algorithms

Article
Optimization Design by Genetic Algorithm
Controller for Trajectory Control of a 3-RRR
Parallel Robot
Lianchao Sheng and Wei Li *
School of Mechatronic Engineering, China University of Mining and Technology, Xuzhou 221116, China;
shenglianchao@163.com
* Correspondence: liweicumt@cumt.edu.cn; Tel.: +86-152-5203-3709

Received: 21 November 2017; Accepted: 12 January 2018; Published: 15 January 2018

Abstract: In order to improve the control precision and robustness of the existing proportion
integration differentiation (PID) controller of a 3-Revolute–Revolute–Revolute (3-RRR) parallel robot,
a variable PID parameter controller optimized by a genetic algorithm controller is proposed in this
paper. Firstly, the inverse kinematics model of the 3-RRR parallel robot was established according to
the vector method, and the motor conversion matrix was deduced. Then, the error square integral
was chosen as the fitness function, and the genetic algorithm controller was designed. Finally, the
control precision of the new controller was verified through the simulation model of the 3-RRR planar
parallel robot—built in SimMechanics—and the robustness of the new controller was verified by
adding interference. The results show that compared with the traditional PID controller, the new
controller designed in this paper has better control precision and robustness, which provides the
basis for practical application.

Keywords: kinematics model; virtual bench; genetic algorithm controller; robustness

1. Introduction
Parallel industrial robots are essentially closed kinematic chain mechanisms, in which the
end-effector is clamped onto the base through several independent kinematic chains [1]. Parallel
industrial robots exhibit the following obvious advantages over the serial series: no cumulative error,
little movement inertia, high structural stiffness, high load capacity, and so on. They are suitable for
fast and accurate operations, and have been widely applied in industrial fields such as integrated
circuit manufacturing and testing, fiber docking, and flight simulators [2,3]. For instance, Lee and
Yien [4] set up a three degrees of freedom (3-DOF) parallel industrial robot in order to operate and carry
large-scale workpieces. Another 3-DOF micro-sized planar parallel industrial robot was developed
for low-torque precision positioning tasks [5]. Combined with the advantages of parallel robots and
their wide applications in all aspects of social life, the research on parallel industrial robots is of
great significance.
A typical 3-DOF planar 3-Revolute–Revolute–Revolute (3-RRR) parallel industrial robot—
consisting of a moving platform which is connected to three identical branches and a static
platform—has the advantages of a simple structure, good symmetry and accurate positioning.
Therefore, many scholars have conducted detailed research on it. Gutierrez [6] performed an exhaustive
singularity analysis of a 3-RRR planar parallel robot. From a kinematic point of view, the optimum
characteristics of the 3-DOF planar parallel industrial robot were studied by Gosselin and Angeles [7].
Serdar Kucuk [8] presented an optimization problem for the 3-degrees-of-freedom RRR fully planar
parallel manipulator (3-RRR) based on the actuator power consumption. The effects of clearance
on the dynamics of a planar 3-RRR parallel manipulator are investigated in [9]. Based on forward

Algorithms 2018, 11, 7; doi:10.3390/a11010007 www.mdpi.com/journal/algorithms

Algorithms 2018, 11, 7 2 of 13

kinematics theory, Flavio and Ron [10] studied the singularity problem, focusing completely on the
planar parallel robot. Sébastien and Ilian [11] studied the accuracy problems on several kinds of typical
3-DOF planar parallel industrial robots. Gao [12] designed a planar 3-RRR parallel industrial robot by
analyzing kinematics, singularity, dynamics, and controlling problems. Zhang et al. [13] presented the
first-order approximation coupling model of planar 3-RRR flexible parallel robots and considered the
rigid body motion constraints, elastic deformation motion constraints, and dynamic constraints of the
moving platform in the coupling model. Yu [14] presented experimental research on the dynamics
of flexible 3-RRR parallel robots and compared the results between the theoretical and experimental
analysis. It can be seen from the above literature that the current research on the 3-RRR parallel robot
mainly focuses on the dynamics model, work space analysis, singular point analysis, and structural
optimization. Good control precision also plays a crucial role in the practical application of a robot.
At present, the research on robot control algorithms mainly focuses on fuzzy control
algorithms [15], adaptive robust control algorithms [16], and genetic control algorithms [17].
Stanet et al. [18] proposed a fuzzy controller for a 3-DOF parallel industrial robot in medical
applications, and achieved more precise results with the fuzzy-based controller than with the classic
PID controller. Noshadi et al. [19,20] introduced a method by which an active force can be used to
control the platform. They also designed a two-level fuzzy tuning resolved acceleration control for
the platforms. In doing so, they suggested that when there is disturbance, the manipulator can still
perform trajectory tracking tasks with a stable response. Pham et al. [21] proposed a robust adaptive
control method based on radial basis function neural network (RBFNN), and studied the trajectory
tracking control of a two-link robot cleaning detection mechanism. In research by Ku et al. [22]
the model predictive control was performed to fulfill the positioning control of a 3-RRR planar
parallel manipulator. Song et al. [23] proposed a new path planning method for a robot based on a
genetic algorithm and the Bezier curve. Wang et al. [24] presented a double global optimum genetic
algorithm–particle swarm optimization for path planning of a welding robot. Lu et al. [25,26] proposed
a new method for the optimal design and tuning of a Proportional–Integral–Derivative-type interval
type-2 fuzzy logic controller for Delta parallel robot trajectory tracking control. The main aims of the
above literature are to improve the robot control accuracy. However, a good control program needs
to not only improve the control accuracy but also have the characteristics of simplicity, applicability,
robustness, and stability.
It is well known that the classic PID controllers have the advantage of simplicity, and the genetic
random search algorithm has a fast search ability. A variable PID parameter controller optimized by
a genetic algorithm has been designed in this paper, differing from previous research by combining
both advantages. The inverse kinematics model of the 3-RRR parallel robot was established according
to the vector method, and the motor conversion matrix was deduced. The control precision of the
new controller was verified through the simulation model of the 3-RRR planar parallel robot—built
in SimMechanics—and the robustness of the new controller was verified by adding interference.
The results show that compared with the traditional PID controller, the new controller designed in
this paper not only has better control precision, stability, and higher robustness but also has the
characteristics of simplicity and applicability, which provides the basis for practical application. The
other parts of this paper are arranged as follows: Section 2 introduces the model of the planar parallel
industrial robot; Section 3 presents the virtual bench setup and genetic algorithm controller; Section 4
discusses the simulation results; finally, Section 5 is the conclusion.

