD controller. Hence, integral action is
mechanical unit labeled A in Fig.12. Analogue control inter- added to the PD controller to achieve an improved control over
face unit labeled A is used to transfer control signals between desired ball position. The quantitative analysis of desired and
computing system and MLS. Advanced PCI1711 I/O card has actual ball position achieved by the controller is presented in
been inserted into a PCI computer slot and connected with Table IV and the control effort analysis of controller is shown
SCSI adapter box using SCSI cable. Mathworks software tools in Table V.

Table III
MEASURED AND DESIRED BALL POSITIONS FOR DIFFERENT CONTROLLERS IN SIMULATION

Controllers
Ball Positions (m)
PD PID FO-PID
Max. 8.12 × 10−3 6.92 × 10−3 5.94 × 10−3
Measured ball position
Min. −4.83 × 10−3 −6.68 × 10−3 −5.65 × 10−3
Max. 5.5 × 10−3 5.5 × 10−3 5.5 × 10−3
Desired ball position
Min. −5.5 × 10−3 −5.5 × 10−3 −5.5 × 10−3
Error 23.06% 19.09% 5.03%
CHOPADE et al.: DESIGN AND IMPLEMENTATION OF DIGITAL FRACTIONAL ORDER PID CONTROLLER · · · 7

Fig. 13. Control system development flow diagram.

Fig. 14. Block diagram of MLS close loop control.

of PD controller while achieving the improvement.

Fig. 16. Experimental PD controller output and object’s trajectory

Fig. 15. (a) Controlled output result of MLS using a PD controller captured on DSO.
(b) Control signal of PD controller.
3) Experimental Results using a FO-PID Controller: The
2) Experimental Results using a PID Controller: Fig.17(a) deviation in ball positions using real time FO-PID controller is
shows measured and desired ball positions using PID con- shown in Fig.19(a). It depicts that error in desired and actual
troller and output of controller c(t) is shown in Fig.17(b). The ball positions has reduced in comparison to both PD or PID
captured controller output signal c(t) and output signal are control actions. The control signal c(t) of FO-PID controller
presented in Fig.18. The deviation in the ball position is mini- is presented in Fig.19(b). It shows that effort required by the
mized to an extent by employing the PID controller. However, controller is least as compared to PD or PID controllers. Plant
the control effort required by controller is still similar to that input signal m(t) and output signal y(t) are presented in Fig.20.
8 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

Table IV
MEASURED AND DESIRED BALL POSITIONS FOR DIFFERENT CONTROLLERS IN REAL TIME IMPLEMENTATION

Controllers
Ball Positions (m)
PD PID FO-PID
Max. 16.8 × 10−3 13.1 × 10−3 12.6 × 10−3
Measured ball position
Min. 8.3 × 10−3 4.85 × 10−3 5.24 × 10−3
Max. 12.5 × 10−3 12.5 × 10−3 12.5 × 10−3
Desired ball position
Min. 5.5 × 10−3 5.5 × 10−3 5.5 × 10−3
Error 29.66% 8.95% 5.75%

PID controller has improved the position accuracy of MLS

compared to PD or PID controllers in real time implementa-
tion. Also, the percentage error is least for FO-PID controller.

Fig. 17. (a) Controlled output result of MLS using a PID controller
(b) Control signal of PID controller. Fig. 20. Experimental FO-PID controller output and object’s trajec-
tory captured on DSO.

Table V
CONTROL EFFORT ANALYSIS OF DIFFERENT
CONTROLLERS IN REAL TIME IMPLEMENTATION

Controllers
Performance Indices
PD PID FO-PID
Error Signal 51.97 14.56 12.79
IAE
Control Signal 208 181 151.5
Error Signal 609 455.7 425.5
ITAE
Control Signal 900.6 797.9 602.5
Fig. 18. Experimental PID controller output and object’s trajectory Error Signal 28.38 4.978 2.488
ISE
captured on DSO. Control Signal 832.6 647.2 347.2

Fig. 21. Control effort analysis.

