Advances in Science, Technology and Engineering Systems Journal Vol. 5, No.

3, 306-316 (2020)
ASTESJ
www.astesj.com
ISSN: 2415-6698
Special Issue on Multidisciplinary Sciences and Engineering

Multimode Control and Simulation of 6-DOF Robotic Arm in ROS

Rajesh Kannan Megalingam*, Raviteja Geesala, Ruthvik Rangaiah Chanda, Nigam Katta

Department of Electronics and Communication Engineering, Amrita Vishwa Vidyapeetham, Amritapuri, 690525, India

ARTICLE INFO ABSTRACT

Article history: The paper proposes the design and simulation of a 6 Degree of Freedom (DOF) robotic
Received: 15 January, 2020 arm, tailored for the coconut crop harvesting, assistive robots like wheelchairs and Home
Accepted: 02 May, 2020 robots, Search and rescue robots for disastrous environments, Collaborative robots for
Online: 29 May, 2020 research use. A kinematics-based solution has been developed for the robotic arm which
makes it easier to operate and use. Keyboard, GUI, Joystick are the three control interfaces
Keywords:
used in the paper. The robotic control interfaces proposed in the paper were developed
Inverse Kinematics
using the Robot Operating System (ROS). The 6- DOF articulated robotic arm was
Robot Operating System (ROS)
designed and visualized in RVIZ. The kinematics helped for the easy manipulation of the
Graphical User Interface
robotic arm with the end effector. The methodologies proposed in the research work are
6 DOF (Degree of freedom)
easy to operate and inexpensive. The designed 6 DOF robotic arm, the first three DOFs
Control Interfaces.
are for positioning of the robot’s arm, while the residual three are used for manipulation
of the gripper.

1. Introduction can be deduced from the kinematic equations. Joint angles

regulate the motion of the robotic arm. Computer-Aided Design
The robotic arm is a programmable mechanical arm that works (CAD) software is the platform to create models with the given
similar to a human arm. Robotic arms played a significant role in set of geometrical parameters. The proposed robotic arm was
the process of industrial automation. The human-like dexterity of designed in Solid Works CAD software considering all the given
these robotic arms makes them efficacious in diverse applications sets of geometrical parameters. Robot Operating System (ROS)
in a variety of industries - manufacturing, atomic power plants, provides an integrated platform for the control of robotic systems.
space exploration, material handling, painting, drilling, ROS is a special kind of framework initially developed with the
agriculture deployments [1], assistive robotics applications [2] purpose of working on robots in the research domains. In order to
and numerous other applications. The robotic arm typically understand how the ROS framework works one should be clear
comprises an end effector that is designed to manipulate and about the concept of communication of messages through topic
govern with the surroundings. The 6 DOF robot arm is designed between nodes. Simulation is one of the ways to optimize the
to manipulate and govern with the surroundings. The 6 DOF is to design and improve the control of robotic systems. ROS provides
pivot in 6 different ways that mimic a human arm. The major a 3-D visualizer (RVIZ) which helps in the visualization of the
issues concerned in an industrial robotic arm are its mechanical pose. With a properly-set URDF (Unified Robot Description
structure and the control mechanism. The control mechanisms can Format) file, one can visualize the robot model in RVIZ. RVIZ is
be effectuated by 3 options: keyboard, joystick and slider-based the simulation software used for the control of the 6-DOF robotic
control. Design of a lightweight robotic arm which can be arm.
compatible with any kind of robotic system. In the research, all The design and development of 7-DOF arms [4] and even
the proffered control mechanisms adopted inverse kinematics, successful simulations on 11-DOF [5] arm are done by other
which makes it easier to control. The proposed control authors with valid proofs of classic IK algorithms [6] in the
mechanisms are compatible with any other complex robotic discrete-time domain for robots. Taking those inputs into
systems of the same degrees of freedom. The dexterity of the consideration the 6-DOF robotic arm in the paper was designed
robotic manipulator depends on the degree of freedom. and tested with various goal positions using different modes of
Kinematic analysis [3] is one of the important steps in the design control. Algorithm developed for the keyboard based control [7]
of the robotic arm. The loop equations of complex robotic systems of the 6 DOF robotic arm using ROS which makes the control
easier. The design and simulation of this robotic arm which is
*
Rajesh Kannan Megalingam, India, +91-9496120900, [email protected]
www.astesj.com 306
https://dx.doi.org/10.25046/aj050340
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
visualized in RVIZ. Simulation of a robot is also possible on the
Gazebo platform [8] with the help of ROS_control packages and
the generation of the robot conFigureuration is in Gazebo, it uses
the URDF file and this is extensive use of a simulation platform
for a real-life scenario. Simulation of the actuators in the arm can
be done using various simulation software’s [9] like Matlab and
Simulink [10], [11] but choosing RVIZ is preferable as there are
many references to proceed. An open-source robot simulator
called USARSim can also be used for both research and education
of a robot’s general architecture [12]. An inverse kinematics
algorithm [13] needs to be developed to achieve the goal positions
in simulation over these platforms using the traditional method
the DH parameter [11], [14] the inverse kinematics solutions [1],
[15] for the arm have been derived. Few other authors used
ADAMS [16] to simulate and evaluate the arm design but this
paper deals with RVIZ simulation with multimode controls. With Figure 1: General Architectural diagram.
the help of ROS platform and MoveIt, the control [17] of the
2.1. Control Block
simulated arm was done and the three control modes [18] were
tested. The sliding mode controller mechanism can be helpful for The control block is the interface between the user and the robotic
regulation of robotic arms with unknown behaviors [19]. Laser arm. It consists of three modes of control. The user can control a
range method [20] can also be used for positioning and tracking robotic arm using any of the below modes.
tasks for a 6-DOF robotic arm. This method helps in the
determination of the distance from the camera to the target. The
• GUI
workspace [21] of the arm and the path for reaching the goal
position varies on the approach and techniques used while • Joystick
modelling the simulation. There are various other control • Keyboard
mechanisms like trajectory tracking [22], master slave control
[23], vibration control of a robotic arm with input constraints. The robotic arms is equipped with a manipulator. A manipulator
Testing in real time doesn't show accurate results due to backlash is a robotic gripper which is used to do dexterity. Figure 2
and electric issues. Considering this problem some authors represents the software architecture of the 6 DOF robotic arm in
proposed a self-learning algorithm [24] that uses positioning error the ROS platform.
after each trail. Simulation gives an idea about the performance of
the algorithm, but in case if real time implementation performance
changes due to several factors. One such thing is the connectivity
and the architecture. Few authors explicitly gave intuition into the
master-slave conFigureuration based tele existence concept
[25]. Considering all this leads, a 6 DOF robotic arm is simulated
and tested by using different control mechanisms.

