Hindawi

Mathematical Problems in Engineering

Volume 2018, Article ID 5794067, 15 pages
https://doi.org/10.1155/2018/5794067

Research Article
Position Tracking Control of PMSM Based on Fuzzy
PID-Variable Structure Adaptive Control

Pei Pei , Zhongcai Pei, Zhiyong Tang , and Han Gu

School of Automation Science and Electrical Engineering, Beihang University, Beijing 100191, China

Correspondence should be addressed to Zhiyong Tang; zyt [email protected]

Received 29 May 2018; Revised 14 August 2018; Accepted 7 September 2018; Published 30 September 2018

Guest Editor: Carlos Llopis-Albert

Copyright © 2018 Pei Pei et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A novel Fuzzy PID-Variable Structure Adaptive Control is proposed for position tracking of Permanent Magnet Synchronous Motor
which will be used in electric extremity exoskeleton robot. This novel control method introduces sliding mode variable structure
control on the basis of traditional PID control. The variable structure term is designed according to the sliding mode surface which
is designed by system state equation, so it could compensate for the disturbance and uncertainty. Considering the chattering of
sliding mode system, the fuzzy inference method is adopted to adjust the parameters of PID adaptively in real time online, which
can attenuate chattering and improve control precision and dynamic performance of system correspondingly. In addition, compared
with the traditional sliding mode control, this method takes the fuzzy PID control item to replace the equivalent control item of
sliding mode variable structure control, which could avoid the control performance reduction resulted from modeling error and
parameter error of system. It is proved that this algorithm can converge to the sliding surface and guarantee the stability of system by
Lyapunov function. Simulation results show that Fuzzy PID-Variable Structure Adaptive Control enjoys better control precision and
dynamic performance compared with traditional control method, and it improves the robustness of system significantly. Finally,
the effectiveness and practicability of the algorithm are verified by the method of Rapid Control Prototyping on the semiphysical
simulation test bench.

1. Introduction precision, and easy engineering implementation is the focus

of research.
Lower extremity exoskeleton system is a kind intelligent Sliding Mode Cotrol (SMC) is widely used as a nonlinear
assist robot, which is wearable, combines the operator with control method currently. By designing the sliding mode
machine, and could be used in medical and military fields surface reasonably, the control performance of system will
wildly [1–3]. Accurate and fast trajectory tracking is one of the not be affected by internal parameter perturbation and exter-
core issues of extremity exoskeleton robot [4, 5]. Therefore, nal disturbance, and the system will have strong robustness
the high precision control of the extremity exoskeleton robot and high control precision. However, the equivalent control
driven by Permanent Magnet Synchronous Motor (PMSM) term of sliding mode control depends on the accuracy of
is one of the core problems of robot control. Conventional system model and parameters, while in practical application,
PID control which has weak robustness is incapable of only an approximate mathematical model of motor can be
dealing with nonlinear control, although it has the advantage obtained. And there are still some uncertain factors, such as
of simple structure and easy implementation. PMSM is a parameter error and external interference, which will reduce
multivariable and strongly coupled nonlinear system, thus the performance of the sliding mode control and even cause
having many factors in its control system, such as current the instability of system [7]. To solve the problem that SMC
coupling, system saturation, parameter perturbation, and cannot obtain the accurate model and parameters of system,
external disturbance, which affect the performance of system some intelligent methods are introduced to compensate the
directly [6]. Therefore, designing a control method with equivalent control item. Reference [8] introduced neural
strong robustness, good dynamic performance, high control network into the variable structure control to compensate
2 Mathematical Problems in Engineering