Fig. 19. (a) Controlled output result of MLS using a FO-PID The control effort required by PD, PID, and FO-PID
controller (b) Control signal of FO-PID controller. controllers is calculated using IAE (Integral Absolute Error),
ITAE, and ISE (Integral Square Error). The analysis has been
carried out for a period of 100s and is tabulated in Table V.
From the data presented in Table IV it is observed that FO- Fig.21 represents the control effort analysis in pictorial form.
CHOPADE et al.: DESIGN AND IMPLEMENTATION OF DIGITAL FRACTIONAL ORDER PID CONTROLLER · · · 9

The error signal is maximum in the case of PD controller

and least in the case of FO-PID controller. The control signal
also follows the same pattern and is least in case of FO-PID
controller, leading to inference that the control effort in terms
of power required by the FO-PID controller to maintain the
ball position is least amongst the three controllers.

From the analysis, it infers that PID controller is better

than PD controller through performance characteristic. FO-
PID controller shows slight improvement over PID controller,
but the effort required is appreciably less for the same im- Fig. 24. Experimental PD controller output and object’s trajectory
provement. Thus proving superiority of FO-PID over integer captured on DSO.
order controllers.

4) Disturbance Injection Analysis of Controllers: The ef-

fect of disturbance is studied by injecting step input to MLS
and effect of increased load is studied by introducing another
metallic ball in levitation system as shown in Fig.22. The step
is applied after interval of 25s on initiation of the input while
another ball is introduced manually after 35s. The measured
and desired ball positions using a PD controller are presented
in Fig.23(a) and the control signal of a controller is shown
in Fig.23(b). PD controller output and object’s trajectory as
captured on DSO is presented in Fig.24.

Fig. 25. (a) Controlled output result of MLS using a PID controller
(b) Control signal of PID controller.

Fig. 22. Levitation of two metallic balls.

Fig. 26. Experimental PID controller output and object’s trajectory

captured on DSO.

Fig. 23. (a) Controlled output result of MLS using a PD controller

(b) Control signal of PD controller.

The instant of step applied in input signal and the instant of

the addition of extra load are demonstrated by circles marked
Fig. 27. (a) Controlled output result of MLS using a FO-PID
on figures. Overshoot is observed at the instant of step and
controller (b) Control signal of FO-PID controller.
after introducing second ball in levitation system.
10 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

The deviation in ball position is higher as load is increased [2] H. K. Chiang, C. A. Chen, and M. Y. Li, “Integral variable-structure
and greater amount of effort (power consumption, as indicated grey control for magnetic levitation system,” IEE Proc. Electric Power
Applications, Vol. 153, No. 6, pp. 809–814, November 2006. doi:
by high switching fluctuations in the voltage graph) is required 10.1049/ip-epa:20060056
by controller to achieve desired ball position.
Similar analysis for PID and FO-PID controllers is pre-
[3] Z. J. Yang, K. Kunitoshi, S. Kanae, and K. Wada, “Adaptive robust
sented in Fig.25 - Fig.28. These figures lead to inference that in output-feedback control of a magnetic levitation system by K-filter
case of PID controller, the deviation in ball position is high and approach,” IEEE Trans. on Industrial Electronics, Vol. 55, No. 1, pp.
greater amount of effort is required by controller to achieve 390–399, January 2008. doi: 10.1109/TIE.2007.896488
ball position as compared to FO-PID controller. Comparison
shows that FO-PID controller requires lesser effort to levitate [4] R. Morales, V. Feliu, and H. Sira-Ramrez, “Nonlinear control for
the object and effect of disturbance is less as compared to PD magnetic levitation systems based on fast online algebraic identification
or PID controllers. of the input gain,” IEEE Trans. on Control Systems Technology, Vol. 19,
No. 4, pp. 757–771, July 2011. doi: 10.1109/TCST.2010.2057511

[5] C. M. Lin, Y. L. Liu, and H. Y. Li, “SoPC-based function-link cerebellar

model articulation control system design for magnetic ball levitation
systems,” IEEE Trans. on Industrial Electronics, Vol. 61, No. 8, pp.
4265–4273, August 2014. doi: 10.1109/TIE.2013.2288201

[6] A. El Hajjaji and M. Ouladsine, “Modeling and nonlinear control of

magnetic levitation systems,” IEEE Trans. on Industrial Electronics, Vol.
48, No. 4, pp. 831–838, August 2001. doi: 10.1109/41.937416

Fig. 28. Experimental FO-PID controller output and object’s trajec- [7] R. C. Kluever, C. A. Kluever, Dynamic Systems: Modeling, Simulation,
and Control, 1st Edition, John Wiley and Sons, April 2015.
tory captured on DSO.