Contributions of the research is as follows: Figure 2: Software Architecture diagram.

• Kinematics based solution for the 6 DOF robotic arm is The main aim is to control the robotic arm’s manipulator position
proposed. in the RVIZ. Figure. 11, describes an RVIZ axis at the base of the
• Validation of three control methodologies of the robotic robotic arm. The user can control the manipulator by using any of
arm. the control modes. These users input x, y, z coordinates will be
• Simulation and evaluation of the 6 DOF robotic arm in published to IK solver node. The IK solver node will compute the
RVIZ. joint states which are required for the manipulator to reach to the
desired coordinates.
2. Architecture
The paper is primarily concerned about the different modes of 2.1.1 Control modes and Interface nodes
control. Figure 1 represents the general architectural diagram As there are different optional modes to command the robotic arm,
which consists of control modes, a processing unit, and a 6 DOF interface node acts as an intermediate module that prioritizes the
robotic arm. ROS is used as the platform to develop software modes. The interface node will act as a bridge between the modes
programs. of control and the computing block IK_solver_node. The user can
Control in testing the 6-DOF robotic arm by simulating it with an define the order of priority for the modes in the definition file in
IK-based solution in RIVZ. The programs and files for the control the interface module. On uploading the entire program, the
modes and processing unit entirely return in CPP. compiler uses the updated definition file. On simultaneous
publishing of data from the three control modes, always the
www.astesj.com 307
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
highest priority mode values will be published to the
IK_solver_node from the interface node.

2.1.2. IK Solver node

The Inverse Kinematics equations are generated by using
Denavit-Hartenberg. An IK function is written in a computing file
where it is passed with x, y, z parameters for computing. And
there is a threshold evaluating block of code where the arms reach
limit is checked whether it could reach the x, y, z coordinates. If
the coordinate is present within the workspace of the arm it
generates the individual joint values else, it generates the joint
values which head to the maximum point it could reach in the Figure 4: representation of 3 axes joystick
direction heading to the input point. The generated joint values for The three axes of the joystick are used to control the x y z
the desired x, y, z coordinate point will be conveyed to the joint coordinates position of the manipulator of 6 DOF robotic arm. If
state publisher. Joint state publisher is the one which updates the the user moves the joystick’s axis in forward direction, x value in
joint values of the robotic arm simulation. the coordinate increases, similarly y and z can also be changed
2.1.3. Graphical User Interface (GUI) with the other two axes of the joystick. In this way, the user can
control the manipulator position.
A user-friendly GUI was designed in QT Designer software [26]
of version 4.8. Qt Designer is used to create user interface files 2.1.5. Keyboard
containing windows and controls. Figure 3 depicts the
representation of the graphical user interface which was The keyboard is another mode of control. The keys ‘a’, ‘s’, ‘d’,
developed in Qt Designer. The designed window consists of 3 fill ‘z’, ‘x’, ‘c’, ‘h’, ‘j’ are used to control the x, y, z values. When a
in blocks, a SEND COORDINATES toggle button, a HOME key from ‘a’ or ‘s’ or ‘d’ on the keyboard is pressed, the x, y, z
toggle button, and a HALT toggle button. coordinate value increases respectively. When keys from ‘z’ or ‘x’
or ‘c’ are pressed then the coordinate variables x, y, z decreases
respectively. ‘h’ and ‘j’ are the home pose and halt buttons. These
coordinates (x, y, z) are published to the interface node. And from
the interface node to the IK solver node for further computation.