for the lack of accurate system models and parameters and parameters of PID. Since fuzzy PID control is a mature
proposed a neural network controller online training method control method applied in engineering, it does not make the
which was applied to the speed loop of electric drive and control structure too complex.
improved the adaptability and robustness greatly based on In this paper, the Fuzzy PID-Variable Structure Adaptive
variable structure control. An adaptive fuzzy sliding mode Control, which is mutual compensation of sliding mode
control strategy was designed to compensate for the influence variable structure control and fuzzy PID control, is pro-
of nonlinear, disturbances, and uncertain parameters in posed for trajectory tracking of PMSM used in extremity
asymmetric hydraulic cylinder position servo control system exoskeleton system. Firstly, this algorithm ensures stability of
in [9]. In this paper, the uncertain factors of the models and system by PID control instead of equivalent control. Then,
parameters were treated as interference. In [10], by using a sliding mode surface based on state equation of system is
the linear feedback RBF, the sliding mode controller was designed for sliding mode variable structure control which
compensated, and the control effect was further improved. compensates for insensitivity of PID control to parameter
Reference [11] proposed a method that achieved good results perturbations and external disturbances. Finally, the fuzzy
combining fast terminal sliding mode (FTSM) with radial reasoning method is added to adjust the parameters of PID
basis function (RBF), which was applied to the vector control adaptively online in real time, which improves robustness
of PMSM. The method of combining neural network control of system again and attenuates chattering of sliding mode
and sliding mode variable structure control was proposed system as well.
in [12, 13] with the main idea of meeting the sliding mode
condition by adjusting the gain coefficient of control law
online. Simply introducing the intelligent control method will 2. Mathematical Model of PMSM
make the controller structure complicated, increase the cost
of controlling, and reduce the usability of control system. The mathematical model of PMSM in the rotating shaft (𝑑, 𝑞)
An ideal controller is designed to strive to reach an optimal is
tradeoff between performance and complexity.
As is known, PID control is a simple, easy to implement, 𝑅𝑠 𝑢
𝑖𝑑̇ = − 𝑖 + 𝑝𝑛 𝜔𝑖𝑞 + 𝑑
and widely used control method. In the face of the traditional 𝐿 𝑑 𝐿
SMC the accurate models and parameters of system cannot
𝑅 𝑝𝑛 𝜓𝑓 𝑢𝑞
be obtained. It is natural to think of the combination of 𝑖𝑞̇ = −𝑝𝑛 𝜔𝑖𝑞 − 𝑠 𝑖𝑑 − 𝜔+
sliding mode variable structure control and PID control. In 𝐿 𝐿 𝐿 (1)
the design of sliding surface, the integral term was introduced 𝑝𝑛 𝜓𝑓 𝐵 𝑇
to form a sliding surface which was similar to PID structure 𝜔̇ = 𝑖 − 𝜔− 𝐿
𝐽 𝑞 𝐽 𝐽
in [14]. Two novel methods, integral approach law and PID-
like approach law, were proposed in [15], to control Electro- 𝜃̇ = 𝜔,
Hydraulic Actuator (EHA). Among these control methods,
PID control was only introduced to design sliding surface;
however, the controller also consists of equivalent control where 𝑢𝑑 and 𝑢𝑞 are voltage component of d axis and
item based on mode of system. In [16], PID control and q axis of the stator, respectively (𝑉), 𝑖𝑑 and 𝑖𝑞 are current
sliding mode variable structure control were combined to component of d axis and q axis of the stator, respectively (𝐴),
be variable structure PID control (VSPID). PID control was 𝑅𝑠 is the resistance of stator winding (Ω), 𝜓𝑓 = √3/2𝜓𝑓𝑚
introduced to replace the variable structure coefficient in is electronic flux linkage, 𝐿 = 𝐿 𝑑 = 𝐿 𝑞 is the inductance
SMC, so that the VSPID became a strict nonlinear controller. of motor, 𝐵 is friction coefficient proportional to rotational
However, there was a risk that only nonlinear control cannot speed, 𝑇𝐿 is load torque (𝑁.𝑚), 𝐽 is moment of inertia
guarantee the optimal performance of the system. A similar (𝑘𝑔.𝑚2 ), 𝜃 = 𝜃𝑒 /𝑝𝑛 is mechanical angle position of motor
approach was proposed in [17], and due to the use of PID rotor (𝑟𝑎𝑑), and 𝜔 = 𝜔𝑒 /𝑝𝑛 is mechanical angle velocity of
control instead of the variable structure factor, it is still a motor rotor (𝑟𝑎𝑑/𝑠).
switch value control. PMSM servo system is a three-closed-loop control system
Since the variable structure control is introduced in the based on flux orientation, including position control loop,
design of controller in this paper, the chattering caused by speed control loop, and current control loop. The three-
variable structure control must be considered. References closed-loop control of PMSM is shown in Figure 1 while
[18, 19] put forward an adaptive fuzzy sliding mode controller, 𝑖𝑑 = 0. Besides PMSM, the system also includes a SVPWM
which realized the self-adaption of variable structure param- inverter, a power rectifier, a position sensor, a current
eters and attenuated the chattering caused by sliding mode controller (ACR), a speed controller (ASR), and a position
control. Reference [20] proposed a method of fuzzy reaching controller (APR). The position controller (APR) is our focus;
law for robot control system, which greatly attenuated the meanwhile the speed controller (ASR) and the position
chattering of traditional sliding mode system and made the controller (APR) are designed based on the traditional PID
system achieve better control performance. In order to give algorithm.
full play to the performance of the variable structure control In order to design the position loop controller con-
item, the fuzzy inference method is introduced to attenuate veniently, it is assumed that the speed control loop, the
the chattering of the system by adaptively modifying the current control loop, and the inverter are ideal. Therefore,
Mathematical Problems in Engineering 3

AC

rectifier

 +  +  + 
d, q

APR ASR SV
ACR   Inverter
- -  =  + - PWM
, 
  -


  
d, q , 
d/dt 
 
,  a, b, c

Sensor PMSM

Figure 1: Control schematic of PMSM.

the mathematical model of the PMSM can be simplified as 󳵻𝐾𝑑 ) is parameters compensation of PID controller. The PID
a second-order differential link [21] as parameters in practical applications can be expressed as

𝜃̇ = 𝜔 𝐾𝑃 = 𝐾𝑝 󸀠 + 󳵻𝐾𝑝
𝑝𝑛 𝜓𝑓 𝐵 𝑇 (2)
𝜔̇ = 𝑖𝑞 − 𝜔 − 𝐿. 𝐾𝑖 = 𝐾𝑖 󸀠 + 󳵻𝐾𝑖 (6)
𝐽 𝐽 𝐽
𝐾𝑑 = 𝐾𝑑 󸀠 + 󳵻𝐾𝑑 ,
It can be also expressed as
where 𝐾𝑝 󸀠 , 𝐾𝑖 󸀠 , and 𝐾𝑑 󸀠 are the preset parameter values
𝑝𝑛 𝜓𝑓 𝐵 𝑇
𝜃̈ = 𝑖 − 𝜃̇ − 𝐿 . (3) of PID controller.
𝐽 𝑞 𝐽 𝐽 The control principle of fuzzy PID control is shown in
Figure 2.
Let 𝑝𝑛 𝜓𝑓 /𝐽 = 𝑓, −𝐵/𝐽 = 𝑔, −𝑇𝐿 /𝐽 = 𝑑, and 𝑖𝑞 = 𝑢. The core of fuzzy PID is the design of fuzzy inference
Then (3) can be simplified as rules. Firstly, fuzzify the input variables 𝑒 and 𝑒𝑐 by member-
ship function, and then the inference is carried out according
𝜃̈ = 𝑔𝜃̇ + 𝑓𝑢 + 𝑑, (4)
to predefined fuzzy rules. Finally, the barycenter method is
used to defuzzify the reasoning results to achieve dynamic
where 𝑢 represents the control input parameter.
real-time update of PID parameters.
The control target of the PMSM position loop is to enable
The fuzzy subset of two-dimensional input variables and
the output to track the reference value quickly and accurately.
three-dimensional output variables is defined as
Let e = 𝜃𝑟 − 𝜃 = 𝑒1 , 𝑒̇ = 𝜃𝑟̇ − 𝜃̇ = 𝑒2 ; then the state equation of
the PMSM can be expressed as {𝑁𝐵, 𝑁𝑀, 𝑁𝑆, 𝑍𝑂, 𝑃𝑆, 𝑃𝑀, 𝑃𝐵} . (7)
𝑒1̇ = 𝑒2 The same membership function is applied to the input
(5) variables and the output variables for real-time calculation
𝑒2 = 𝜃𝑟̇ − 𝜃.̇ and online adjustment. The membership functions 𝑁𝐵 and
𝑃𝐵 adopt Gaussian distribution function, while the rest adopt
3. Design of Fuzzy PID-Variable Structure triangular distribution function [22], as is shown in Figure 3.
Adaptive Controller The fuzzy rules of 󳵻𝐾𝑝 , 󳵻𝐾𝑖 , and 󳵻𝐾𝑑 are shown in
Table 1, where the first column is the fuzzy rules of 󳵻𝐾𝑝 ,
3.1. Design of Fuzzy PID Controller. The fuzzy PID controller the second column is the fuzzy rules of 󳵻𝐾𝑖 , and the third
is designed as a system with two-dimensional input and column is the fuzzy rules of 󳵻𝐾𝑑 . 49 fuzzy control rules can
three-dimensional output. The inputs are the error (𝑒) and the be obtained from Table 1, which will be adopted for fuzzy
error change rate (𝑒𝑐 ) of system. The output (󳵻𝐾𝑝 , 󳵻𝐾𝑖 , and controller designing.
4 Mathematical Problems in Engineering