[8] G. F. Franklin, J. D. Powell, and A. Emami-Naeini, Feedback Control

V. C ONCLUSION of Dynamic Systems, 3rd Edition, Addison-Wesley, Reading, MA, 1994.
In this paper, digital FO-PID controller is applied on MLS
to improve the positional accuracy and control effort. A new
discrete optimal pole-zero approximation method is proposed [9] T. H. Wong, “Design of a magnetic levitation control system - An
for realization of controller. This method provides the optimal undergraduate project,” IEEE Trans. on Education, Vol. 29, No. 4, pp.
number of pole-zero pairs to maintain the phase value within 196–200, November 1986. doi: 10.1109/TE.1986.5570565
the tolerance, of around 1o . dPSO method is used for tuning
the parameters of controller. The performance analysis for [10] R. Sinha and M. L. Nagurka, “Analog and labview-based control of
integer and fractional order controllers have been carried out a maglev system with NI-ELVIS,” ASME International Mechanical
Engineering Congress and Exposition, Orlando, Florida, USA, pp. 741–
in both simulation and experimentation. The results show that 746, November 2005.
a better control over position accuracy with lesser efforts (over
conventional methods) can be achieved. In practical terms, this
efficiency improvement translates to better fuel efficiency. This [11] S. Saha, S. Das, R. Ghosh, B. Goswami, R. Balasubramanian, A. K.
Chandra, A. Gupta, “Design of a fractional order phase shaper for iso-
paper provides a basis for evaluating the utility of fractional damped control of a PHWR under step-back condition,” IEEE Trans.
order control to improve the performance of power conversion on Nuclear Science, Vol. 57, No.3, pp. 1602–1612, June 2010. doi:
systems and precision robotic applications. 10.1109/TNS.2010.2047405

Table VI
PARAMETERS OF THE MLS [12] S. Das, S. Das, and A. Gupta, “Fractional order modeling of a PHWR
under step-back condition and control of its global power with a robust
Symbol Parameters Values controller,” IEEE Trans. on Nuclear Science, Vol. 58, No. 5, pp. 2431–
2441, October 2011. doi: 10.1109/TNS.2011.2164422
i Input Current in the Coil [0 − 3]A
u Input Voltage [0 − 5]V
m Mass of the Steel Sphere 20 × 10−3 kg [13] S. Saha, S. Das, R. Ghosh, B. Goswami, R. Balasubramanian, A. K.
Chandra, S. Das, A. Gupta, “Fractional order phase shaper design with
k Magnetic Constant 8.54 × 10−5 kg Bode integral for iso-damped control system,” ISA Trans., Vol. 49, No.
k1 Input Conductance 0.3971/Ω 2, pp. 196–206, April 2010. doi:10.1016/[Link].2009.12.001
g Gravitational Acceleration 9.81m/s2
[14] S. Das, “Fuel efficient nuclear reactor control,” International Conference
R EFERENCES on Nuclear Engineering, Beijing, China, May 16-20, 2005.

[1] C. M. Lin, M. H. Lin, and C. W. Chen, “SoPC-based adaptive

PID control system design for magnetic levitation system,” IEEE [15] K. J. Astrom and T. Hagglund, PID Controllers: Theory, Design and
Systems Journal, Vol. 5, No. 2, pp. 278–287, June 2011. doi: Tuning, Instrument Society of America, North Carolina, 1995.
10.1109/JSYST.2011.2134530
CHOPADE et al.: DESIGN AND IMPLEMENTATION OF DIGITAL FRACTIONAL ORDER PID CONTROLLER · · · 11

[16] I. Podlubny, Fractional Differential Equations, Academic Press, New [31] I. Petras, “Fractional order feedback control of a DC motor,” Journal of
York, 1999. Electrical Engineering, Vol. 60, No. 3, pp. 117128, March 2009.

