In this paper, a decentralized adaptive fuzzy control
scheme for second order systems has developed.
Simulation results for robot manipulator show that the
proposed control scheme works well, even if the ideal
model is not in concordance with the real inverse
dynamics. An important feature of this study is that it has
transferred the proposed fuzzy feedback controller to a
closed-form relation between the inputs and the output,
leading to a computationally efficient adaptive fuzzy
logic controller. The rule base consists of only four rules
and has a PD-like structure. The gains are tuned on-line
based on the gradient method. This feedback controller is
inherently bounded; the upper and lower bounds can be
arbitrary selected by suitably adjust its parameters.
Finally, it can be concluded that using the proposed
control approach presents a convenient option for
controlling a large class of nonlinear MIMO second order
systems.
7 trang |
Chia sẻ: huongthu9 | Lượt xem: 505 | Lượt tải: 0
Bạn đang xem nội dung tài liệu An Efficient Adaptive Fuzzy Control Scheme for Industrial Manipulators, để tải tài liệu về máy bạn click vào nút DOWNLOAD ở trên
An Efficient Adaptive Fuzzy Control Scheme for
Industrial Manipulators
Abdel Badie Sharkawy, Douglas A. Plaza, and Daniel E. Ochoa
Escuela Superior Politécnica del Litoral, ESPOL, Guayaquil, Ecuador.
Email: {asharkaw, dochoa, douplaza}@fiec.espol.edu.ec
Abstract—This paper develops a generalized adaptive fuzzy
control scheme for MIMO nonlinear second order systems.
Here, the example robotic manipulators is used to illustrate
the control algorithm. The controller for each degree of
freedom (DOF) consists of a feedback fuzzy PD systems
used to keep the closed-loop stable. The rule base consists of
only four rules per each DOF. Furthermore, the fuzzy
feedback system is decentralized and simplified leading to a
computationally efficient control scheme. The proposed
control scheme has the following advantages: 1) it needs no
exact dynamics of the system and the computation is time-
saving because of the simple structure of the fuzzy systems;
and 2) the controller is robust against various uncertainties.
The computational complexity of the proposed control
scheme has been analyzed and compared with previous
works. Computer simulations show that this controller is
effective in achieving the control goals.
Index Terms—robot manipulators, fuzzy pd feedback
control, closed-loop stability, computational complexity
I. INTRODUCTION
Generally speaking, multiple-input multiple-output
(MIMO) systems usually have characteristics of
nonlinear dynamics coupling. Therefore, the difficulty in
controlling MIMO systems is how to overcome the
coupling effects between the degrees of freedom. The
computational burden and dynamic uncertainty associated
with MIMO systems make model-based decoupling
impractical for real-time control.
Adaptive control has been studied for many decades to
deal with constant or slowly changing unknown
parameters. Applications include manipulators, ship
steering, aircraft control and process control. Although
the perfect knowledge of the inertia parameters can be
relaxed via adaptive technique, its real practical
usefulness is not really clear and the obtained controllers
may be too complicated to be easily implemented, [1].
Also, because of many design parameters, like learning
rates and initialization of the parameters to be adapted,
etc., most existing methodologies have limitations.
Moreover, owing to the different characteristics among
design parameters, attaining a complete learning, while
considering an overall performance goal, is an extremely
difficult task, [2]
Manuscript received January 1, 2015; revised June 21, 2015.
Fuzzy controllers have demonstrated excellent
robustness in both simulations and real-life applications.
However, it has been proved that standard fuzzy logic
controllers are not suitable for loop controllers, [3]. This
fact is referred to that there are many tuning parameters
in membership functions and control rules. Furthermore,
standard fuzzy logic controller has a long computation
time since it performs fuzzification, inference, and
defuzzification processes in determining control inputs.
Thus, it is difficult for control inputs of standard fuzzy
logic control to be computed within the sampling time of
a loop controller. For this reason, complexity reduction of
fuzzy feedback controllers was the topic of many
researchers; for instance see [4].