2. The Model of the Planar 3-RRR Parallel Robot

It is essential to establish the kinematics model for researching the planar parallel industrial
robot control strategies. In this section, the inverse kinematics model of the 3-RRR parallel robot is
established according to the vector method, and the motor conversion matrix is deduced, which is the
foundation for the subsequent control.
Algorithms 2018, 11, 7 3 of 13
Algorithms 2018, 11, 7 3 of 13
Algorithms 2018, 11, 7 3 of 13

2.1. The Structure

2.1. The Structure ofof the
the 3-RRR
3-RRR Planar
Planar Parallel
Parallel Robot
Robot
2.1. The Structure of the 3-RRR Planar Parallel Robot
The
The three-dimensional
three-dimensional map map of of the
the 3-RRR
3-RRR parallel
parallel robot robot is is shown
shown in in Figure
Figure 1.1. The
The schematic
schematic
The three-dimensional map of the 3-RRR parallel robot is shown in Figure 1. The schematic
diagram
diagram of the 3-RRR planar parallel robot is shown in Figure 2. The outline of the 3-RRR planar
of the 3-RRR planar parallel robot is shown in Figure 2. The outline of the 3-RRR planar
diagram of the 3-RRR planar parallel robot is shown in Figure 2. The outline of the 3-RRR planar
parallel
parallel robot robot isis composed
composed of of thethe regular
regular triangle
triangle of of thethe moving
moving platformplatform CC11CC22C C33;; the
the fixed
fixed base base
parallel robot is composed of the regular triangle of the moving platform C1C2C3; the fixed base
A A A ; and three identical branches A B C , A B C , A B C .
A1A2A3; and three identical branches A1B1C1, A2B2C2, A3B3C3. As Figure 2 shows, the three vertices of
1 2 3 1 1 1 2 2 2 3 3 3 As Figure 2 shows, the three vertices
A1A2A3; and three identical branches A1B1C1, A2B2C2, A3B3C3. As Figure 2 shows, the three vertices of
of
thethe regular
regular triangle
triangle A1,AA
areare 1 ,2,AA23,,A 3 , where
where thethe active
active revolute
revolute jointsjointsareare
set.set.
TheThe points
points of B of1, BC11,, C B22,,
B12,, C
the regular triangle are A1, A2, A3, where the active revolute joints are set. The points of B1, C1, B2, C2,
C B33 ,are
B32,, C C3 theare the passive
passive revolute
revolute joints,
joints, where
where A1BA11=B1A= 2BA 2 BA
2 = B3A=3 B
2 3= l11, B
=1C
l13, B = 1BC21C=
2 =BB2C
3C2 3== Bl23.CThe
3 = lradius
2 . The
B3, C3 are the passive revolute joints, where A1B1 = A2B2 = A3B3 = l1, B1C1 = B2C2 = B3C3 = l2. The radius
radius of the circle of the motor in the working platform is R. The
of the circle of the motor in the working platform is R. The radius of the circle of the joints in the radius of the circle of the joints in
of the circle of the motor in the working platform is R. The radius of the circle of the joints in the
the platform is r.
moving platform is r. The 3-DOF planar motion can be achieved using the three driving motors in
moving The 3-DOF planar motion can be achieved using the three driving motors in
moving platform is r. The 3-DOF planar motion can be achieved using the three driving motors in
the
the three
three active
active joints.
joints.
the three active joints.

Figure 1.
Figure1. Three-dimensional map
Three-dimensionalmap
1.Three-dimensional
mapof the
Figure
of
ofthe 3-Revolute–Revolute–Revolute
the3-Revolute–Revolute–Revolute (3-RRR) parallel
3-Revolute–Revolute–Revolute (3-RRR)
(3-RRR) parallel robot.
robot.
parallel robot.

Figure 2. Diagram of the 3-RRR parallel robot.

Figure
Figure2.
2. Diagram
Diagramof
ofthe
the3-RRR
3-RRRparallel
parallelrobot.
robot.

2.2. The Kinematic Model of the 3-RRR Planar Parallel Robot

2.2. The Kinematic Model of the 3-RRR Planar Parallel Robot
According to the geometric relationships shown in Figure 2, the coordinates of the motor in the
According to the geometric relationships shown in Figure 2, the coordinates of the motor in the
static coordinate system are shown as follows,
static coordinate system are shown as follows,
Algorithms 2018, 11, 7 4 of 13

2.2. The Kinematic Model of the 3-RRR Planar Parallel Robot

According to the geometric relationships shown in Figure 2, the coordinates of the motor in the
static coordinate system are shown as follows,
" √ √ #
3
0 3R 2 R
h i
A= A1 A2 A3 = 3
(1)
0 0 2R

The coordinates of the center of the moving platform in the moving coordinate system are shown
as follows, " √ √ #
− 23 r 23 r 0
h i
P = C1 C2 C3 = (2)
− 12 r − 12 r r
The spatial transformation matrix T for transforming the moving coordinate system into the static
coordinate system is shown as follows,
 
cos β cos γ sin β cos γ sin α − sin γ cos α sin β cos γ cos α + sin γ sin α x
 sin γ cos β sin γ sin β sin α + cos γ cos α sin γ sin β cos α − cos γ sin α y 
T= (3)
 