[17] S. Das, (a) Functional Fractional Calculus for System Identification and [32] S. Cuoghi and L. Ntogramatzidis, “Direct and exact methods for the
Controls, Springer Science and Business Media, 2007, doi: 10.1007/978- synthesis of discrete-time proportional-integral-derivative controllers,”
3-540-72703-3; and (b) Functional Fractional Calculus, Springer Sci- IET Control Theory and Applications, Vol. 7, No. 18, pp. 21642171,
ence and Business Media, 2011. doi: 10.1007/978-3-642-20545-3 December 2013. doi: 10.1049/iet-cta.2013.0064

[18] I. Podlubny, “Fractional-order systems and P I λ Dµ controllers,” IEEE [33] Y. Q. Chen, I. Petras, and D. Xue, “Fractional order control-a tutorial,”
Trans. on Automatic Control, Vol. 44, No. 1, pp. 208–214, January 1999. American Control Conference, St. Louis, MO, USA, June 10-12, pp.
doi: 10.1109/9.739144 1397–1411, 2009. doi: 10.1109/ACC.2009.5160719

[19] A. Charef, H. H. Sun, Y. Y. Tsao, and B. Onaral, “Fractal system as [34] Y. Jin, Y. Q. Chen, and D. Xue, “Time-constant robust analysis of
represented by singularity function,” IEEE Trans. on Automatic Control, a fractional order [proportional derivative] controller,” IET Control
Vol. 37, No. 9, pp. 14651470, September 1992. doi: 10.1109/9.159595 Theory and Applications, Vol. 5, No. 1, pp. 164172, January 2011. doi:
10.1049/iet-cta.2009.0543

[20] A. Oustaloup, F. Levron, B. Mathiew, and F. M. Nanot, “Frequency-band

complex noninteger differentiator: characterization and synthesis,” IEEE [35] C. A. Monje, B. M. Vinagre, V. Feliu, and Y. Q. Chen, “Tuning and
Trans. on Circuits and Systems I: Fundamental Theory and Applications, auto-tuning of fractional order controllers for industry applications,”
Vol. 47, No. 1, pp. 25–39, January 2000. doi: 10.1109/81.817385 Control Engineering Practice, Vol. 16, No. 7, pp. 798812, July 2008.
doi: 10.1016/[Link].2007.08.006

[21] Feedback Instruments Ltd., Magnetic levitation control experiments,

manual: 33-942S Ed01 122006, Feedback Part No. 1160–33942S. [36] J. Zhong and L. Li, “Tuning fractional-order controllers for a
solid-core magnetic bearing system,” IEEE Trans. on Control Sys-
tems Technology, Vol. 23, No. 4, pp. 1648–1656, July 2015. doi:
[22] R. Pintelon and J. Schoukens, System Identification: A Frequency 10.1109/TCST.2014.2382642
Domain Approach, Wiley-IEEE Pres, ISBN 978-0-470-64037-1, 2012.

[37] F. Padula and A. Visioli, “Optimal tuning rules for proportional-

[23] C. Yeroglu and N. Tan, “Note on fractional-order proportional-integral- integral derivative and fractional-order proportional-integral-derivative
differential controller design,” IET Control Theory and Applications, controllers for integral and unstable processes,” IET Control Theory and
Vol. 5, No. 17, pp. 19781989, November 2011. doi: 10.1049/iet- Applications, Vol. 6, No. 6, pp. 776786, April 2012. doi: 10.1049/iet-
cta.2010.0746 cta.2011.0419

[24] D. Valerio and J. Sa da Costa, “Introduction to single-input, single- [38] S. Das, S. Saha, S. Das, and A. Gupta, “On the selection of tuning
output fractional control,” IET Control Theory and Applications, Vol. 5, methodology of FOPID controllers for the control of higher order
No. 8, pp. 10331057, May 2011. doi: 10.1049/iet-cta.2010.0332 processes,” ISA Transactions, Vol. 50, No. 3, pp. 376–388, July 2011.
doi: 10.1016/[Link].2011.02.003