In this paper, we focus on the design of an adaptive
fuzzy feedback controller based on the Lyapunov
synthesis. Only four rules constitute the rule base for each
DOF. Furthermore, the fuzzy feedback controller is
decentralized and simplified leading to a computationally
efficient adaptive fuzzy control scheme. To demonstrate
the proposed approach, we use the example of robotics
because it is a well-known example of nonlinear MIMO
second order systems.
The paper is outlined as follows: in Section 2, the robot
model and the nominal value of its parameter are
introduced. This model is used to generate simulation
data instead of experimental data from real robot platform.
In Section 3, the fuzzy feedback controller is derived
based on the Lyapunov direct method. Furthermore, the
controller is simplified, i.e. it has a closed form
mathematical relation with only three parameters need to
be tuned and the controller gain is adaptively determined
on-line so as to minimize a performance index. Section 4
discusses the computational complexity of the proposed
control scheme in comparison with previous works.
Simulation results are demonstrated in Section 5. Finally,
some concluding remarks are given in Section 6.
II. ROBOT MODELING AND THE CONTROL STATEMENT
Without the loss of generality, we take the two-link
rigid robot shown in Fig. 1, as an example to demonstrate
the proposed control scheme. The inverse dynamic model
is expressed as:
)(),()( GCMu (1)
Journal of Automation and Control Engineering Vol. 4, No. 3, June 2016
©2016 Journal of Automation and Control Engineering 213
doi: 10.18178/joace.4.3.213-219
where
nR is the joint angular position vector of the
robot;
nRu is the vector of applied joint torques (or
forces);
nnRM )( is the inertia matrix, positive
definite;
nRC ),( is the effect of Coriolis and
centrifugal torques; and
nRG )( is the gravitational
torques. The physical properties of the above model (1)
can be found in [5]; however, they are not needed here.
1
2
1l
cel
1m
em
X
Y
1cl
e
g
unknown load
Figure 1. An articulated two-link manipulator.
For the robot shown in Fig. 1, (1) can be rewritten as:
2
1
2
1
1
212
2
1
2221
1211
2
1
0
)(
G
G
h
hh
MM
MM
u
u
where
)sin(2)cos(2 2423111 aaaM , 222 aM ,
)sin()cos( 242321221 aaaMM ,
)cos()sin( 2423 aah
)cos()cos( 212111 bbG ,
)cos( 2122 bG
With
2
1
22
1111 lmlmIlmIa eceeec ,
2
2 ceee lmIa ,
)cos(13 ecee llma , )sin(14 ecee llma
1111 glmglmb ec , .2 ceeglmb
The nominal parameters of the two-link manipulator
are chosen as follows:
kgm 51 , kgme 5.2 , ml 0.11 , mlc 5.01
mlce 5.0 ,
030e ,
2
1 36.0 kgmI ,
2 24.0 kgmIe
Position control, or also the so-called regulation
problem is one of the most relevant issues in the
operation of robot manipulators. This is a particular case
of the motion control or trajectory control. The primary
goal of motion control in joint space is to make the robot
joints track a given time-varying desired joint position,
Tddd ],[ 21 .
Several control architectures related to robot control
can be found in literature ranging from the simple PD,
learning based, adaptive, and adaptive/learning hybrid
controllers. The reader is referred to [6], [7] and the
references included. The main advantage of the PD
controller is that it can easily be implemented on simple
microcontroller architectures. On the other hand, the
performance obtained from PD controllers is not
satisfying for most of the sensitive applications [7].
III. DECENTERALIZED ADAPTIVE FUZZY CONTROL
The performance of any fuzzy logic controller is
greatly dependent on its inference rules. In most cases,
the closed-loop control performance and stability are
enhanced if more rules are added to the rule base of the
fuzzy controller. However, a large set of rules requires
more on-line computational time and more parameters
need to be adjusted. Adjustment of the fuzzy system may
be achieved using GAs [8]. However, GAs cannot be
used on-line and perfect mathematical model or
experimental data should be available. In this Section, a
robust adaptive PD-type fuzzy feedback controller is
driven for a class of MIMO second order nonlinear
systems.