− sin β cos β sin α cos β cos α z

 
0 0 0 1

where α, β, γ are the rotation angles around X, Y and Z, respectively. x, y, z are the distances along
the X, Y and Z axes, respectively. Due to the parallel robot studied in this paper, they can only be
translated along the x, y axis and rotated around the z axis,

α = 0, β = 0, z = 0 (4)

T can be expressed as follows,

 
cos γ − sin γ 0 x
 sin γ cos γ 0 y 
T= (5)
 
0 0 1 0

 
0 0 0 1

The coordinates of the moving platform center in the static coordinate system can be expressed
as follows, " # " #
C P
=T× (6)
1 1

Combining Equations (2) and (5), C can be expressed as follows,

"

#
h i x − r cos π
6 − γ x + r cos γ − π6 x − r sin γ
C= C1 C2 C3 =

(7)
y − r sin π
6 +γ y + r sin γ − π6 y + r cos γ

For any one branch Ai Bi Ci , the parallel robot satisfies the following motion constraint of the
closed-loop vector. The vector equation is expressed as follows [8],
* * * *
OCi = OAi + Ai Bi + Bi Ci (8)

where i = 1, 2, 3.
Combined with Equation (8), θ Bi is the angle between the link Bi Ci and the positive direction of
the X axis.
Algorithms 2018, 11, 7 5 of 13

Algorithms 2018, 11, 7 5 of 13

 xCi = xAi + l1 cos θ Ai + l2 cos θ Bi
 (9)
 yCi = y Ai + l1 sin θ Ai + l2 sin θ Bi
(
Equation (9) can be expressed as xfollows,
Ci = x Ai + l1 cos θ Ai + l2 cos θ Bi
(9)
yCi = y Ai + l1 sin θ Ai + l2 sin θ Bi
fi1 sin θ Ai + fi 2 cos θ Ai + fi 3 = 0 (10)
Equation (9) can be expressed as follows,
where

fi1 = 2l1 ( y Aif i1−sin

yCi θ) Ai + f i2 cos θ Ai + f i3 = 0 (10)
fi 2 = 2l1 ( xAi − xCi )
where
fi 3 = xCi2 + yCi2 + xAi2f i1+=
y Ai2 2l−12(yxCiAix− y 2 y)Ci y Ai + l12 − l22
Ai − Ci
f i2 = 2l1 ( x Ai − xCi )
Let ti = tan (θi / 2) , fsubstituting
i3 = x 2 + y2 Equation
+ x 2 + y2(10), − 2x according to the universal
Ci x Ai − 2yCi y Ai + l − l
2 2 formula of the
Ci Ci Ai Ai 1 2
trigonometric function, Equation (10) can be expressed as follows,
Let ti = tan(θi /2), substituting Equation (10), according to the universal formula of the
trigonometric function, Equation( f(10)
i 3 − fi 2 ) ti + 2 fi1ti + ( fi 3 + fi 2 ) = 0
can be 2
expressed as follows, (11)

( f i3 −the
According to the rooting formula, )t2i + can
f i2roots 2 f i1 tbe
i +expressed
( f i3 + f i2 )as=follows,
0 (11)


According to the rooting formula,  − f can
the roots + be fi12 + fi 22 − fi 32  as follows,
expressed
θ Ai1 = 2arctan  i1 
   fi 3 −√fi 2 
 1  2 2 2
 θ = 2arctan − f i1 + f i1 + f i2 − f i3 (12)
 2
Ai  − f − f 2 √ f i3 − 2f i2 2 
i1 + f2i 2 − 2 f i 3 2  (12)
θAi =2 2arctan  i1
− f i1 − f i1 + f i2 − f i3
 θ Ai = 2arctan fi 3 −f i3fi− 
 2 f i2


Similarly,θ Bi
Similarly, θ Bi can
can be
be obtained.
obtained.
In In order
order to to build
build thethe genetic
genetic algorithm
algorithm controller,
controller, it is
it is necessary
necessary to to obtain
obtain thethe transformation
transformation
matrix of the motor physical model. The physical model of the three-phase asynchronous
matrix of the motor physical model. The physical model of the three-phase asynchronous motor motor
is is
shown
shown in in Figure
Figure 3. 3.

Figure
Figure 3. Physical
3. Physical model
model of the
of the three-phase
three-phase asynchronous
asynchronous motor.
motor.

The
Themodel
modelof ofthethethree-phase
three-phaseasynchronous
asynchronousmotor
motorcancanbebeestablished
establishedbybycoordinate
coordinate
transformation. The transformation matrix from the three-phase coordinate system to the two-phase
transformation. The transformation matrix from the three-phase coordinate system to the two-phase
stationary coordinate
stationary coordinatesystem
systemthrough thethe
through geometric relationship
geometric can
relationship bebe
can expressed
expressedasas
follows [27],
follows [27],
r " 1
#
2 1 −
√2
−√21
C3s/2s = 3 (13)
3 0 2 − 23
  1 1 
1 − −
2 2 2 
C3 s/2s =   (13)
3 3 3
0 2 −
Algorithms 2018, 11, 7 2  6 of 13

The transformation matrix from the two-phase stationary coordinate system to the two-phase
rotation
Thecoordinate system
transformation can be
matrix expressed
from as follows
the two-phase [27],
stationary coordinate system to the two-phase
rotation coordinate system can be expressed as follows [27],
 cos θe sin θe 
C2s/2r =
"
θe θ #
(14)
- sin θ cossin
cos θe e e
C2s/2r = (14)
− sin θe cos θe
3. Virtual Bench Setup and Genetic Algorithm Controller
3. Virtual Bench Setup and Genetic Algorithm Controller
To verify the control effect of the PID controller optimized by the genetic algorithm, a virtual
bench Toisverify the control
constructed effect of the PIDAt
in SimMechanics. controller
the sameoptimized by the genetic
time, to simulate algorithm,
the gap a virtual
error between thebench
two
is constructed in SimMechanics. At the same time, to simulate the gap error between
revolute joints in the actual system in the establishment of a three-dimensional model, the axis the two revolute
joints in the
distance actual the
between system
twoin the establishment
revolute joints is setoftoa0.5
three-dimensional
mm. As Figure 4model,
shows,the axis
the distance
virtual between
bench of the
the two 3-RRR
planar revolute joints isindustrial
parallel set to 0.5 robot
mm. As Figure 4 shows,
is translated from theits virtual benchinto
Pro/E form of the planar 3-RRR
Simulink in the
parallel industrial
SimMechanics robot is translated from its Pro/E form into Simulink in the SimMechanics toolbox.
toolbox.