Table 1: Fuzzy rules table.

ec
e
NB NM NS ZO PS PM PB
NB PS PM NB PS NB NB PB NB NB PM NB NM PM NB NS ZO NB ZO ZO PM ZO
NM PB PS NB PB NM NB PB NM NM PS NM NS PS NM NS ZO NM ZO ZO PS ZO
NS PM PS NM PM NM NM PM NM NS PS NM NS ZO NM ZO NS NM PS NS PS PS
ZO PM ZO NM PM NS NS PS NS NS ZO NS ZO NS NS PS NM NS NM NM ZO NM
PS PS ZO NS PS ZO NS ZO ZO ZO NS ZO PS NS ZO PS NM ZO PM NM ZO PM
PM ZO PB ZO ZO PS ZO NS PS PS NM PS PS NM PS PM NM PS PB NB PB PB
PB ZO PB ZO ZO PM ZO NM PM PM NB PM PM NM PM PM NB PM PB NB PM PB

e
Fuzzy
ec Inference

ΔKp ΔKi ΔKd

d/dt PID y
yr + System
Controller
-

Figure 2: Control schematic of Fuzzy PID.

1
0.8
membership

0.6
0.4
0.2
0
−3 −2 −1 0 1 2 3
Variables

Figure 3: Membership function curve.

The process of adaptively adjusting PID parameters method as

should follow the following principles:
∑𝑛𝑖=1 𝑢𝑖 𝐴 (𝑢𝑖 )
(1) 𝐾𝑝 : When 𝑒 is a positive and larger value, 󳵻𝐾𝑝 is 𝑢 (𝑡) = , (8)
∑𝑛𝑖=1 𝐴 (𝑢𝑖 )
positive, that is, to increase 𝐾𝑝 . However, when 𝑒
is negative and its absolute value is increasing, the where 𝑢𝑖 is the ith element of 𝑢 and 𝐴(𝑢𝑖 ) is the 𝑢𝑖 th
system is in overshoot state, so 𝐾𝑝 should be reduced. membership function of A on 𝑢(𝑡).
When 𝑒 is near zero, if 𝑒𝑐 is negative, the system Then, the output of Fuzzy PID controller can be expressed
overshoot will become larger and larger, so 𝐾𝑝 should as
be reduced; if 𝑒𝑐 is positive, in order to reduce the
deviation, 󳵻𝐾𝑝 should be positive. 𝑑𝑒
𝑢𝑃𝐼𝐷 = 𝐾𝑝 𝑒 + 𝐾𝑖 ∫ 𝑒 + 𝐾𝑑 , (9)
(2) 𝐾𝑖 : Following the principle of integral separation, its 𝑑𝑡
regulation is similar to that of 𝐾𝑝 , but the situation
of integral saturation should be prevented, for fear where 𝐾𝑃 = 𝐾𝑝 󸀠 + 󳵻𝐾𝑝 , 𝐾𝑖 = 𝐾𝑖 󸀠 + 󳵻𝐾𝑖 , and 𝐾𝑑 = 𝐾𝑑 󸀠 + 󳵻𝐾𝑑 .
of system overshoot. If 𝐾𝑝 increases, 𝐾𝑖 should be
reduced accordingly. 3.2. Design of Sliding Mode Variable Structure Controller.
(3) 𝐾𝑑 : When 𝑒 is positive and increasing, increasing 𝐾𝑝 According to state equations (5), the sliding mode surface is
may cause differential spillover. Therefore, if the larger defined as
𝐾𝑝 is taken, the smaller 𝐾𝑑 should be taken.
𝑠 = 𝑐𝑒1 + 𝑒2 , (10)
Finally, the fuzzy quantity of the output variable is
defuzzified as the output of the controller by barycenter where c is a positive constant.
Mathematical Problems in Engineering 5

Fuzzy PID control

Fuzzy
Inference
ΔKp ΔKi ΔKd
ec
de/dt
yr e PID Controller y
+ +
PMSM
- +

c1 + Sign(·) H
+
S
variable structure control

Figure 4: Control schematic of Fuzzy PID-Variable Structure Adaptive Control.