3. Design and Implementation

Robotic arms play a vital role in the industrial and social
applications, where the design of the arm plays an important role
in its function. According to the functionality, there are six main
types of industrial robots: Cartesian, SCARA, cylindrical, delta,
polar and vertically articulated. The design of the robotic arm that
was developed for the research is not compromised in terms of the
workspace. Any arm with revolute joints between all its base
members falls in the category of the articulated robotic arm. An
articulated arm has the ability to rotate all the joints (number of
joints is from two to ten). The reason behind the choice for the
Figure 3: Representation of the graphical user interface.
articulated robotic arm for the research is that it has more work
The user can enter their x y z coordinates to which the user wanted envelope compared to SCARA, Delta, and Cylindrical.
the robotic arm to reach. On clicking the SEND COORDINATE
button, the values will get published to the IK solver node. For Arm base is called the waist which is vertical to the ground and
resetting the robot, a HOME toggle button is designed. Whenever the upper body of the robot base is connected to the waist through
the user presses the toggle button, the robotic arm will reach the a revolute joint which rotates along the axis of the waist and to
which the second DOF (shoulder) is attached as similar to the arm
predefined pose. A HALT toggle button is also designed which
in Figure 5. Third DOF which resembles the elbow of a human
will stop the motion of the robotic arm in case of emergency. arm. The fourth DOF gives yaw movement, fifth gives the pitch
2.1.4. 3- Axis Joystick and the last DOF gives the roll movement. The 6-DOF articulated
robotic arm was designed on Solid works and dimensions are
The Joystick is another option for controlling the robotic arm. mentioned in the Table 1. The first 3-DOFs give positional (i.e. x,
Figure 4 depicts a model of the 3-axes joystick. Arduino is the y, and z) coordinates to the end-effector and the last 3-DOFs give
microcontroller used for taking the joystick values. The joystick the orientation of the end-effector. In the Table 1 base footprint is
will be connected to Arduino analog pins externally; through the base mounting stand for the robotic arm. The arm has 6 links
analog pins the joystick values are fetched. namely base_link, Link1, Link2, Link3, Link4, Link5 and
www.astesj.com 308
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
followed by the gripper. Link1 is the link between joints J1 and
J2, Link2 is between joints J2 and J3. Similarly, for Link3, Link4,
Link5 are the links with their respective joints.