Figure 4.
Figure 4. The
The virtual bench of
virtual bench of the
the planar
planar 3-RRR
3-RRR parallel
parallel industrial
industrial robot.
robot.

According to the industrial robot structure in Figure 1, a virtual bench can be established as
According to the industrial robot structure in Figure 1, a virtual bench can be established as shown
shown in Figure 4. The parallel industrial robot is composed of three elements, six links, one moving
in Figure 4. The parallel industrial robot is composed of three elements, six links, one moving platform,
platform, and three driving joints. The moving platform is connected by three branches which consist
and three driving joints. The moving platform is connected by three branches which consist of two
of two individual links. Each of the two links are connected by revolute joints. The three driving joints
individual links. Each of the two links are connected by revolute joints. The three driving joints on
on the static platform can constitute an equilateral triangle. Similarly, the three connection joints on
the static platform can constitute an equilateral triangle. Similarly, the three connection joints on the
the moving platform can also constitute an equilateral triangle.
moving platform can also constitute an equilateral triangle.
The control scheme is significant for achieving the good stability and precise positioning needed
The control scheme is significant for achieving the good stability and precise positioning needed
by the 3-RRR planar parallel industrial robot. As we already know, the traditional PID controllers
by the 3-RRR planar parallel industrial robot. As we already know, the traditional PID controllers have
have good control of linear objects and their great advantages are simplicity and convenience.
good control of linear objects and their great advantages are simplicity and convenience. However,
However, when the control reference is complicated in the trajectory tracking control, or when the
when the control reference is complicated in the trajectory tracking control, or when the dynamical
dynamical parameters are changeable and complex, the classic PID control may not rapidly present
parameters are changeable and complex, the classic PID control may not rapidly present a good result.
a good result. In order to make the controller perform better, the genetic algorithm is introduced in
In order to make the controller perform better, the genetic algorithm is introduced in this paper. It is a
this paper. It is a random search algorithm of natural selection and genetic mechanism which
random search algorithm of natural selection and genetic mechanism which possesses the advantage
possesses the advantage of simplicity, versatility, robustness, and speed compared to the
of simplicity, versatility, robustness, and speed compared to the conventional method. The design flow
conventional method. The design flow chart of the genetic algorithm controller is shown in Figure 5.
chart of the genetic algorithm controller is shown in Figure 5.
The design process of the genetic algorithm is as follows:
The design process of the genetic algorithm is as follows:
Step 1: Parameter encoding: Due to the need-tuning parameters in the real domain, the simple binary
Step 1: Parameter
encodingencoding: Due six
and encoder, to the need-tuning
sub-strings can parameters
be used as inthethe
sixreal domain, the
parameters simple binary
influencing each
encoding and encoder, six sub-strings can be used as the six parameters influencing each other.
They then formed a chromosome. The precision of the PID solution parameters is set to 4 digits
after the decimal point, and the solution space of the PID parameters can be divided into
(1000 − 0) × (104 ) = 10,000,000 points. Because 223 < 10,000,000 < 224 , 24-bit binary numbers are
needed to represent these solutions. In other words, a solution code is a 24-bit binary string.
 Algorithms 2018, 11, 7 7 of 13

Algorithms 2018, 11, 7 7 of 13

other. They then formed a chromosome. The precision of the PID solution parameters is set
to 4 digits after the decimal point, and the solution space of the PID parameters can be divided
Algorithmsother. They
2018, 11, 7 then formed a chromosome. The precision of the PID solution parameters 7isofset 13
into (1000 − 0) × (104) = 10,000,000 points. Because 223 < 10,000,000 < 224, 24-bit binary numbers
to 4 digits after the decimal point, and the solution space of the PID parameters can be divided
are needed to represent these solutions. In other words, a solution code is a 24-bit binary
into (1000 − 0) × (104) = 10,000,000 points. Because 223 < 10,000,000 < 224, 24-bit binary numbers
string. these
Initially, Initially, these
binary binary
strings are strings
randomly aregenerated.
randomly One generated. One such
such binary stringbinary string
represents a
are needed to represent these solutions. In other words, a solution code is a 24-bit binary
represents astring
chromosome chromosome
where the string
length where thechromosome
of the length of thestring
chromosome
is 24. string is 24.
string. Initially, these binary strings are randomly generated. One such binary string
Step
Step 2: 2:The
The fitness
fitness function
function is the
is the basisbasis
for for evaluating
evaluating the selection,
the selection, where where the error
the error integral
integral can becan be
used
represents a chromosome string where the length of the chromosome string is 24.
asused as the performance
the performance of the system,
of the system, such assuch as Integrator
Integrator Error
Error (IE), (IE), Integrator
Integrator Absolute
Absolute Error
Error (IAE),
Step 2: The fitness function is the basis for evaluating the selection, where the error integral can be
(IAE), IntegralError
Integral Square Error (ISE), Integration Time andError
Absolute Error (ITAE), etc. Each
used asSquare
the performance (ISE),
ofIntegration
the system,Time such and Absolute
as Integrator (ITAE),
Error (IE), etc. Each
Integrator formula
Absolute has
Error
itsformula
own focus. hasThis
its own
paperfocus.
wantsThis paper awants
to achieve smaller to dynamic
achieve aerror,
smaller
so thedynamic
integral error,
squareso the
error
(IAE), Integral Square Error (ISE), Integration Time and Absolute Error (ITAE), etc. Each
integral
isformula
used as the square error
fitness is used as the fitness Zfunction:
function:
has its own focus. This paper wants ∝ to achieve a smaller dynamic error, so the
∝ F = [e (t)]2 (15)
integral square error is used asFthe = fitness
[e（ t） function:
]2 0 (15)
0
∝
Step 3: Genetic algorithm parameters: FGroup
Step 3: Genetic algorithm parameters: Group =  [e（ t）]2nn==20,
size
size
Crossover probability Pc = 0.8, Mutation
20, Crossover probability Pc = 0.8,(15)
Mutation
0
probability Pm = 0.005.
probability Pm = 0.005.
Step
Step 4: 3: Genetic algorithm parameters: Group size n = 20, Crossover probability Pc = 0.8, Mutation
Step 4:If Ifthe
theerror
erroris isminimum,
minimum,then
thenrunning
runningisisstopped,
stopped,as
asthe
theoutput
output has
has reached
reached the
the optimal
optimal
probability
parameters. Pm = 0.005.
Otherwise, Steps 1 to 3 are repeated.
parameters. Otherwise, Steps 1 to 3 are repeated.
Step 4: If the error is minimum, then running is stopped, as the output has reached the optimal
parameters. Otherwise, Steps 1 to 3 are repeated.