[25] D. Chen, Y. Q. Chen, and D. Xue, “Digital fractional order Savitzky-

Golay differentiator,” IEEE Trans. on Circuits And Systems II: Express [39] S. Das, I. Pan, S. Das, and A. Gupta, “A novel fractional or-
Briefs, Vol. 58, No. 11, pp. 758–762, November 2011. doi: 10.1109/TC- der fuzzy PID controller and its optimal time domain tuning based
SII.2011.2168022 on integral performance indices,” Engineering Applications of Arti-
ficial Intelligence, Vol. 25, No. 2, pp. 430–442, March 2012. doi:
10.1016/[Link].2011.10.004
[26] M. O. Efe, “Fractional order systems in industrial automation-a survey,”
IEEE Trans. on Industrial Informatics, Vol. 7, No. 4, pp. 582–591,
November 2011. doi: 10.1109/TII.2011.2166775 [40] S. Das, I. Pan, S. Das, and A. Gupta, “Improved model reduction
and tuning of fractional-order P I λ Dµ controllers for analytical rule
extraction with genetic programming,” ISA Transactions, Vol. 51, No.
[27] J. A. T. Machado, “Discrete-time fractional-order controllers,” Frac- 2, pp. 237–261, March 2012. doi: 10.1016/[Link].2011.10.004
tional calculus and Applied Analysis, Vol. 4, No. 1, pp. 47–66, January
2001.
[41] S. Das, I. Pan, K. Halder, S. Das, and A. Gupta, “LQR based improved
discrete PID controller design via optimum selection of weighting
[28] A. S. Dhabale, R. Dive, M. V. Aware, and S. Das, “A new method for matrices using fractional order integral performance index,” Applied
getting rational approximation for fractional order differintegrals,” Asian Mathematical Modelling, Vol. 37, No. 6, pp. 4253–4268, March 2013.
Journal of Control, Vol. 17, No. 6, pp. 21432152, November 2015. doi: doi: 10.1016/[Link].2012.09.022
10.1002/asjc.1148

[42] S. Das, I. Pan, and S. Das, “Performance comparison of optimal

[29] Y. Q. Chen and K. L. Moore, “Discretization schemes for fractional- fractional order hybrid fuzzy PID controllers for handling oscillatory
order differentiators and integrators,” IEEE Trans. on Circuits and fractional order processes with dead time,” ISA Transactions, Vol. 52,
Systems I: Fundamental Theory and Applications, Vol. 49, No. 3, pp. No. 4, pp. 550–566, July 2013. doi: 10.1016/[Link].2013.03.004
363–367, March 2002. doi: 10.1109/81.989172

[43] S. Das, I. Pan, S. Das, and A. Gupta, “Master-slave chaos synchro-

[30] I. Pan and S. Das, Gain and Order Scheduling for Fractional Order nization via optimal fractional order P I λ Dµ controller with bacterial
Controllers, Intelligent Fractional Order Systems and Control, Springer foraging algorithm,” Nonlinear Dynamics, Vol. 69, No. 4, pp. 2193–
Berlin Heidelberg, pp. 147157, 2013. doi: 10.1007/978-3-642-31549-7 2206, September 2012. doi: 10.1007/s11071-012-0419-x
12 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

[44] S. Saha, S. Das, S. Das, and A. Gupta, “A conformal mapping based Anjali S. Junghare received the B.E. and [Link]
fractional order approach for sub-optimal tuning of PID controllers with degree from Visvesvaraya Regional College of En-
guaranteed dominant pole placement,” Communications in Nonlinear gineering (VRCE), Nagpur, India, in 1981 and 1985,
Science and Numerical Simulation, Vol. 17, No. 9, pp. 3628–3642, respectively. She received the Ph.D. degree from
September 2012. doi: doi:10.1016/[Link].2012.01.007 VRCE in 2007. She is currently an associate pro-
fessor in the Department of Electrical Engineering,
Visvesvaraya National Institute of Technology, Nag-
[45] S. Das, I. Pan, K. Halder, S. Das, and A. Gupta, “Impact of fractional pur, India. She has 13 years of industrial experience
order integral performance indices in LQR based PID controller de- and 19 years of academic experience.
sign via optimum selection of weighting matrices,” IEEE International Her research areas are control system, fractional
Conference on Computer Communication and Informatics, Coimbatore, order controllers, fuzzy Logic and its applications.
India, Jan. 1012, pp. 1–6, 2012. doi: 10.1109/ICCCI.2012.6158892