Assembly file in Solid works is converted into a URDF file using

Solid works with a maximum reach length of 1.248 m. Figure 5
and Figure 6 represents the design of the model of a 6-DOF
robotic arm in Solid works. URDF is the conventional format for
describing robots in ROS. The format is used to simulate a virtual
arm. The joint positions, link lengths, and global origin are
defined in the Solid works before exporting the assembly file of
Solid works into URDF for simulation. The designed arm is
simulated and controlled in Rviz using ROS. The evaluation of
the designed URDF is done using MoveIt setup assistant which is
ROS platform. In this section, the author first loaded the URDF
file and created the collision matrix. MoveIt divides the arm joints
and manipulators into groups, which helps to have a safer control
of the robotic arm by resolving errors due to singularity issues.
Figure 6. Top view of the Robotic arm
Table 1: Joint link lengths.
4. Robotic Arm Kinematics
i joint parent child Link Shift in
length Origin The kinematics of the robotic arm are categorized into forward
“(m)” “(m)” kinematics and Inverse kinematics. To calculate the kinematics
solution the concepts of reference frames, Euler angles and
0 Fixed Base base_link 0 0
Homogenous transforms are important.
base footprint
1 J1 base_link link1 0.15 0 4.1. Reference Frame representation
2 J2 link1 link2 0.14 0
3 J3 link2 link3 0.45 0 Reference frames play an important role for the kinematics [27]
4 J4 link3 link4 0.216 0.04 of any arm. The general concept of reference frames is explained
with the illustration in Figure 7. Consider point P, in the 2-D
5 J5 link4 link5 0.091 0
Cartesian as shown in Figure 7, is expressed with vector U relative
6 J6 link5 link6 0.106 0.03
to coordinate frame B. The same point P needs to be represented
7 gripper link6 Gripper 0.095 0
in frame A with vector V as shown in equations (1), (2) and (3).
link
The matrix representation of the point P in frame A is given in (3).
Total 1.248 Any point in frame B is multiplied with 𝑏𝑏𝑏𝑏𝑏𝑏 will project it onto
frame A.
𝑣𝑣𝑥𝑥 𝑎𝑎�𝑥𝑥 𝑏𝑏�𝑥𝑥 𝑎𝑎�𝑥𝑥 𝑏𝑏�𝑦𝑦 𝑣𝑣𝑥𝑥
�𝑣𝑣 � = � �� � (1)
𝑦𝑦 𝑎𝑎� 𝑏𝑏�
𝑦𝑦 𝑦𝑦 𝑎𝑎� 𝑏𝑏� 𝑈𝑈𝑦𝑦
𝑦𝑦 𝑦𝑦
𝑎𝑎�𝑦𝑦 = 𝑏𝑏�𝑥𝑥 sin 𝜙𝜙 = 𝑏𝑏�𝑦𝑦 cos 𝜙𝜙
𝑎𝑎�𝑥𝑥 = 𝑏𝑏� cos 𝜙𝜙 = −𝑏𝑏�𝑦𝑦 sin 𝜙𝜙 (2)
𝑣𝑣𝑥𝑥 cos 𝜙𝜙 − sin 𝜙𝜙 𝑈𝑈𝑥𝑥
�𝑣𝑣 � = � ̇ �� � (3)
𝑦𝑦 𝑠𝑠i𝑛𝑛𝑛𝑛 cos 𝜙𝜙 𝑈𝑈𝑦𝑦
𝑎𝑎 cos 𝜙𝜙 − sin 𝜙𝜙
𝑏𝑏𝑅𝑅 = � ̇ �
𝑠𝑠i𝑛𝑛𝑛𝑛 cos 𝜙𝜙

Figure 5. Front view of the Robotic arm

www.astesj.com 309
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
4.2. Euler angles respect to some other unknown frame. In Figure 9 point in frame,
B is transformed onto frame A. And the homogeneous transform
For a rigid body in space as shown in Figure 8 a box is considered is generated using rotational and translational matrices.
as a rigid body, it has 3DOF for rotation those are roll, pitch, and
yaw. These are called the Euler angles [28] and the representation Equation (7) the term 𝑇𝑇�𝑃𝑃𝐴𝐴� is the representation of point P vector
𝐴𝐴0
of this rotational DOF. Euler angles are a system to describe a with respect to frame A and this is 3*1 matrix. On the other end
sequence or a composition of rotations. Conventionally, the
in the term 𝑅𝑅𝐵𝐵𝐴𝐴 is the rotational matrix of frame B on to frame A
movements about the three axes of rotations and their associated
and this is a 3*3 matrix. 𝑇𝑇� 𝐴𝐴 𝐵𝐵0 Is the translation vector and this
angles are described by the 3D rotation matrices. Equations (4), �𝐴𝐴
0
(5) and (6) show the matrix representation of roll, pitch, and yaw. is 3*1 matrix. The 𝑇𝑇�𝑃𝑃/
𝐵𝐵
is a representation of point P vector with
𝐵𝐵0
These are 3D- rotation matrices and anticlockwise in direction.
respect to B frame.
𝑐𝑐𝑐𝑐𝑐𝑐 𝛹𝛹 − 𝑠𝑠𝑠𝑠𝑠𝑠 𝛹𝛹 0
𝑅𝑅𝑥𝑥 (𝛹𝛹) =� 𝑠𝑠𝑠𝑠𝑠𝑠 𝛹𝛹 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 0� (4) 𝑇𝑇�𝑃𝑃𝐴𝐴� 𝑅𝑅B𝐴𝐴 𝑇𝑇� 𝐴𝐴 𝐵𝐵0 𝑇𝑇�𝑃𝑃/
𝐵𝐵
� 𝐴𝐴0 �= � �𝐴𝐴
0 �� 𝐵𝐵 0 � (7)
0 0 1 1 1 1
0 0 0
𝑐𝑐𝑐𝑐𝑐𝑐 𝜙𝜙 − 𝑠𝑠𝑠𝑠𝑠𝑠 𝜙𝜙 0
𝑅𝑅𝑦𝑦 (𝜙𝜙)=� 𝑠𝑠𝑠𝑠𝑠𝑠 𝜙𝜙 (5) 4.4. Forward kinematics
𝑐𝑐𝑐𝑐𝑐𝑐 𝜙𝜙 0�
0 0 1 The forward kinematics [27] deals with location and pose of the
𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃 − 𝑠𝑠𝑠𝑠𝑠𝑠 𝜃𝜃 0 end effector with the joint variables and parameters. Here the end-
𝑅𝑅𝑧𝑧 (𝜃𝜃)=� 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑐𝑐𝑐𝑐𝑐𝑐 𝜃𝜃 0� (6) effector frame is mapped to the base joint of the arm using
0 0 1 homogenous reference frames as discussed above. Denavit–
Hartenberg parameters [14], [11] (also called DH parameters) are
the four parameters associated with a particular convention for
attaching reference frames to the links of a spatial kinematic chain,
or robot manipulator. These are calculated for all the joints using
the Figure 10 as a reference and data given in Table 2. Origin O(i)
is the intersection between x(i) and z(i) axes. The joint angle (θ(i))
is t, the angle from axis x(i-1) to x(i) about z(i) using the right-
hand rule. In joint 3 it has an offset of 180-degree constant
between x (2) and x(3) and in joint 4 it has an offset of 90-degree
constant between x(3) and x(4). Link twist (α(i-1)) is the angle
from z(i-1) to z(i) measured about x(i-1) using right hand rule. In
Table2 link length (a(i-1)) is equal to the difference between z(i-
1) and z(i) along the x(i) direction. Link offset (d(i)) in Table 2 is
the difference between x(i-1) and x(i) along the z(i) axis. Using
equation 8, which is the general representation of the transform
matrix from one frame to another. So by using the data available
in Table 2 transform matrix for link 0 and link 1 is given in 𝑇𝑇01 in
(8) Similarly, (9) is the transform for link 1 and link 2.
𝑇𝑇12 , 𝑇𝑇23 , 𝑇𝑇34 , 𝑇𝑇45 , 𝑇𝑇56 , 𝑇𝑇67 These are also similar kinds of transform
matrices with their respective links. Finally, the transform matrix
from base joint to the end effector is derived by 𝑇𝑇07 shown in
(16). The orientation difference between the definitions of the
gripper in URDF to the DH conventions so the rotation need to be
done around y axis (17) and then need to be done around z axis
(18) that is R_c shown in (19).