Figure 5. The design flow chart of the genetic algorithm controller.

Figure 5. The design flow chart of the genetic algorithm controller.
Figure 5. The design flow chart of the genetic algorithm controller.
The stability of the controller is of great importance to the reliability of the control system. To
further
Thetheoretically
stability of analyze
ofthe the stability
thecontroller
controller of theimportance
PID controller theoptimized by aofgenetic algorithm for
The stability isisofofgreat
great
importance to to reliability
the reliability of thethe control
control system.
system. To
the 3-RRR
To further parallel robot,
theoretically the Routh
analyze stability criterion is introduced for stability verification of the PID
further theoretically analyze the the stability
stability of the
of the PIDPID controller
controller optimized
optimized by abygenetic
a genetic algorithm
algorithm for
controller
for the 3-RRRoptimized
parallel by a genetic
robot, the algorithm.
Routh stabilityThe systemis PID
criterion closed-loop
introduced for controlverification
stability block diagramofPIDis
the
the 3-RRR parallel robot, the Routh stability criterion is introduced for stability verification of the
shown
PID in Figure
controller 6.
optimized
controller optimized by abygenetic
a genetic algorithm.
algorithm. TheThe system
system PIDclosed-loop
PID closed-loopcontrol
controlblock
block diagram
diagram is is
shown in
shown in Figure
Figure 6.
6.

Figure 6. System proportion integration differentiation (PID) closed-loop control block diagram.

Figure 6. System proportion integration differentiation (PID) closed-loop control block diagram.
Figure 6. System proportion integration differentiation (PID) closed-loop control block diagram.
Algorithms 2018, 11, 7 8 of 13

According to the closed-loop PID control block diagram shown in Figure 6, the characteristic
equation of the closed-loop control transfer function can be expressed as follows,

bK p Kd s3 + bK p + aK p Kd s2 + aK p + bK p Ki s + aK p Ki = 0



(16)

The Routh-table of the characteristic equation is expressed as follows,

s3 bK p Kd aK p + bK p Ki
s2 bK p + aK p Kd aK p Ki
(bK p +aK p Kd )( aK p +bK p Ki )−(bK p Kd )( aK p Ki ) (17)
s2 bK p + aK p Kd
s0 aK p Ki

In order to build PID controllers conveniently, the circular equation is divided into two input
equations ( XP = 0.03 − 0.03 cos(50t) (Equation (a)), YP = 0.03 sin(50t) (Equation (b))). Therefore, two
PID controllers optimized by a genetic algorithm are set. The parameters in system transfer function
are given as a = 0.7374, b = 0.000036.
When there is no interference added, the PID control parameters of the PID controller optimized
by a genetic algorithm for Equation (Equation (a)) are Kp1 = 991.1899, Ki1 = 152.2033, Kd1 = 9.9589.
The PID control parameters of the PID controller optimized by a genetic algorithm for Equation
(Equation (b)) are Kp2 = 0.3, Ki2 = 799.9985, Kd2 = 0.0003.
When there is interference added, the PID control parameters of the PID controller optimized by
a genetic algorithm for Equation (Equation (a)) are Kp1 = 100, Ki1 = 198.9589, Kd1 = 9.7398. The PID
control parameters of the PID controller optimized by a genetic algorithm for Equation (Equation (b))
are Kp2 = 0.3, Ki2 = 799.4141, Kd2 = 0.0003.
Substituting the above parameters into the Routh-table Equation (17), it can be seen that none of
the coefficients of the eigenvalue equation is zero, and the sign of each coefficient is positive; therefore,
according to the Routh stability criterion, the system can be proved to be stable.