𝑠 = 𝑠̇ = 0 can be obtained according to the condition that The output of Fuzzy PID-Variable Structure Adaptive
the system maintains the motion on the sliding surface. That Controller is
is,
𝑢 = [(𝐾𝑝 󸀠 + 󳵻𝐾𝑝 ) 𝑒 + (𝐾𝑖 󸀠 + 󳵻𝐾𝑖 ) ∫ 𝑒
𝑠̇ = 𝑐𝑒̇ + 𝑒̈ = 𝑐𝑒̇ + (𝜃𝑟̈ − 𝜃)̈ , (11)
(14)
𝑑𝑒
where 𝜃𝑟 is the desired trajectory. + (𝐾𝑑 󸀠 + 󳵻𝐾𝑑 ) ] + ℎ sign (𝑠) ,
From (4), (10), and (11), we have 𝑑𝑡
1 where 𝑠 is the sliding mode surface, [(𝐾𝑝 󸀠 + 󳵻𝐾𝑝 )𝑒 + (𝐾𝑖 󸀠 +
𝑢∗ = (𝑐𝑒̇ + 𝜃𝑟̈ − 𝑔𝜃̇ − 𝑑) . (12)
𝑓 󳵻𝐾𝑖 ) ∫ 𝑒 + (𝐾𝑑 󸀠 + 󳵻𝐾𝑑 )(𝑑𝑒/𝑑𝑡)] is the approximate linear
control item, and ℎ sign (𝑠) is the nonlinear control item.
From (12), it can be seen that the result of equivalent
control quantity 𝑢∗ includes 𝑓, 𝑔, and 𝑑. 𝑓 and 𝑔 are deter-
mined by the system parameters 𝐵 and 𝐽. 𝑑 is determined 4. Proof of Stability
by the system parameters 𝐽 and 𝑇𝐿 . It can be inferred that The controller designed by (14) is used to control PMSM posi-
the equivalent control amount will be changed with system tion loop described in (4) and (5). If the control parameters
parameter perturbation and load disturbance, and then the are selected according to the constraints of (15), the system
input will make the moving point away from the sliding state can converge to the sliding surface and ensure global
surface, which will even lead to negative effect on servo stability of the system.
system.
The switch item should be designed as 𝛼3 < 𝑔𝛼2 + 𝑓ℎ + 𝛼4
(15)
𝑢𝑟 = 𝑘 sign (𝑠) , (13) 𝑓𝑖 𝑘𝑑𝑖 > 𝑐𝑖 + 𝑔𝑖 + 𝑝𝑖 ,
where 𝑘 > 0 is the variable structure gain; 𝑠 is the sliding where 𝛼1 , 𝛼2 , 𝛼3 , and 𝛼4 are the upper bounds of given
mode surface; and sign (∙) is the sign function. position target (𝜃𝑟 ), derivative of given position target (𝜃𝑟̇ ),
the second derivative of given position target (𝜃𝑟̈ ), and load
3.3. Design of Fuzzy PID-Variable Structure Adaptive Con- disturbance d, respectively, and 𝑝 is the parameter of positive
troller. Considering that the equivalent control item in SMC definite matrix defined later. It can be also expressed as
will affect the servo control performance due to modeling 󵄩󵄩 󵄩󵄩
error and inaccurate parameters of system, the effect of the 󵄩󵄩𝜃𝑟 󵄩󵄩 ≤ 𝛼1
equivalent control item in SMC is compensated by PID con- 󵄩󵄩 ̇ 󵄩󵄩
trol item, and then the effect of the nonlinear is compensated 󵄩󵄩𝜃𝑟 󵄩󵄩 ≤ 𝛼2
(16)
by the variable structure control item. Meanwhile, the fuzzy 󵄩󵄩󵄩𝜃𝑟̈ 󵄩󵄩󵄩 ≤ 𝛼3
inference method is used to adjust the parameters of PID in 󵄩 󵄩
real time online, which can attenuate chattering. The fuzzy ‖𝑑‖ ≤ 𝛼4 ,
PID controller can ensure the stability of system, limit the
error to a narrow range, and attenuate the chattering of SMC where ‖ ∙ ‖ is the Euclidean norm.
as well. On the other hand, the variable structure control In the actual control system, the position target input
item is used to compensate the effect of nonlinear, which of the system is bounded, and the perturbation of system
can suppress the effect of parameter perturbation and load parameters and the disturbance of external load will change
disturbance. The schematic diagram of Fuzzy PID-Variable within a certain range. Thus the assumption in (16) is
Structure Adaptive Control is shown in Figure 4. reasonable in the actual control system.
6 Mathematical Problems in Engineering

For n-dimensional system the sliding surface can be In the process of stability analysis, as there is no integral
expressed as term in the design of sliding mode surface, the integral term
in fuzzy PID can be ignored. Then (26) can be obtained:
𝑇
S = [𝑠1 𝑠2 . . . 𝑠𝑛 ] 𝑛 = 1, 2 . . . 𝑛 (17)
𝑠𝑖 𝑠𝑖̇ = (𝑐𝑖 𝑒1𝑖 + 𝑒2𝑖 ) [𝜃𝑟𝑖̈ − 𝑔𝑖 𝜃𝑟𝑖̇ − 𝑓𝑖 ℎ𝑖 sign (𝑠𝑖 ) − 𝑑𝑖 ]
𝑠𝑖 = 𝑐𝑖 𝑒1𝑖 + 𝑒2𝑖 . (18)
− 𝑐𝑖 𝑓𝑖 𝑘𝑝𝑖 𝑒1𝑖 2
The control item in (13) can be expressed as (26)
+ (𝑐𝑖 2 + 𝑐𝑖 𝑔𝑖 − 𝑐𝑖 𝑓𝑖 𝑘𝑑𝑖 − 𝑓𝑖 𝑘𝑝𝑖 ) 𝑒1𝑖 𝑒2𝑖
𝑑𝐸
𝑈 = (𝐾𝑃 𝐸 + 𝐾𝐼 ∫ 𝐸 + 𝐾𝐷 ) + 𝐻 sign (𝑆) . (19)
𝑑𝑡 + (𝑐𝑖 + 𝑔𝑖 − 𝑓𝑖 𝑘𝑑𝑖 ) 𝑒2𝑖 2 .

To construct a control matrix for an 𝑛-dimensional Substituting (26) into (24), the following equation can be
system reasonably, set obtained:
𝑛
𝐾𝑃 = diag {𝐾𝑝1 𝐾𝑝2 . . . 𝐾𝑝𝑛 } 𝑉̇ = ∑ {(𝑐𝑖 𝑒1𝑖 + 𝑒2𝑖 ) [𝜃𝑟𝑖̈ − 𝑔𝑖 𝜃𝑟𝑖̇ − 𝑓𝑖 ℎ𝑖 sign (𝑠𝑖 ) − 𝑑𝑖 ]
𝑖=1
𝐾𝐼 = diag {𝐾𝑖1 𝐾𝑖2 . . . 𝐾𝑖𝑛 } (27)
− 𝑐𝑖 𝑓𝑖 𝑘𝑝𝑖 𝑒1𝑖 2 + (𝑐𝑖 + 𝑔𝑖 − 𝑓𝑖 𝑘𝑑𝑖 + 𝑝𝑖 ) 𝑒2𝑖 2
𝐾𝐷 = diag {𝐾𝑑1 𝐾𝑑2 . . . 𝐾𝑑𝑛 }
𝑇
(20) + (𝑐𝑖 2 + 𝑐𝑖 𝑔𝑖 − 𝑐𝑖 𝑓𝑖 𝑘𝑑𝑖 − 𝑓𝑖 𝑘𝑝𝑖 + 𝑞𝑖 ) 𝑒1𝑖 𝑒2𝑖 } .
𝐸 = [𝑒1 𝑒2 . . . 𝑒𝑛 ]
Because of (23), we can obtain
𝐻 = diag {ℎ1 ℎ2 . . . ℎ𝑛 }
𝑛
sign (𝑆) = diag {sign (𝑠1 ) sign (𝑠2 ) . . . sign (𝑠𝑛 )} . 𝑉̇ = ∑ {𝑠𝑖 [𝜃𝑟𝑖̈ − 𝑔𝑖 𝜃𝑟𝑖̇ − 𝑓𝑖 ℎ𝑖 sign (𝑠𝑖 ) − 𝑑𝑖 ]
𝑖=1 (28)
Construct a positive definite Lyapunov function as
− 𝑐𝑖 𝑓𝑖 𝑘𝑝𝑖 𝑒1𝑖 2 + (𝑐𝑖 + 𝑔𝑖 − 𝑓𝑖 𝑘𝑑𝑖 + 𝑝𝑖 ) 𝑒2𝑖 2 } .
𝑇 𝑇̇ 𝑄𝑃 𝐸
𝑆𝑇 𝑆 [𝐸 𝐸 ] [ 𝑃 𝐺 ] [ 𝐸̇ ] (21) Combining (16), then (28) can be expressed as
V= + .
2 2 𝑛
󵄩 󵄩
In order to guarantee the positive definiteness of the 𝑉̇ ≤ −∑ [󵄩󵄩󵄩𝑠𝑖 󵄩󵄩󵄩 (𝑔𝑖 𝛼2 + 𝑓𝑖 ℎ𝑖 + 𝛼4 − 𝛼3 )]
Lyapunov function (21), the definition of the symmetry 𝑖=1