Table 2: DH-Parameters for the arm.

Links O(i) θ(i) α(i-1) a(i-1) d(i)

0-1 1 θ1 0 0 0.15
1-2 2 θ2 -90 0 0.14
2-3 3 θ3+180 180 0.45 0.45
3-4 4 θ4+90 -90 0.04 0
4-5 5 θ5 -90 0 0.091
4.3. Homogeneous transforms 5-6 6 θ6 90 0.03 0.106
6-7 7 θ7 0 0 0.095
Homogeneous transform is the representation of the transform
matrix of one frame which is both rotated and translated with

www.astesj.com 310
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
𝐶𝐶𝐶𝐶𝐶𝐶(−𝜋𝜋/2) 0 𝑆𝑆𝑆𝑆𝑆𝑆(𝜋𝜋/2) 0
0 1 0 0
R_y= � � (17)
−𝑆𝑆𝑆𝑆𝑆𝑆(−𝜋𝜋/2) 0 𝐶𝐶𝐶𝐶𝐶𝐶(𝜋𝜋/2) 0
0 0 0 1
𝐶𝐶𝐶𝐶𝐶𝐶(𝜋𝜋) −𝑆𝑆𝑆𝑆𝑆𝑆(𝜋𝜋) 0 0
𝑆𝑆𝑆𝑆𝑆𝑆(𝜋𝜋) 𝐶𝐶𝐶𝐶𝐶𝐶(𝜋𝜋) 0 0
R_z= � � (18)
0 0 1 0
0 0 0 1
R_c = R_y*R_z (19)
4.5. Inverse kinematics (IK)