4. Simulation and Results Discussion

To verify the control effect of the PID controller optimized by a genetic algorithm, the circular
trajectory which is commonly used in industrial applications is selected as the trajectory. The system
control block diagram is shown in Figure 7. The trajectory equation is as follows,

XP = 0.03 − 0.03 cos(50t)

YP = 0.03sin(50t) (18)
θ = π/6

The radius of the circle where the active revolute joints are located is R = 0.25 m. The radius of
the circle where the triangle of the mobile platform is located is r = 0.0825 m; the length of the links
l1 = l2 = 0.16 m; the concentrated mass of the links mg = 0.2 kg; the concentrated mass of the moving
platform mp = 1 kg. To better describe the trajectory tracking error, ei (i = 1, 2) are defined as the
tracking errors. The control parameters of the classic PID controller are Kp 1 = 10, Ki1 = 100, Kd1 = 5,
Kp2 = 0.35, Ki2 = 500, Kd2 = 0.0003.
 q
2
2
 e =
1 x p − xd + y p − yd
q
2
2 (19)
 e =
2 x g − xd + y g − yd

where

xd is the x coordinate of the desired trajectory.

yd is the y coordinate of the desired trajectory.
Algorithms 2018, 11, 7 9 of 13

xp is the x coordinate of the trajectory obtained by the classic PID controller.

yp is the y coordinate of the trajectory obtained by the classic PID controller.
xg is the x coordinate of the trajectory obtained by the PID controller optimized by a genetic algorithm.
yg is the y coordinate of7the trajectory obtained by the PID controller optimized by a genetic
Algorithms 2018, 11, 9 of 13algorithm.

The trajectory trackingoferrors

xg is the x coordinate of theobtained
the trajectory classic byPIDtheand the PID optimized
PID controller optimized by by a genetic
a genetic algorithm are
algorithm.
y is the y coordinate of the trajectory obtained by the PID controller optimized
shown in Figure 8. It is found that the PID controller optimized by the genetic algorithm has a higher
g by a genetic algorithm.
Algorithms 2018, 11, 7 9 of 13
control accuracy. The trajectories
The trajectory tracking errors of the end-effector
of the classic PID andusing theoptimized
the PID classic PID and the
by a genetic PID optimized
algorithm
xare
by a genetic shown
g isalgorithm
the in Figure
x coordinate 8.the
areofshownIt is found
trajectory that the9.
obtained
in Figure PIDIt controller
by the
canPID optimized
becontroller
found bythe
the running
optimized
that genetic algorithm
by a genetic has aof the PID
algorithm.
trajectory
yhigher
g is the y control accuracy.
coordinate The trajectories
of the trajectory obtainedofbythe
theend-effector using
PID controller the classic
optimized PID and
by a genetic the PID
algorithm.
controller optimized by a genetic algorithm is closer to the desired trajectory; therefore, the genetic
optimized by a genetic algorithm are shown in Figure 9. It can be found that the running trajectory
The trajectory
algorithm-optimized PIDtracking
controller errorsdesigned
of the classic PID
thisand the PID
hastooptimized by a genetic algorithm accuracy.
of the PID controller optimized by a genetic in
algorithmpaper
is closer athe
higher
desiredtrajectory
trajectory; tracking
therefore,
are shown in Figure 8. It is found that the PID controller optimized by the genetic algorithm has a
Moreover,thethegenetic
PID controller obtained
algorithm-optimized PIDbycontroller
a genetic algorithm
designed saveshas
in this paper more time
a higher than the
trajectory PID controller
tracking
higher control accuracy. The trajectories of the end-effector using the classic PID and the PID
obtained byaccuracy.
the Moreover,
trial and the PID
error controller
method, obtained
which is by a genetic
simple and algorithm
practical.saves more time than the PID
optimized by a genetic algorithm are shown in Figure 9. It can be found that the running trajectory
controller obtained by the trial and error method, which is simple and practical.
of the PID controller optimized by a genetic algorithm is closer to the desired trajectory; therefore,
the genetic algorithm-optimized PID controller designed in this paper has a higher trajectory tracking
accuracy. Moreover, the PID controller obtained by a genetic algorithm saves more time than the PID
controller obtained by the trial and error method, which is simple and practical.

Figure 7. System control block diagram.

Figure 7. System control block diagram.
To further verify the robustness of the controller designed in this paper, an artificial disturbance
is added to the model. The curve of the disturbance is as follows:
To further verify the robustnessFigure
of the controller
7. System controldesigned in this paper, an artificial disturbance is
block diagram.
x d = 0.003cos(50
added to the model. The curve of the disturbance t)
is as follows: (20)
d = 0.003s in(50t )
To further verify the robustness of theycontroller designed in this paper, an artificial disturbance
is added to the model. The curve of thexdisturbance is
= 0.003 cos as(follows:
50t)
d
(20)
d = 0.003
ydx= 0.003cos(50t) )
sin(50t
(20)
yd = 0.003s in(50t )

Figure 8. The trajectory tracking errors of the classic PID and the PID optimized by a genetic algorithm.

Figure 8. The trajectory tracking errors of the classic PID and the PID optimized by a genetic algorithm.
Figure 8. The trajectory tracking errors of the classic PID and the PID optimized by a genetic algorithm.
Algorithms 2018, 11, 7 10 of 13
Algorithms 2018, 11, 7 10 of 13

Algorithms 2018, 11, 7 0.03 10 of 13

0.02 Desired
0.03
Classic
Genetic algorithm
0.01
0.02 Desired
Y(m) Classic
0 Genetic algorithm
0.01

-0.01
Y(m)

-0.02
-0.01

-0.03
-0.02
0 0.01 0.02 0.03 0.04 0.05 0.06
-0.03 X(m)

Figure 9. Trajectories of the end-effector using the classic PID and the PID optimized by a genetic
Figure 9. Trajectories of the 0 end-effector
0.01 using0.03
0.02 the classic
0.04 PID
0.05 and0.06
the PID optimized by a
algorithm.
genetic algorithm. X(m)