matrix [ 𝑄 𝑃
𝑃 𝐺 ] is
𝑛
− ∑ [(𝑐𝑖 𝑓𝑖 𝑘𝑝𝑖 ) 𝑒1𝑖 2 ] (29)
𝑄 = diag (𝑞1 𝑞2 . . . 𝑞𝑛 ) 𝑖=1

𝑛
𝑃 = diag (𝑝1 𝑝2 . . . 𝑝𝑛 ) (22) − ∑ [(−𝑐𝑖 − 𝑔𝑖 + 𝑓𝑖 𝑘𝑑𝑖 − 𝑝𝑖 ) 𝑒2𝑖 2 ] .
𝑖=1
𝐺 = diag (𝑔1 𝑔2 . . . 𝑔𝑛 ) ,
It can be deduced that 𝑉̇ is a negative definite matrix when
where 𝑞𝑖 , 𝑝𝑖 , and 𝑔𝑖 satisfy the following relationship: (15) is satisfied. According to Lyapunov’s second method, it
can be concluded that the system is globally stable based on
𝑐𝑖 2 − 𝑐𝑖 𝑓𝑖 𝑘𝑑𝑖 +𝑐𝑖 𝑔𝑖 − 𝑓𝑖 𝑘𝑝𝑖 + 𝑞𝑖 = 0. (23) Fuzzy PID-Variable Structure Adaptive Control.
Differentiating 𝑉, we obtain
5. Simulation Verification
𝑄 𝑃 𝐸̇
𝑉̇ = 𝑆𝑇 𝑆̇ + [𝐸𝑇 𝐸𝑇̇ ] [ ][ ] The model of the position-speed-current three-closed-loop
𝑃 𝐺 𝐸̈ control system for PMSM vector control is built based on
(24) Matlab/Simulink. The speed loop and the current loop adopt
𝑛
2 the conventional PID control while the position loop adopts
= ∑ (𝑠𝑠̇ + 𝑞𝑖 𝑒1𝑖 𝑒2𝑖 + 𝑝𝑖 𝑒2𝑖 ) .
𝑖=1
the Fuzzy PID-Variable Structure Adaptive Control which is
introduced in this paper. The simulation model is shown in
From (4), (5), (11), and (14), (25) can be obtained: Figure 5.
The parameters of the simulation model are shown in
𝑠𝑖̇ = 𝜃𝑟𝑖̈ − 𝑔𝑖 𝜃𝑟𝑖̇ − 𝑓𝑖 𝑘𝑝𝑖 𝑒1𝑖 + (𝑐𝑖 + 𝑔𝑖 − 𝑓𝑖 𝑘𝑑𝑖 ) 𝑒2𝑖 Table 2.
(25) Figures 6–10 show the step response. The input reference
− 𝑓𝑖 𝑘𝑖𝑖 ∫ 𝑒1𝑖 − 𝑓𝑖 ℎ𝑖 sign (𝑠𝑖 ) − 𝑑𝑖 . position is a step signal of 𝜃𝑟 = 0.6(𝑟𝑎𝑑), and the load
disturbance of 50𝑁𝑚 is added at 0.75𝑠. The simulation results
Mathematical Problems in Engineering 7

Fuzzy-PID

SMC

Figure 5: Simulink model.

Step Response
0.7
0.6
0.5
Pos (rad)

Fuzzy-PID-SMC PID
0.4 Fuzzy-PID

0.3
0.2
0.1
0
0 0.5 1 1.5
Time (s)

Figure 6: Step response (a).

Step Response
0.615
0.61
0.605
Pos (rad)

0.6
0.595
0.59 Fuzzy-PID PID
0.585 Fuzzy-PID-SMC
0.58
0.575
0.57
0.565
0.012 0.0125 0.013 0.0135 0.014 0.0145 0.015 0.0155 0.016
Time (s)

Figure 7: Step response (b).

Step Response
0.615
0.61
0.605
Pos (rad)

0.6 Fuzzy-PID-SMC
0.595
0.59
0.585
0.58 Fuzzy-PID PID
0.575
0.57
0.565
0.748 0.75 0.752 0.754 0.756 0.758 0.76 0.762 0.764
Time (s)

Figure 8: Step response (c).

8 Mathematical Problems in Engineering

Table 2: The parameters of the simulation model.

Parameter Value Parameter Value

𝑅𝑠 2.875Ω 𝐾𝑝0 700
𝐿 0.00153H 𝐾𝑖0 6
𝜓𝑓 0.175Wb 𝐾𝑑0 0.1
J 0.0008Kgm2 𝑐 268.5
𝑃𝑛 4 ℎ 2.8
𝐾𝑇 1.05

Step Response Error

0.04
0.03
Pos E (rad)

0.02
0.01 Fuzzy-PID-SMC Fuzzy-PID

0
−0.01 PID

−0.02
0.012 0.0125 0.013 0.0135 0.014 0.0145 0.015 0.0155 0.016
Time (s)

Figure 9: Step response error (a).