A. Construction of Fuzzy Feedback Controllers
In this Sub-section we apply the fuzzy synthesis to the
design of stable controllers. To this end, consider a class
of MIMO nonlinear second order systems whose dynamic
equation can be expressed as:
),,()( uxxftx (2)
where ),,( uxxf is an unknown continuous function, u
is the feedback control input and Tnxxxtx ],,,[)( 21 is
the state vector and T
21 ],...,,[ nxxx
dt
dx
x . We now seek a
smooth Lyapunov function nn RRV : for the
continuous feedback model (1) that is positive definite,
i.e. 0)( xV when 0x and 0)( xV when 0x , and
grows to infinity: )(xV as xxT . Obviously,
this holds for a generalized Lyapunov candidate function
of the following quadratic form:
xxxxtxV TT
2
1
2
1
),( (3)
Differentiating (3) with respect to time gives
nnnn xxxxxxxxxxxxtxV 22112211),(
from which
)()()(),( 22221111 nnnn xxxxxxxxxxxxtxV
This is equal to
nVVVtxV
21),( (4)
Journal of Automation and Control Engineering Vol. 4, No. 3, June 2016
©2016 Journal of Automation and Control Engineering 214
where
nixxxxtxV iiiii ,2,1 ,),(
Then the standard results in Lyapunov stability theory
imply that the dynamic system (2) has a stable
equilibrium exx if each iV
in (4) is 0 along the
system trajectories. To achieve this, we have chosen the
control )(xui to be proportional to ix .
Next, our controller design is achieved if we determine
a fuzzy control )(xu
iFB
so that:
nixuxxxtxV iiiiii ,2,1 ,0)(),( (5)
where i is a positive constant. The results of Wang [9]
state that, a fuzzy system that would approximate (5)
exists. To this end, one would consider the state vector
)(tx and )(tx to be the inputs to the fuzzy system. The
output of the fuzzy system is the feedback control iu . A
possible form of the control rules is:
IF ix is (lv) and/or ix is (lv)
THEN iu is (lv), ni ,2,1
where the (lv) are linguistic values (e.g. positive,
negative). These rules constitute the rule base for a
Mamdani-type fuzzy controller.
In the above formulation, two basic assumptions have
been made. They are:
The knowledge of the state vector. It is assumed to
be available from measurements.
The control input, u is proportional to x . This
assumption can be justified for a large class of
second order nonlinear mechanical systems, [10].
For instance, here in robotics, it means that the
acceleration of links is proportional to the input
torque.
These two assumptions represent the basic knowledge
about the system which is needed to derive the control
rules. Clearly, the exact mathematical model is not
needed.
B. Adaptive Fuzzy Feedback Control Design
Robots are familiar examples of trajectory-following
mechanical systems. Their nonlinearities and strong
coupling of the robot dynamics present a challenging
control problem. In practice, the load may vary while
performing different tasks, the friction coefficients may
change in different configurations and some neglected
nonlinearities as backlash may appear. Therefore, the
control objective is to design a stable fuzzy controller so
that the link movement follows the desired trajectory in
spite of such effects.
Consider a class of robots whose vector of generalized
coordinates is denoted by Tn 21 where i ,
ni ,,1 are the joint parameters. We consider the state
variables of the robot as )(t
and )(t , which are
usually available as feedback signals. Define the tracking
error vectors
)(te
and
)(te
as:
)()()( ttte d , and )()()( ttte d (6)
where
d and
d are vectors of the desired joint
position and velocity, respectively. We now apply the
approach presented in the previous Sub-section in order
to find a fuzzy controller that achieves tracking to the
robotic system under consideration. To this end, let us
choose the following Lyapunov function candidate
)(
2
1
eeeeV TT (7)
Differentiating with respect to time and using (4) gives
iiiii eeeeV
To enforce asymptotic stability, it is required to find u
so that
0 iiiii eeeeV (8)
In some neighborhood of the equilibrium of (7).
Taking the control u to be proportional to e , Eqn (8)
can be rewritten as:
0
iFBiiiii
ueeeV (9)
where i is positive constant, ni ,,1 . Sufficient
conditions for (9) to hold can be stated as follows.
if, for each ],,1[ ni , ie and ie have opposite
signs and iiu is zero, inequality (9) holds;
if ie and ie are both positive, then (9) will hold if
iiu is negative; and
if ie and ie are both negative, then (9) will hold if
iiu is positive. ],,1[ ni denotes the joint
number.