The trajectory
Figure tracking
9. Trajectories of theerrors of the using
end-effector classic
thePID andPID
classic theand
PIDthe
optimized by a genetic
PID optimized algorithm
by a genetic
are algorithm.
The shown
trajectoryin Figure
tracking 10. The maximum
errors error isPID
of the classic still and
less than 1 mm.
the PID It is found
optimized bythat the controller
a genetic algorithm
designed
are shown in this paper
in Figure 10. The stillmaximum
has a highererroraccuracy when
is still lessthe interference
than 1 mm. It is is
added.
found The trajectories
that of
the controller
The trajectoryusing
tracking errors of theandclassic PID optimized
and the PID optimized by a genetic algorithm
designed in this paper still has a higher accuracy when the interference is added. The trajectoriesin
the end-effector the classic PID the PID by a genetic algorithm are shown of the
are shown
Figure 11.inIt Figure
can be 10. found Thethatmaximum error trajectory
the running is still lessofthan 1 mm. It is found that theby controller
end-effector using the classic PID and the PID optimized bythe PID
a geneticcontroller
algorithm optimized
are shown a genetic
in Figure 11.
designed
algorithm in this paper to
is closer stillthe
hasdesired
a highertrajectory.
accuracy when the interference
Therefore, the genetic is added. The trajectoriesPID
algorithm-optimized of
It can
thebeend-effector
found thatusing the running trajectory of
thethe PID controller optimized by a genetic algorithm is
controller designed inthe thisclassic
paper PID
has aand
higher PID optimized
trajectory by aaccuracy
tracking genetic and
algorithm
has betterare shown in
robustness
closer to
Figurethe
when 11.
desired
the It
trajectory.
can be found
interference
Therefore, the genetic algorithm-optimized PID
that the running trajectory of the PID controller optimized by a genetic
is added.
controller designed in
this paper
algorithmhas is
Figure a 12
higher
showstrajectory
closer tothe the tracking
desired
torque curves accuracy
trajectory.
produced byand
Therefore, hasthe
the classic better robustness
genetic
PID when
algorithm-optimized
controller and the PID thecontroller
interference
PID
is added.
controller
optimized designed in this
by a genetic paper hasThese
algorithm. a higher
torquetrajectory
curves tracking accuracy
are measured using and has better
a torque robustness
transducer built
when
Figure the12interference
shows the
in SimMechanics. istorque
It can added.
be foundcurves
that, produced
compared with by the theclassic PIDPID
traditional controller and
controller, thethe PIDcurves
torque controller
optimized Figure
produced
by a by12 shows
the PID
genetic the torque curves
controller
algorithm. produced
optimized
These torque a by
bycurves theare
genetic classic PID controller
algorithm
measured are aand
smoother
using thethe
at
torque PID controller
beginning
transducer of in
built
optimized
operation; by a genetic
therefore, algorithm.
the new These
controller torque
has curves
better are
stability measured
at the using
beginning.
SimMechanics. It can be found that, compared with the traditional PID controller, the torque curves a torque transducer built
in SimMechanics. It can be found that, compared with the traditional PID controller, the torque curves
produced by the PID controller optimized by a genetic algorithm are smoother at the beginning of
produced by the PID controller optimized by a genetic algorithm are smoother at the beginning of
operation; therefore, the new controller has better stability at the beginning.
operation; therefore, the new controller has better stability at the beginning.

Figure 10. The errors of using the classic PID and the PID optimized by a genetic algorithm in the
presence of disturbances.

Figure
Figure 10. 10.
TheThe errors
errors of of usingthe
using theclassic
classic PID
PID and
andthe
thePID
PIDoptimized
optimizedby by
a genetic algorithm
a genetic in thein the
algorithm
presence of disturbances.
presence of disturbances.
Algorithms 2018, 11, 7 11 of 13
Algorithms 2018, 11, 7 11 of 13
Algorithms 2018, 11, 7 11 of 13

0.03
0.03

0.02 Desired
0.02 Desired
Classic
Classic algorithm
Genetic
0.01 Genetic algorithm
0.01

0
0

-0.01
-0.01

-0.02
-0.02

-0.03
-0.03
0 0.01 0.02 0.03 0.04 0.05 0.06
0 0.01 0.02 0.03
X(m) 0.04 0.05 0.06
X(m)
Figure 11. Trajectory of the end-effector using the classic PID and the PID optimized by a genetic
Figure Trajectory
11.11.
Figure ofofthe
Trajectory end-effector
the end-effectorusing
using the
the classic
classic PID and the
PID and the PID
PIDoptimized
optimizedbybya agenetic
genetic
algorithm in the presence of disturbances.
algorithm in the presence of disturbances.
algorithm in the presence of disturbances.

Figure 12. The torque curves produced by the classic PID control and the GA–PID control.
Figure
Figure 12.12.
TheThe torquecurves
torque curvesproduced
producedby
bythe
the classic
classic PID control
control and
andthe
theGA–PID
GA–PIDcontrol.
control.
Algorithms 2018, 11, 7 12 of 13

5. Conclusions
In this paper, a variable PID parameter controller optimized by a genetic algorithm controller
is proposed. The inverse kinematics model of the 3-RRR parallel robot was established according
to the vector method, and the motor conversion matrix was deduced. The gap error between the
two revolute joints in the actual system was simulated by the establishment of a three-dimensional
model; the axis distance between the two revolute joints was set to 0.5 mm. The simulation model
of the 3-RRR planar parallel robot—built by SimMechanics—was used to verify the new controller.
The results show that, compared with the traditional PID controller, the new controller designed in this
paper has better control precision, and when interference is added to the system, the new controller
has better robustness and stability, which provides the basis for practical application.