Figure 10: Cartesian representation of joints with rotational angles for DH- Inverse kinematics analysis is for obtaining the Joint angles by
parameters. using end effector Cartesian space or position coordinates. Since
the last three joints J4, J5, J6 in Figure5 and Figure 6 are revolute
𝑐𝑐𝜃𝜃𝑖𝑖 −𝑠𝑠𝜃𝜃𝑖𝑖 0 𝑎𝑎𝑖𝑖−1 joints with joint axis intersection at J5 which would be wrist
𝑠𝑠𝜃𝜃 𝑐𝑐𝑐𝑐 𝑐𝑐𝜃𝜃𝑖𝑖 𝑐𝑐𝛼𝛼𝑖𝑖−1 −𝑠𝑠𝛼𝛼𝑖𝑖−1 −𝑠𝑠𝛼𝛼𝑖𝑖−1 𝑑𝑑𝑖𝑖 center (WC). Thus the kinematic of the IK is now evaluated by
𝑇𝑇𝑖𝑖𝑖𝑖−1 = � 𝑖𝑖 𝑖𝑖−1 � (8)
𝑠𝑠𝜃𝜃𝑖𝑖 𝑠𝑠𝛼𝛼𝑖𝑖−1 𝑐𝑐𝜃𝜃𝑖𝑖 𝑠𝑠𝛼𝛼𝑖𝑖−1 𝑐𝑐𝛼𝛼𝑖𝑖−1 𝑐𝑐𝛼𝛼𝑖𝑖−1 𝑑𝑑𝑖𝑖 calculating Inverse position and Inverse Orientation.
0 0 0 1
4.5.1. Inverse Position
c𝜃𝜃𝑖𝑖 = cos 𝜃𝜃𝑖𝑖
The inverse position problem is for obtaining the first three joint
𝑠𝑠𝜃𝜃𝑖𝑖 = sin 𝜃𝜃𝑖𝑖 angles (𝜃𝜃1, 𝜃𝜃2, 𝜃𝜃3). To evaluate Inverse Position the end effector
c𝛼𝛼𝑖𝑖−1 = cos 𝛼𝛼𝑖𝑖−1 position (Px, Py, Pz) and Orientation (Roll, Pitch, Yaw) need to
be taken from the test case input data. Thus the rotation matrix for
𝑐𝑐𝜃𝜃1 −𝑠𝑠𝜃𝜃1 0 0 the end effector R_rpy shown (20). After the error correction the
𝑠𝑠𝜃𝜃 𝑐𝑐𝜃𝜃1 0 0 actual end effector rotation matrix R_end is given in (21). In (22)
𝑇𝑇01 = � 1 � (9)
0 0 1 0 ⋅ 15 the obtained matrix 𝑅𝑅06 is the rotation part of the full
0 0 0 1 homogeneous transform matrix𝑇𝑇0𝐸𝐸𝐸𝐸 which is a transform matrix
𝑐𝑐𝜃𝜃2 −𝑠𝑠𝜃𝜃2 0 0 of end effector to the base joint. Using these concepts the joint
0 0 1 0 ⋅ 14 angles (𝜃𝜃1, 𝜃𝜃2, 𝜃𝜃3) are obtained.
𝑇𝑇12 = � � (10)
−𝑠𝑠𝜃𝜃2 −𝑐𝑐𝜃𝜃2 0 0
0 0 0 1 R_rpy = Rot (Z, yaw)*Rot(Y, pitch)*Rot(X, Roll) (20)

−𝑐𝑐𝜃𝜃3 𝑠𝑠𝜃𝜃3 0 0.45 R_end = R_rpy*R_c (21)