Using these observations and assuming that i is
positive small number, one can easily obtain the four
rules listed below in Table I.
TABLE I. FUZZY RULES FOR THE FUZZY FEEDBACK CONTROLLER
P N
P
N
ie
ie
Nu
PuZu
Zu ni ,1
In this Table, P, N, denote respectively positive,
negative errors; Pu , Nu and Zu are respectively positive,
negative and zero control inputs. These rules are simply
the fuzzy partitions of ie , ie
and iu
which follow
directly from the stabilizing conditions
of the Lyapunov
function, (7).
Journal of Automation and Control Engineering Vol. 4, No. 3, June 2016
©2016 Journal of Automation and Control Engineering 215
In concluding words, the presented approach
transforms classical Lyapunov synthesis from the world
of exact mathematical quantities to the world of words
[10]. This combination provides us with a solid analytical
basis from which the rules are obtained and justified.
Relative to other works, this number of rules is quite
small. For example, in [11], the rule base of a two-link
robot consists of 625 rules. After introducing a rule base
reduction approach, the authors in [11] reach to a rule
base consists of 160 rules, which is hard to be
implemented.
To complete the design, we must specify the fuzzy
system with which the fuzzy feedback computes the
control signal. The Gaussian membership defining the
linguistic terms in the rule base is chosen as follows:
2)(
),()( z
ax
zpositive eaxGx
),()( znegative axGx
)0,()( xGxzero
where
0za and z stands for control variable, the
product for “and” and center of gravity inferencing. For
some positive constant i
k
,
],,1[ ni
denotes the joint
number, the above four rules can be represented by the
following mathematical expression:
),(),(
))(,())(,(
),(),(
))(,())(,(
22
22
11
11
iiii
iiiiii
iiii
iiiiii
i
aeGaeG
kaeGkaeG
aeGaeG
kaeGkaeG
u
This yields the fuzzy feedback controller
)]2(tanh)2(tanh[ 21 iiiiii eaeaku , ni ,,1 (10)
In (10), the inputs are the error in position ie and the
error in velocity ie and the output is the control input of
joint i; i.e. it is a PD-type fuzzy feedback controller. The
following remarks are in order:
The fuzzy controller in (10) is a special case of
fuzzy systems, where Gaussian membership
functions are used to introduce the input variables
( ie and ie ) to the fuzzy network. Also, the
fuzzification and defuzzification methods used in
this study are not unique; see [10] for other
alternatives. For example, using different
membership functions (e.g. triangular, trapezoidal
etc.) will result in a different fuzzy controller.
However, the controller in (10) is a simple one and
the closed form relation between the inputs and
the output makes it computationally inexpensive.
Only three parameters per each DOF need to be
tuned, namely, they are ik , ia1 and ia2 . This
greatly simplifies the tuning procedure; since the
search space is quite small relative to other works.
For instance, the fuzzy controller in [12] needs 45
parameters to be tuned for a one DOF system.
This controller is inherently bounded since
1)(tanh x .
Finally, the fuzzy PD gain, i.e. ik , ],,1[ ni is
chosen so as to minimize the following quadratic
performance index:
2)]([
2
1
kurJ iii (11)
where input ir is a constant. According to the gradient
method, the learning algorithm of the parameter ik in the
feedback fuzzy controller (10) can be derived as follows:
)]2tanh()2[tanh( 21 iiiiii
ii
ii
i
i
i
ececur
ku
uJ
k
J
k
(12)
Thus, the fuzzy feedback controller uses the ie , ie and
iu to compute (12) and update the control gain ik given
that .0)0( ik The overall closed-loop control system is
shown in Fig. 2.
Fuzzy Feedback
controller (10)
d
i
d
i
,
ii
,
Robot link i
_
ii ee ,
The update law
(12)
iu
Figure 2. Configuration of the proposed decentralized fuzzy control
scheme of joint i.