Step Response Error

0.04
0.03
Pos E (rad)

0.02 PID

0.01
0 Fuzzy-PID-SMC
Fuzzy-PID
−0.01
−0.02
0.745 0.75 0.755 0.76 0.765 0.77
Time (s)

Figure 10: Step response error (b).

Step Response Speed

30
25
speed (rad/s)

20
15
10
5
0
−5
0 0.5 1 1.5
Time (s)

Figure 11: Step response speed under Fuzzy PID-Variable Structure Adaptive Control.

are compared with conventional PID control and fuzzy PID Control responds fastest and converges to the reference
control. position in the shortest time. In addition, the simulation
From Figure 6, it can be seen that all the three control results show that the chattering of variable structure control
algorithms can obtain better results based on the condition is attenuated obviously due to the adaptive adjustment of PID
that the parameters of the control system are appropriate. control parameters online. Figure 11 is the speed curve (after
However, it can be seen from the details that compared a low-pass filter) of PMSM controlled by Fuzzy PID-Variable
with PID control and fuzzy PID control, the Fuzzy PID- Structure Adaptive Controller. Figure 12 shows the current
Variable Structure Adaptive Control can accelerate system curve (after a low-pass filter) of the 𝑞 axis (the deceleration
convergence, reduce system response time, and obtain better ratio of 50 is considered).
dynamic performance and robustness. Especially after adding Figures 13–15 show the sinusoidal response. The input
load disturbances, the Fuzzy PID-Variable Structure Adaptive reference position of system is a sinusoidal signal (amplitude:
Mathematical Problems in Engineering 9

Step Response Iq
1.2
1
0.8

Iq (A)
0.6
0.4
0.2
0
−0.2
0 0.5 1 1.5
Time (s)

Figure 12: Step response 𝑖𝑞 under Fuzzy PID-Variable Structure Adaptive Control.

Sin Response
0.2
0.15
0.1
Pos (rad)

0.05 Fuzzy-PID
Fuzzy-PID-SMC
0 PID

−0.05
−0.1
−0.15
−0.2
0 0.5 1 1.5
Time (s)

Figure 13: Sinusoidal response (a).

−3 Sin Response
2 x 10
1.5
Pos (rad)

1
0.5
0
Fuzzy-PID-SMC Fuzzy-PID PID
−0.5
−1
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
−4
Time (s) x 10

Figure 14: Sinusoidal response (b).

Sin Response
0.145
0.14
0.135
Pos (rad)

0.13
0.125 Fuzzy-PID PID
0.12
0.115 Fuzzy-PID-SMC
0.11
0.105
0.1
0.75 0.755 0.76 0.765 0.77
Time (s)

Figure 15: Sinusoidal response (c).

0.15𝑟𝑎𝑑, frequency: 10𝐻𝑧), and the load disturbance of Controller. Figure 17 shows the current curve (after a low-
50𝑁𝑚 is added at 0.75𝑠. Compared with Fuzzy PID-Variable pass filter) of the 𝑞 axis (the deceleration ratio of 50 is
Structure Adaptive Control, the other two algorithms exhibit considered).
phase lag. Especially after adding load disturbance, PID By analyzing the step response and sinusoidal response
control and fuzzy PID control exhibit a large static deviation. with load disturbance, the Fuzzy PID-Variable Structure
In addition, the simulation results show that the chattering Adaptive Control can improve the servo precision, the
of variable structure control is attenuated obviously due to dynamic performance, and the robustness of system. At the
the adaptive adjustment of PID control parameters online. same time, the chattering of variable structure control is
Figure 16 is the speed curve (after a low-pass filter) of also attenuated obviously because the PID parameters can be
PMSM controlled by Fuzzy PID-Variable Structure Adaptive adjusted online adaptively by fuzzy inference.
10 Mathematical Problems in Engineering

Sin Response speed

2
1.5

Speed (rad/s)
1
0.5
0
−0.5
−1
−1.5
0 0.5 1 1.5
Time (s)

Figure 16: Sinusoidal response speed under Fuzzy PID-Variable Structure Adaptive Control.

Sin Response Iq
1
0.8
0.6
Iq (A)

0.4
0.2
0
−0.2
0 0.5 1 1.5
Time (s)

Figure 17: Sinusoidal Response 𝑖𝑞 under Fuzzy PID-Variable Structure Adaptive Control.

Encoder module

PWM module
Pos
contr
oller

AD module

Figure 18: RCP model.

6. Experiment Verification Adaptive Control are compared by a step signal. The step
response testing is performed with a constant torque of 5Nm,
The Rapid Control Prototype (RCP) is a kind of semiphysical and the step motion mode is a reciprocating motion with the
simulation. After the mathematical simulation of the system range of 100∘ to 200∘ . The experimental results are shown
control model meets the desired effect, the control algorithm in Figures 21–25; Figure 21 is the position tracking curve,
model of the servo system will be extracted individually. Figure 22 is the position tracking error curve, and Figure 23
Then the Real-Time Driver (RTD) of actual controlled object, is the speed curve. Figure 24 is the original curve of torque
feedback original, and drive unit are added to the control sensor under Fuzzy PID-Variable Structure Adaptive Control
algorithm, which will constitute a closed-loop system [23, (it is basically consistent with the curve of torque sensor
24]. Before the structure design and processing completion under PID control), and Figure 25 is the curve of torque
of extremity exoskeleton, the effectiveness of Fuzzy PID- sensor after a low-pass filter.
Variable Structure Adaptive Control is verified on a semi- It can be seen from Figures 21 and 22 that the positioning
physical simulation test bench. The RCP model is shown in accuracy of conventional PID control and Fuzzy PID-Variable
Figure 18. Structure Adaptive Control is almost the same. However, it
The control system test bench is shown in Figure 19. is evident that the dynamic response of Fuzzy PID-Variable
The control principle of test bench is shown in Figure 20. Structure Adaptive Control outperforms that of conventional
PID control. Because of the unidirectional constant torque,
6.1. Static Testing. The step response performances of con- the forward step response and reverse step response of
ventional PID control and Fuzzy PID-Variable Structure the same control will be different. Through comparison,
Mathematical Problems in Engineering 11

Drive Unit Load

Motor PMSM
Controller control card

Figure 19: The control system test bench.

PC/Windows

Simulator/QNX

I/O Cards
Position Position
feedback feedback
current current
feedback PWM∗6 PWM∗6
feedback

Drive Drive
Unit Unit

V W
U V W U

Torque
LOAD Sensor PMSM

Figure 20: The control principle.