𝑠𝑠𝜃𝜃 𝑐𝑐𝜃𝜃3 0 0 𝑟𝑟11 𝑟𝑟12 𝑟𝑟13 𝑝𝑝𝑥𝑥
𝑇𝑇23 = � 1 � (11)
0 0 −1 −0 ⋅ 45 𝑟𝑟21 𝑟𝑟22 𝑟𝑟23 𝑝𝑝𝑦𝑦 𝑅𝑅06 𝑟𝑟 0 0/𝐸𝐸𝐸𝐸
0 0 0 1 𝑇𝑇0𝐸𝐸𝐸𝐸 = � �= � � (22)
𝑟𝑟31 𝑟𝑟32 𝑟𝑟33 𝑝𝑝𝑧𝑧 0 0 0 1
𝑠𝑠𝜃𝜃4 𝑐𝑐𝜃𝜃4 0 0 0.04 0 0 0 1
0 0 1 0 4.5.2. Inverse Orientation
𝑇𝑇34 = � � (12)
−𝑐𝑐𝜃𝜃4 −𝑠𝑠𝜃𝜃4 0 0
0 0 0 1 Inverse orientation problem is for obtaining the final three joint
angles (𝜃𝜃4, 𝜃𝜃5, 𝜃𝜃6). The resultant transform is obtained using the
𝑐𝑐𝜃𝜃5 −𝑠𝑠𝜃𝜃5 0 0
individual DH transforms. Hence the resultant rotation is
0 0 1 0.0911
𝑇𝑇45 = � � (13) calculated as shown in (23). Since the overall the RPY (roll, pitch
−𝑠𝑠𝜃𝜃5 −𝑐𝑐𝜃𝜃5 0 0
and yaw) rotation between base_link and gripper_link is equal to
0 0 0 1
the product of individual rotations between the links as shown in
𝑐𝑐𝜃𝜃6 −𝑠𝑠𝜃𝜃6 0 0.03 (24). Multiply inverse matrix of 𝑅𝑅30 (inv( 𝑅𝑅30)) on either sides in (24)
0 c𝜃𝜃6 −1 −0.106 the resultant is shown in (25). The resultant matrix (25) on RHS
𝑇𝑇56 = � � (14)
𝑐𝑐𝜃𝜃6 𝑐𝑐𝜃𝜃6 0 0 (right hand side of the equation) does not have any variables after
0 0 0 1 substituting the joint angles, and hence comparing LHS (left hand
𝑐𝑐𝜃𝜃7 −𝑠𝑠𝜃𝜃7 0 0 side of the equation) with RHS will result the equations for the
𝑐𝑐𝜃𝜃7 𝑐𝑐𝜃𝜃7 0 0 last three joint angles (𝜃𝜃4, 𝜃𝜃5, 𝜃𝜃6). Solving the obtained equations
𝑇𝑇67 = � � (15)
0 𝑐𝑐𝜃𝜃7 1 0.095 which results the joint angles.
0 0 0 1
𝑅𝑅60 = 𝑅𝑅10 ∗ 𝑅𝑅21 ∗ 𝑅𝑅32 ∗ 𝑅𝑅43 ∗ 𝑅𝑅54 ∗ 𝑅𝑅65 (23)
𝑇𝑇70 = 𝑇𝑇10 ∗ 𝑇𝑇21 ∗ 𝑇𝑇32 ∗ 𝑇𝑇43 ∗ 𝑇𝑇54 ∗ 𝑇𝑇65 ∗ 𝑇𝑇76 (16)
R_end = 𝑅𝑅60 (24)
www.astesj.com 311
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
𝑅𝑅63 = inv( 𝑅𝑅30 )*R_end (25) and fail cases obtained while conducting the survey by 20
individuals. The test results are plotted in Figure 13. Considering
5. Experiment and results these errors, joystick control was avoided for the hardware test.
The robotic arm in the research is designed for the manipulation The hardware test has been implemented using GUI as a control
tasks for various applications such as Coconut crop harvesting, interface since the simulation results of keyboard based control
Assistive robots like self-driving wheel chairs and Home robots, and GUI based control are the same. Table 6 represents the
Search and Rescue robots for disastrous environments, outcomes of IK solver when implemented on hardware using GUI
Collaborative robots for research use and etc. The evaluation of based control. It is observed that the output in the hardware testing
the design and software was first done on RVIZ simulation with using GUI based control is similar to the simulation testing using
few goal positions, these goal positions are the most desired GUI.
positions of the end effector during the general operations. Once
the software simulation has given satisfactory results, the testing
was done on hardware which was already built. The designed 6
DOF robotic arm is simulated in RVIZ. An axis plugin has been
added to visualize a coordinate axis at the base of the model
robotic arm. The blue line indicates the z-axis, the red line
indicates the x-axis and green line indicates the y-axis. The silver-
coloured arm represents the goal pose of the arm for the user input
values.
Figure 11 represents the different poses of the robotic arm
when different coordinates are input to IK solver. Different x, y,
and z coordinates are published to the interface module to
manipulate the pose of the model arm. Table 3, 5 and 6 represents
the outcomes of the IK solver when different goal points are input
into the solver using different control methods on the RVIZ
simulation platform. The output joint state values of all the goals
positions in Figure 11 are given in Table 3, 5 and 6. The input
point is considered with respect to the RVIZ [17] world axis frame.
With the IK based approach, the robotic arm was manipulated at
the desired points in the RVIZ world frame simulation. The three
control modes have been tested on 16 different goal positions. The
input data from the interface block enters the IK solver node where
the required joint values are calculated to achieve the respective
input pose. These generated joint values are passed to Joint state
publisher and robot state publisher where the required Tf and
combined joint state array will be passed to the RVIZ for the
simulation.
On comparing the accuracy between the three methods keyboard
and GUI were able to position the arm to the desired goal pose.
Joystick failed in some cases as the control mode was difficult to
manipulate the robotic arm. Since the Joystick PWM values need
to be transferred to ROS via a microcontroller, due to the
disturbances in the hardware, in a few cases joystick input readings
varied. There was 1%-3% error in the input and hence the output
was also affected due to the change in the input values. Comparing
the results from Table 3, 4, 5 it is clear that GUI and Keyboard are
two ideal methods to accurately position the robotic arm at a
desired position.
The Figure 12 and Figure 13 illustrates the error percentage values
while conducting a simulation experiment through a joystick. A
survey has been conducted by testing the system with a 20 number
of individuals. Based on the outputs that were obtained, an
analysis has been made on the error percentage in reaching the
goal position. The error is calculated by comparing the correct
outcomes obtained by an individual to the total no of test cases in
Figure 12. And an analysis has been made which showed the error Figure 11: The goal pose of the robotic arm on different user inputs.
percentage for each goal pose. This is calculated based on the pass
www.astesj.com 312
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)

Figure 12: Error percentage for individual person

Figure 13 Error Percentage for individual goal position

Table 3: Outcomes of the joint state values from the IK solver node in simulation for the user input coordinates using GUI based control.