IV. COMPUTATIONAL ASPECTS
In general, control algorithms for closed loop control
should require short computation time due to limited
memory of low-cost microprocessors. This Section
discusses the computational complexity of the feedback
controller and compares it with that of a self-tuning fuzzy
controller proposed in [13]. It is shown that the proposed
control scheme is computationally very efficient.
Naturally, the computational burden can be evaluated
in terms of required mathematical multiplication and
addition operations. The computation of the fuzzy
feedback controller can be divided into two parts:
computation of (10) and computation of the adaptive gain;
ik , (12). For the sake of comparison, Table II
demonstrates the computational complexity of our
scheme with the self-tuning fuzzy controller proposed in
[13]. The comparison is fair since the feedback controller
in [13] is essentially a PD fuzzy controller with self-
tuning mechanism. In [13], the rule base has been
transformed to a decision table and is used by a back-
propagation algorithm to adjust the scaling factors of the
fuzzy system. The difference resides in the fact that the
rule base in [13] consists of 49 rules for one DOF system
and the mapped elements ( e and e ) are obtained by
interpolation. Furthermore, the tuning procedure is
Journal of Automation and Control Engineering Vol. 4, No. 3, June 2016
©2016 Journal of Automation and Control Engineering 216
composed of two stages and some learning steps are
needed by the second stage, while the tuning system
using (12) is much simpler. Simulation results, in the
coming Section, show that it is also efficient.
TABLE II. COMPUTATIONAL COMPLEXITY OF THE FUZZY FEEDBACK
CONTROLLER
Self-tuning fuzzy
controller [13]
The proposed
fuzzy controller
Addition 97n 6n
Multiplication 113n 15n
V. SIMULATION RESULTS
The purpose of the simulation is to investigate the
robustness of the proposed control scheme. The robot
system considered in the simulation is the two-link robot
presented in Section 2. Through the simulations, the
physical insight of the behavior is revealed. In the coming
results, it is assumed that )1(5.01
td e ,
)1(2
td e and initial positions of joints
rad 15/)0()0( 21 , which are equivalent to the
initial position errors, since the desired positions are
0)0()0( 21
dd . Also, the robot is initialized at rest, i.e.
the initial velocities of joints sec/
0)0()0( 21 rad
o .
This initialization imposes a large initial velocity error
since ,2/)0(1 e
sec/ )0(2 rade . The input
torque has been saturated to mNuu .
300, 21 . With
these initialization conditions, one can expect uneasy
transient stage.
0 1 2 3
200
400
600
800
Time
in
seconds
N
.m 1k
2k
Figure 5.
Record of the
adaptive control gains during motion.
The input torques are shown in Fig. 3 and Fig. 4 shows
the evolution of the tracking errors. They show that the
errors have converged to zero. Note that the transient
period is less than 0.5 seconds. Otherwise, it is interesting
to notice how the control gains evolve with time. Fig.
5
depicts the evolution of these parameters with time. They
have been initialized as mNk .
600)0(1
and
mNk .
275)0(2 .
In order to observe how the controller behaves in the
presence of various uncertainties, two types of
uncertainties are considered, namely, unmodeled
nonlinear friction and unknown payloads.
A.
Unmodeled Friction
At the off-line training stage of our simulation, we
obtain the training samples from the robot model in (1),
which does not consider the nonlinear friction. In order to
examine the performance of the controller in the presence
of unmodeled nonlinear friction, the following
unmodeled nonlinear friction is added at the control stage:
sd FFF
where dF
and sF
are the dynamic and static friction
torques, respectively. They can be expressed by:
2
1
22
11
)cos(
0
0 )cos(
x
x
xd
xd
Fd
and
)sgn(
)sgn(
22
11
xc
xc
Fs
We use 30 ,50 21 dd
and 12 ,18 21 cc . Results
are shown in Fig. 6 and
Fig.
7. It can be noticed that the
transient period has increased relative to the cases when
the friction was not considered. Also the input torques is
relatively higher during this period. Nevertheless,
convergence of the tracking errors has been achieved.
0 1 2 3
-0.4
-0.2
0
0.2
0.4
Time
in
seconds
E
rr
o
r
in
ra
d
ia
n
1e
2e
Figure 6.