Step Response
220
200
Position (°)

180
160 Fuzzy-PID-SMC

140
Desired Step PID
120
100
80
18 20 22 24 26 28 30 32 34 36 38 40
Time (S)

Figure 21: Step response position tracking.

12 Mathematical Problems in Engineering

Step Response Error

0.5
0.4

Position Error (°)

0.3
0.2
0.1
0
−0.1
−0.2 PID
−0.3 Fuzzy-PID-SMC
−0.4
−0.5
18 20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 22: Step response position tracking error.

Step Response Speed

50
40
30
Speed (RPM)

20
10
0
−10
−20
−30 PID
Fuzzy-PID-SMC
−40
−50
18 20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 23: Step response speed.

Torque data
40
30
Torque (Nm)

20
10
0
−10
−20
−30
−40
18 20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 24: Step response original torque data.

Torque Data
8
7
Torque (Nm)

6
5
4
3
2
1
0
18 20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 25: Step response torque data after low-pass filter.

it is observed that the forward step response and reverse a constant torque of 5Nm. The experimental results are shown
step response under Fuzzy PID-Variable Structure Adaptive in Figures 26–30; Figure 26 is the position tracking curve,
Control are more consistent. Figure 27 is the position tracking error curve, and Figure 28
is the speed curve. Figure 29 is the original curve of torque
6.2. Dynamic Testing. The dynamic response performances sensor under Fuzzy PID-Variable Structure Adaptive Control
of conventional PID control and Fuzzy PID-Variable Struc- (it is basically consistent with the curve of torque sensor
ture Adaptive Control are compared by sinusoidal response under PID control), and Figure 30 is the curve of torque
by a sinusoidal signal (amplitude: 100∘ , frequency: 2Hz) with sensor after a low-pass filter.
Mathematical Problems in Engineering 13

Sin Response
280
260
240

Position (°)
220
200
180
160
140 PID
120 Fuzzy-PID-SMC
100
80
19 20 21 22 23 24 25
Time (s)

Figure 26: Sinusoidal response position tracking.

Sin Response Error

2.5
2
Position Error (°)

1.5
1
PID
0.5
0
−0.5
−1 Fuzzy-PID-SMC
−1.5
−2
−2.5
20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 27: Sinusoidal response position tracking error.

Sin Response Speed

40
30
Speed (RPM)

20
Fuzzy-PID-SMC
10 PID

0
−10
−20
−30
−40
20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 28: Sinusoidal response speed.

Torque data
50
40
Torque (Nm)

30
20
10
0
−10
−20
−30
−40
−50
20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 29: Sinusoidal response original torque data.

Torque data
6
5.8
5.6
Torque (Nm)

5.4
5.2
5
4.8
4.6
4.4
4.2
4
20 22 24 26 28 30 32 34 36 38 40
Time (s)

Figure 30: Sinusoidal response torque data after low-pass filter.

14 Mathematical Problems in Engineering

Both conventional PID control and Fuzzy PID-Variable Acknowledgments

Structure Adaptive Control achieve good performances
under the proper parameter as shown Figure 26. Since the This work was supported by the Chinese National Science
conventional PID control performances are good enough, the Foundation (no. 51075017).
advantages of Fuzzy PID-Variable Structure Adaptive Control
will be more persuasive when we illustrate the tracking errors References
in Figure 27. It is observed that the dynamic response of Fuzzy
PID-Variable Structure Adaptive Control outperforms that [1] X. Zhang, Z. Yue, and J. Wang, “Robotics in lower-limb
of conventional PID control, especially at the near place of rehabilitation after stroke,” Behavioural Neurology, vol. 2017,
the position curve crossing time axis (at this time, the speed Article ID 3731802, 13 pages, 2017.
reaches the maximum.). [2] S. Yu, Z. Wang, and K. Zhang, “Sequential time-dependent
There is no obvious chattering in the response curve reliability analysis for the lower extremity exoskeleton under
of static testing and dynamic testing under Fuzzy PID- uncertainty,” Reliability Engineering & System Safety, vol. 170,
Variable Structure Adaptive Control, which indicates that it pp. 45–52, 2018.
is feasible to introduce fuzzy inference into PID control to [3] Z. Li, W. Ma, Z. Yin, and H. Guo, “Tracking control of time-
attenuate chattering. Finally, we conclude the advantages of varying knee exoskeleton disturbed by interaction torque,” ISA
the proposed Fuzzy PID-Variable Structure Adaptive Control Transactions, vol. 71, pp. 458–466, 2017.
method as high precision, good robustness, and simplicity [4] B. Brahmi, M. Saad, C. Ochoa Luna, M. H. Rahman, and A.
based on experiments. Brahmi, “Adaptive Tracking Control of an Exoskeleton robot
with Uncertain Dynamics Based on Estimated Time Delay
Control,” IEEE/ASME Transactions on Mechatronics, 2018.
[5] Q. Wu, X. Wang, F. Du, and R. Xi, “Modeling and position
7. Conclusion control of a therapeutic exoskeleton targeting upper extremity
The Fuzzy PID-Variable Structure Adaptive Control algo- rehabilitation,” Proceedings of the Institution of Mechanical
Engineers, Part C: Journal of Mechanical Engineering Science,
rithm for the position tracking control of PMSM which
vol. 231, no. 23, pp. 4360–4373, 2017.
will be used in electric extremity exoskeleton robot is pro-
posed. The controller consists of two parts which are mutual [6] J. Qian, A. Xiong, and W. Ma, “Extended state observer-
based sliding mode control with new reaching law for pmsm
compensation: one is a nonlinear part (Variable structure
speed control,” Mathematical Problems in Engineering, vol. 2016,
control), and the other is an approximate linear part (fuzzy Article ID 6058981, 10 pages, 2016.
PID control). Variable structure control has high robustness.
[7] S. Wang, Y. Gao, J. Liu, and L. Wu, “Saturated sliding mode
Fuzzy PID control is used to compensate the equivalent
control with limited magnitude and rate,” IET Control Theory
control item in SMC and attenuate chattering. The simulation & Applications, vol. 12, no. 8, pp. 1075–1085, 2018.
results show that the algorithm can improve control accuracy,
[8] A. V. Topalov, G. L. Cascella, V. Giordano, F. Cupertino, and
dynamic performance, and robustness. The chattering of
O. Kaynak, “Sliding mode neuro-adaptive control of electric
sliding mode system is attenuated to some extent. At the same drives,” IEEE Transactions on Industrial Electronics, vol. 54, no.
time, the Fuzzy PID-Variable Structure Adaptive Control 1, pp. 671–679, 2007.
algorithm has the advantages of simple structure and easy
[9] J. F. Wu, J. Z. Wang, and L. P. Wang, “Adaptive fuzzy sliding
engineering realization as well. Finally, the effectiveness mode control of asymmetrical hydraulic cylinder based on
and practicability of the algorithm are verified based on a DSP,” Advanced Materials Research, vol. 655, pp. 1474–1478,
semiphysical simulation test bench with the method of RCP. 2013.
Undoubtedly, the control algorithm proposed in this
[10] J. C. Zhou, J. P. Tian, and A. H. Tang, “Adaptive Sliding Mode
paper as the lower level control method can be applied Variable Structure Control for Turn Over Robot Based on RBF
in our designed electric exoskeleton successfully. And it is Neural Network,” Automation Instrumentation, pp. 34–36, 2018.
also the foundation and origin of designing the higher level
[11] L. Qi and H. Shi, “Adaptive position tracking control of perma-
control method in our later works on this electric extremity nent magnet synchronous motor based on RBF fast terminal
exoskeleton. sliding mode control,” Neurocomputing, vol. 115, pp. 23–30, 2013.
[12] F.-J. Lin and R.-J. Wai, “Sliding-mode-controlled slider-crank
mechanism with fuzzy neural network,” IEEE Transactions on
Data Availability Industrial Electronics, vol. 48, no. 1, pp. 60–70, 2001.
[13] D. X. Phu, T. D. Huy, V. Mien, and S.-B. Choi, “A new
The simulation and experimental data used to support the composite adaptive controller featuring the neural network and
findings of this study are available from the corresponding prescribed sliding surface with application to vibration control,”
author upon request. Mechanical Systems and Signal Processing, vol. 107, pp. 409–428,
2018.
[14] T. Su, W. Kuo, V. Giap, H. Q. Vu, and Q. Nguyen, “Active
Conflicts of Interest Magnetic Bearing System Using PID-surface Sliding Mode
Control,” in Proceedings of the 2016 Third International Confer-
The authors declare that there are no conflicts of interest ence on Computing Measurement Control and Sensor Network
regarding the publication of this paper. (CMCSN), pp. 5–8, Matsue, Japan, May 2016.
Mathematical Problems in Engineering 15