[46] S. Das, I. Pan, S. Das, and A. Gupta, “Genetic algorithm based improved
sub-optimal model reduction in nyquist plane for optimal tuning rule
extraction of PID and P I λ Dµ controllers via genetic programming,”
IEEE International Conference on Process Automation, Control and
Computing, July 20, pp. 1–6, 2011. doi: 10.1109/PACC.2011.5978962

[47] A. Rajasekhar, S. Das, and A. Abraham, “Fractional order PID con-

troller design for speed control of chopper fed DC motor drive using Mohan V. Aware (M’08, SM’14) received the B.E.
artificial bee colony algorithm,” IEEE World Congress on Nature and degree in Electrical Engineering from College of
Biologically Inspired Computing, August 12, pp. 269–266, 2013. doi: Engineering, Amravati, India, in 1980, the [Link]
10.1109/NaBIC.2013.6617873 degree from the Indian Institute of Technology,
Bombay, in 1982 and the Ph.D. degree for research
work on “Direct Torque Controlled Induction Motor
[48] R. Song and Z. Chen, “Design of PID controller for maglev system based Drives” from Nagpur University, India, in 2002.
on an improved PSO with mixed inertia weight,” Journal of Networks, From 1982 to 1989, he was a Design Officer with
Vol. 9, No. 6 , pp. 1509–1517, January 2014. doi: 10.4304/jnw.9.6.1509- Crompton Greaves Ltd., Nasik, India. From 1989 to
1517 1991, he was a Development Engineer with Nippon
Denro India pvt. Ltd. During 2001-2002, he was
a Research fellow with Electrical Engineering Department, Hong Kong
[49] A. Q. Badar, B. S. Umre, and A. S. Junghare, “Reactive power Polytechnic University, Hong Kong. Presently, he is a Professor in Electri-
control using dynamic particle swarm optimization for real power cal Engineering Department, Visvesvaraya National Institute of Technology,
loss minimization,” International Journal of Electrical Power and Nagpur, India.
Energy Systems, Vol. 41, No. 1, pp. 133136, Octomber 2012. doi: His research areas are electrical drives, distributed generation with energy
10.1016/[Link].2012.03.030 storage and power electronics. He has published more than 150 technical
papers in different journals and conferences.

Amit S. Chopade (S’16) received the B.E. degree

from Nagpur University, India, in 2012. Currently,
he is working as Junior Research Fellow on BRNS
Research Project “Development of Industrial Frac-
tional Order PID Controller” at Visvesvaraya Na-
tional Institute of Technology, Nagpur, India.
His research interests include electrical drives, en-
ergy efficient systems, and applications of fractional Shantanu Das works as a Scientist Bhabha Atomic
order controllers. Research Centre(BARC), India. He graduated in
Electrical Engineering and Electronics Engineering,
from BITS Pilani and thereafter working as scien-
tist at BARC, since 1984, in the area of Nuclear
Reactor Control and Safety. He is doing develop-
ment on Fractional Calculus since about 1998, in
Swapnil W. Khubalkar (S’15) received the B.E. order to understand natural laws and to engineer
and [Link] degree from Nagpur University, In- for betterment in controls, signal processing and
dia, in 2010 and 2012, respectively. Currently he systems identification. Since 2004, he is working on
is working toward Ph.D. in the area of fractional development of meta-material science to understand
order controllers at Visvesvaraya National Institute exotic properties of electro magnetism and to manipulate electromagnetic
of Technology, Nagpur, India. flow for usage is electronics systems, and also in application of microwave
His research interests include electrical drives, power for material processing. He is Honorary Senior Research Professor
controls, and applications of fractional order con- at Department of Physics, Jadavpur University, Adjunct Professor at DIAT
trollers. Pune, and under UGC Visiting Fellow at Department of Applied Mathematics,
Calcutta University. He has several publications and patents on all these topics.