The tracking errors
in the presence of unmodeled friction.
0 1 2 3
-400
-200
0
200
400
Time in seconds
N
.m
1u
2u
Figure 3. The control effort.
0 1 2 3
-0.4
-0.2
0
0.2
0.4
Time in seconds
E
rr
o
r
in
ra
d
ia
n
1e
2e
Figure 4. The tracking errors
Journal of Automation and Control Engineering Vol. 4, No. 3, June 2016
©2016 Journal of Automation and Control Engineering 217
0 1 2 3
-200
-100
0
100
200
300
Time in seconds
N
.m
1u
2u
Figure 7. The control input in the presence of unmodeled friction.
B. Unknown Payload
In robot systems, the unknown payload is one of the
major dynamic uncertainties. Compared with the
parameter uncertainties and unmodeled friction, the
influence of unknown payload is much greater. The
coming results are obtained when the mass and inertia of
the base and elbow links (carrying the payload) have been
increased to 150%. This increase in the mass and inertia
of the two links is supposed to be unknown. Fig. 8 shows
that input torque is relatively high. Also, the tracking
errors exhibit larger overshoot during the transient period,
Fig. 9. However, convergence of errors to a narrow
region close to zero has taken place.
0 1 2 3
-400
-200
0
200
400
Time in seconds
N
.m
1u
2u
Figure 8.
The input torques when the payload
increases to 150%.
0 1 2 3
-0.6
-0.4
-0.2
0
0.2
0.4
Time
in
seconds
E
rr
o
rs
in
ra
d
ia
n
1e
2e
Figure 9.
The
tracking errors in the presence of 150% increase in the
payload.
VI. CONCLUSIONS
In this paper, a decentralized adaptive fuzzy control
scheme for second order systems has developed.
Simulation results for robot manipulator show that the
proposed control scheme works well, even if the ideal
model is not in concordance with the real inverse
dynamics. An important feature of this study is that it has
transferred the proposed fuzzy feedback controller to a
closed-form relation between the inputs and the output,
leading to a computationally efficient adaptive fuzzy
logic controller. The rule base consists of only four rules
and has a PD-like structure. The gains are tuned on-line
based on the gradient method. This feedback controller is
inherently bounded; the upper and lower bounds can be
arbitrary selected by suitably adjust its parameters.
Finally, it can be concluded that using the proposed
control approach presents a convenient option for
controlling a large class of nonlinear MIMO second order
systems.
ACKNOWLEDGMENT
This work was supported by a grant from the
Ecuadorian Government, SENESCYT.
REFERENCES
[1] S. Tong, S. Sui, and Y. Li, “Adaptive fuzzy decentralized tracking
fault-tolerant control for stochastic nonlinear large-scale systems
with unmodeled dynamics,” Information Sciences, vol. 289, pp.
225–240, 2014.
[2] S. Tong, Y. Li, Y. Li, and Y. Liu, “Observer-based adaptive fuzzy
backstepping control for a class of stochastic nonlinear strict-
feedback systems,” IEEE Trans. on Systems, Man, and
Cybernetics—Part b: Cybernetics, vol. 41, no. 6, pp. 1693-1704,
December, 2011.
[3] M. Margaloit and G. Langholz, “Fuzzy control of a benchmark
problem: Computing with words approach,” IEEE Trans. on Fuzzy
Systems, vol. 12, no. 2, pp. 230-235, 2004,.
[4] Y. H. Kim, S. C. Ahn, and W. H. Kwon, “Computational
complexity of general fuzzy logic control and its simplification for
a loop controller,” Fuzzy Sets and Systems, vol. 111, pp. 215-224,
2000.
[5] T. Hseng, S. Li, and Y. C. Huang, “MIMO adaptive fuzzy
terminal sliding-mode controller for robotic manipulators,”
Information Sciences, vol. 180, pp. 4641-4660, 2010.
[6] S. Liuzzo and P. Tomei “A global adaptive learning control for
robotic manipulators,” Automatica, vol. 44, pp. 1379-1384, 2008.