[15] Z. Tang, P. Pei, M. Wen, and Z. Shi, “Two novel reaching laws
and their use in sliding mode control of EHA,” in Proceedings
of the 2015 IEEE International Conference on Information and
Automation, ICIA 2015 - In conjunction with 2015 IEEE Interna-
tional Conference on Automation and Logistics, pp. 1441–1446,
China, August 2015.
[16] E. M. Jafarov, M. N. A. Parlakçi, and Y. Istefanopulos, “A new
variable structure PID-controller design for robot manipula-
tors,” IEEE Transactions on Control Systems Technology, vol. 13,
no. 1, pp. 122–130, 2005.
[17] W. H. Jiang, Y. C. Fan, Z. Liu, and D. Zeng, “Design of servo
positioning system based on PID variable structure,” Applied
Mechanics and Materials, vol. 321–324, pp. 1463–1467, 2013.
[18] S. Liu and L. Ding, “Application of adaptive fuzzy sliding mode
controller in PMSM servo system,” in Proceedings of the 1st
International Conference on Computing Control and Industrial
Engineering, CCIE 2010, pp. 95–98, China, June 2010.
[19] A. Ayadi, M. Smaoui, S. Aloui, S. Hajji, and M. Farza, “Adaptive
sliding mode control with moving surface: Experimental val-
idation for electropneumatic system,” Mechanical Systems and
Signal Processing, vol. 109, pp. 27–44, 2018.
[20] Z. M. Chen, W. J. Meng, and J. G. Wang, “Fuzzy reaching
law sliding mode control of robot manipulators,” IEEE Pacific-
Asia Workshop on Computational Intelligence and Industrial
Application, vol. 2, pp. 393–397, 2008.
[21] S. S. Yeoh, T. Yang, L. Tarisciotti, C. I. Hill, S. Bozhko,
and P. Zanchetta, “Permanent-Magnet Machine-Based
Starter–Generator System With Modulated Model Predictive
Control,” IEEE Transactions on Transportation Electrification,
vol. 3, no. 4, pp. 878–890, 2017.
[22] C.-C. Tsai, F.-C. Tai, Y.-L. Chang, and C.-T. Tsai, “Adaptive
predictive PID control using fuzzy wavelet neural networks
for nonlinear discrete-time time-delay systems,” International
Journal of Fuzzy Systems, vol. 19, no. 6, pp. 1718–1730, 2017.
[23] A. Deep, N. K. Pilli, A. K. Chauhan, and S. K. Singh, “Design
and development of rapid prototype controlled PMSM drive,”
in Proceedings of the 2015 IEEE IAS Joint Industrial and
Commercial Power Systems / Petroleum and Chemical Industry
Conference, ICPSPCIC 2015, pp. 134–139, India, November 2015.
[24] W. Zhang, Y. Zhang, R. Wang, and X. Pan, “A model-based
DSP control platform for rapid prototype of SVPWM,” in
Proceedings of the 2010 IEEE 10th International Conference on
Signal Processing, ICSP2010, pp. 2523–2526, China, October
2010.
Advances in Advances in Journal of The Scientific Journal of
Operations Research
Hindawi
Decision Sciences
Hindawi
Applied Mathematics
Hindawi
World Journal
Hindawi Publishing Corporation
Probability and Statistics
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018

International
Journal of
Mathematics and
Mathematical
Sciences

Journal of

Hindawi
Optimization
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at

www.hindawi.com

International Journal of
Engineering International Journal of
Mathematics
Hindawi
Analysis
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of Advances in Mathematical Problems International Journal of Discrete Dynamics in

Complex Analysis
Hindawi
Numerical Analysis
Hindawi
in Engineering
Hindawi
Differential Equations
Hindawi
Nature and Society
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

International Journal of Journal of Journal of Abstract and Advances in

Stochastic Analysis
Hindawi
Mathematics
Hindawi
Function Spaces
Hindawi
Applied Analysis
Hindawi
Mathematical Physics
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018