Acknowledgments: This work is supported by the National Natural Science Foundation of China (Grant
No. U1610111 and No. 51775543), and the Project Funded by the Priority Academic Program Development
of Jiangsu Higher Education Institutions (PAPD).
Author Contributions: Wei Li and Lianchao Sheng put forward the initial ideas; Lianchao Sheng conceived and
designed the simulations model; Lianchao Sheng analyzed the data; Lianchao Sheng wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Merlet, J.P. Parallel Robots; Kluwer Academic Publishers: Norwell, MA, USA, 2000.
2. Salas, F.G.; Santibáñez, V.; Llama, M.A. Fuzzy-Tuned PD Tracking Control of a 3-RRR Parallel Manipulator:
Stability Analysis and Simulations. Intell. Autom. Soft Comput. 2014, 20, 159–182. [CrossRef]
3. Fichter, E.F. A Stewart Platform-Based Manipulator: General Theory and Practical Construction. Int. J. Robot
Res. 1986, 5, 157–182. [CrossRef]
4. Lee, K.M.; Yien, C. Design and control of a prototype platform manipulator for work holding and work
handling applications. J. Mech. Work. Technol. 1989, 20, 305–314. [CrossRef]
5. Wang, R.; Zhang, X. Parameters optimization and experiment of a planar parallel 3-dof nanopositioning
system. IEEE Trans. Ind. Electron. 2017, 99, 2388–2396. [CrossRef]
6. Gutierrez, M.N.C. Dimensional synthesis of 3rrr planar parallel robots for well-conditioned workspace.
IEEE Lat. Am. Trans. 2015, 13, 409–415. [CrossRef]
7. Gosselin, C.; Angeles, J. The optimum kinematic design of a planar three-degree-of-freedom parallel
manipulator. J. Mech. Trans. Autom. Design 1988, 110, 35–41. [CrossRef]
8. Kucuk, S. Energy minimization for 3-rrr fully planar parallel manipulator using particle swarm optimization.
Mech. Mach. Theory 2013, 62, 129–149. [CrossRef]
9. Varedi-Koulaei, S.M.; Daniali, H.M.; Farajtabar, M.; Fathi, B.; Shafiee-Ashtiani, M. Reducing the undesirable
effects of joints clearance on the behavior of the planar 3-rrr parallel manipulators. Nonlinear Dyn. 2016, 86,
1–16. [CrossRef]
10. Firmani, F.; Podhorodeski, R.P. Singularity analysis of planar parallel manipulators based on forward
kinematic solutions. Mech. Mach. Theory 2009, 44, 1386–1399. [CrossRef]
11. Kroger, R.; Binder, S. Accuracy analysis of 3-dof planar parallel robots. Mech. Mach. Theory 2008, 43, 445–458.
12. Gao, M.W. Design, Analysis and Control on the Planar Parallel Robot. Master’s Thesis, South China
University of Technology, Guangzhou, China, 2012.
13. Zhang, Q.; Fan, X.; Zhang, X. Dynamic analysis of planar 3-rrr flexible parallel robots with dynamic stiffening.
Shock Vib. 2014, 1, 154–159. [CrossRef]
14. Yu, Y.Q.; Du, Z.C.; Yang, J.X.; Li, Y. An experimental study on the dynamics of a 3-rrr flexible parallel robot.
IEEE Trans. Robot. 2011, 27, 992–997. [CrossRef]
15. Moezi, S.A.; Rafeeyan, M.; Zakeri, E.; Zare, A. Simulation and experimental control of a 3-rpr parallel robot
using optimal fuzzy controller and fast on/off solenoid valves based on the pwm wave. ISA Trans. 2016, 61,
265. [CrossRef] [PubMed]
16. Shang, C.; Tao, G.; Meng, D. Adaptive robust trajectory tracking control of a parallel manipulator driven by
pneumatic cylinders. Adv. Mech. Eng. 2016, 8. [CrossRef]
Algorithms 2018, 11, 7 13 of 13

17. Abu-Dakka, F.J.; Assad, I.F.; Alkhdour, R.M.; Abderahim, M. Statistical evaluation of an evolutionary
algorithm for minimum time trajectory planning problem for industrial robots. Int. J. Adv. Manuf. Technol.
2017, 89, 389–406. [CrossRef]
18. Stan, S.D.; Gogu, G.; Manic, M.; Balan, R.; Rad, C. Fuzzy Control of a 3 Degree of Freedom Parallel Robot.
In Technological Developments in Education and Automation; Springer: Dordrecht, The Netherlands, 2010;
pp. 437–442.
19. Noshadi, A.; Mailah, M.; Zolfagharian, A. Active force control of 3-RRR planar parallel manipulator.
In Proceedings of the IEEE International Conference on Mechanical and Electrical Technology, Singapore,
10–12 September 2010; pp. 77–81.
20. Noshadi, A.; Mailah, M.; Zolfagharian, A. Intelligent active force control of a 3-rrr parallel manipulator
incorporating fuzzy resolved acceleration control. Appl. Math. Model. 2012, 36, 2370–2383. [CrossRef]
21. Pham, C.V.; Wang, Y.N. Robust adaptive trajectory tracking sliding mode control based on neural networks
for cleaning and detecting robot manipulators. J. Intell. Robot. Syst. 2015, 79, 101–114. [CrossRef]
22. Kuo, Y.L.; Lin, T.P.; Wu, C.Y. Experimental and numerical study on the semi-closed loop control of a planar
parallel robot manipulator. Math. Probl. Eng. 2014, 5, 1–9. [CrossRef]
23. Song, B.; Wang, Z.; Sheng, L. A new genetic algorithm approach to smooth path planning for mobile robots.
Assem. Autom. 2016, 36, 138–145. [CrossRef]
24. Wang, X.; Shi, Y.; Ding, D.; Gu, X. Double global optimum genetic algorithm–particle swarm
optimization-based welding robot path planning. Eng. Optim. 2016, 48, 299–316. [CrossRef]
25. Lu, X.; Liu, M. Optimal design and tuning of pid-type interval type-2 fuzzy logic controllers for delta parallel
robots. Int. J. Adv. Robot Syst. 2016, 13, 96. [CrossRef]
26. Lu, X.G.; Liu, M.; Liu, J.X. Design and optimization of interval type-2 fuzzy logic controller for delta parallel
robot trajectory control. Int. J. Fuzzy Syst. 2017, 19, 190–206. [CrossRef]
27. Yuan, L. Modern Permanent Magnet Synchronous Motor Control Theory and MATLAB Simulation; Beijing
University of Aeronautics and Astronautics Press: Beijing, China, 2016.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).