[7] S. Yamacli and H. Canbolat, “Simulation of a SCARA robot with
PD and learning controllers,” Simulation Modelling Practice and
Theory, vol. 16, pp. 1477-1487, 2008.
[8] A. B. Sharkawy, “Genetic fuzzy self-tuning PID controllers for
antilock braking systems,” Engineering Applications of Artificial
Intelligence, vol. 23, no. 7, pp. 1041-1052, 2010.
[9] L. X. Wang, A Course in Fuzzy Systems and Control, Upper
Saddle River, NJ: Prentice-Hall, ch. 4, 1997, pp. 76-81.
[10] L. A. Zadeh, “Fuzzy logic = computing with words,” IEEE Trans.
Fuzzy Systems, vol. 4, no. 2, pp. 103-111, 1996.
[11] H. Bezine, N. Derbel, and A. M. Alimi, “Fuzzy control of robot
manipulators: Some issues on design and rule base size
reduction,” Engineering Applications of Artificial Intelligence, vol.
15, pp. 401-416, 2002.
[12] T. L. Seng, M. Khalid, and R. Yusof, “Tuning of a neuro-fuzzy
controller by genetic algorithm,” IEEE Transaction on Systems,
Man, and Cybernetics, Part B: Cybernetics, vol. 29, no. 2, pp.
226-239, April, 1999.
[13] C. T. Chao and C. C. Teng, “A PD-like self-tuning fuzzy controller
without steady-state error,” Fuzzy Sets and Systems, vol. 87, pp.
141-154, 1997.
Journal of Automation and Control Engineering Vol. 4, No. 3, June 2016
©2016 Journal of Automation and Control Engineering 218
Abdel Badie Sharkawy
was born in Assiut
Egypt on December 4th, 1958 and got PhD in
Robotic Control in 1999 from
the Slovak
Technical University in Bratislava, Slovakia.
The major area of interest
are modeling and
control using intelligent systems. He was a
senior lecturer within the department of
mechatronics engineering, the Hashemite
University, Jordan for three years (2001-2004)
and within the electrical engineering
department, Al-Tahady University, Sirte, Libya
during the fall semester, 2005. He was a professor at mechanical
engineering department, Assiut University, Egypt (April 2012-
July
2014). Now, he is a visiting professor (Prometeo Program) in the
Faculty of Electrical and Computer Engineering, ESPOL, Guayaquil,
Ecuador. His research interests include adaptive fuzzy identification and
control, automotive control systems, robotics (modeling and control),
and the use of neural networks in the control of mechanical systems.
Douglas A. Plaza born in Guayaquil, Ecuador
on 10th June 1977. He obtained his Ph.D.
degree in Electromechanical Engineering from
Ghent University, Ghent, Belgium
in 2013
and
a master degree in Industrial Control
Engineering from Universidad de Ibague,
Ibague, Colombia
in 2008.
Currently, he is a
lecturer of control theory associated to the
Faculty of Electrical and Computing
Engineering at Escuela Superior Politecnica del
Litoral ESPOL. His research interests include: Kalman filtering,
sequential Monte Carlo methods, model predictive control and non-
linear control.
Daniel E. Ochoa
was born in Guayaquil-
Ecuador on October 17th, 1975. He was
awarded an Computer Engineering degree at
Escuela Superior Politécnica del Litoral
(ESPOL), Guayaquil-Ecuador in 2000 and a
PhD in Computer Science in 2011 at Gent
University in Gent-Belgium.
His interest are
computer vision and robotics.
He worked as
research assistant at IPI group in Gent
University and currently he is the head of the
Computer Vision and Robotic center at ESPOL where he also works as
a professor since 2013.
He has published several indexed publications
and has been reviewer of national and international scientific journals.
Most of his research work has been done in biological image analysis.
He is a founding
member and head of the robotic and intelligent systems
network of Ecuador.
Journal of Automation and Control Engineering Vol. 4, No. 3, June 2016
©2016 Journal of Automation and Control Engineering 219
Các file đính kèm theo tài liệu này:
- an_efficient_adaptive_fuzzy_control_scheme_for_industrial_ma.pdf