Goal X Y Z DOF1 DOF2 DOF3 DOF4 DOF5 DOF6

Count position position position (radians) (radians) (radians) (radians) (radians) (radians)

Goal 1 0.301 0.651 0.525 1.2704 -0.9539 0.7554 -1.0660 1.3424 -1.185
Goal 2 0.402 0.730 0.376 1.0098 -1.222 0.9088 1.2786 0.476 1.4327
Goal 3 0.133 - 0.116 0.699 -0.5268 0.5868 -0.567 -0.6876 0.3067 0.0028
Goal 4 -0.392 0.532 0.825 2.2363 -0.6110 0.9545 0.043966 0.0151 0.6119
Goal 5 0.675 - 0.095 0.220 -0.0797 -1.0676 0.2312 -0.8217 0.4022 0.9090
Goal 6 - 0.019 0.242 1.020 1.4644 0.1755 0.3900 1.0424 0.5786 -1.297
Goal 7 0.204 0.644 0.529 1.3112 0.6556 0.3388 -0.9354 0.2892 1.14233
Goal 8 0.428 - 0.020 0.760 0 0 0 0 0 0

www.astesj.com 313
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
Goal 9 0.255 - 0.267 0.067 -0.89364 -1.23433 -0.70224 0.579329 0.924559 -0.41793
Goal 10 0.043 0.027 1.079 1.12537 0.424841 0.580440 -0.63930 0.2128 -0.99506
Goal11 - 0.017 - 0.110 0.759 1.396672 1.164312 -0.62972 0.222626 0.77854 0.505854
Goal 12 0.671 - 0.020 0.510 0.028259 -0.67949 0.223439 -0.42861 0.52486 0.095769
Goal 13 0.141 - 0.338 1.038 2.177904 0.639303 0.979159 1.05975 1.08808 -0.56990
Goal 14 0.281 - 0.123 0.844 2.831651 1.213924 -0.16744 0.57619 0.93674 -1.19414
Goal 15 0.303 - 0.716 0.108 -1.25223 -1.48647 0.71624 1.17593 0.725 -0.75893
Goal 16 -0.116 0.235 1.001 2.050420 0.0916879 0.3374000 0.156057999 1.03124 0.649038000

Table 4: Outcomes of the joint state values from the IK solver node in simulation for the user input coordinates using Keyboard.

Goal X Y Z DOF1 DOF2 DOF3 DOF4 DOF5 DOF6

Count position position position (radians) (radians) (radians) (radians) (radians) (radians)

Goal 1 0.301 0.651 0.525 1.2704 -0.9539 0.7554 -1.0660 1.3424 -1.185
Goal 2 0.402 0.730 0.376 1.0098 -1.222 0.9088 1.2786 0.476 1.4327
Goal 3 0.133 - 0.116 0.699 -0.5268 0.5868 -0.567 -0.6876 0.3067 0.0028
Goal 4 -0.392 0.532 0.825 2.2363 -0.6110 0.9545 0.043966 0.0151 0.6119
Goal 5 0.675 - 0.095 0.220 -0.0797 -1.0676 0.2312 -0.8217 0.4022 0.9090
Goal 6 - 0.019 0.242 1.020 1.4644 0.1755 0.3900 1.0424 0.5786 -1.297
Goal 7 0.204 0.644 0.529 1.3112 0.6556 0.3388 -0.9354 0.2892 1.14233
Goal 8 0.428 - 0.020 0.760 0 0 0 0 0 0
Goal 9 0.255 - 0.267 0.067 -0.89364 -1.23433 -0.70224 0.579329 0.924559 -0.41793
Goal 10 0.043 0.027 1.079 1.12537 0.424841 0.580440 -0.63930 0.2128 -0.99506
Goal 11 - 0.017 - 0.110 0.759 1.396672 1.164312 -0.62972 0.222626 0.77854 0.505854
Goal 12 0.671 - 0.020 0.510 0.028259 -0.67949 0.223439 -0.42861 0.52486 0.095769
Goal 13 0.141 - 0.338 1.038 2.177904 0.639303 0.979159 1.05975 1.08808 -0.56990
Goal14 0.281 - 0.123 0.844 2.831651 1.213924 -0.16744 0.57619 0.93674 -1.19414
Goal 15 0.303 - 0.716 0.108 -1.25223 -1.48647 0.71624 1.17593 0.725 -0.75893
Goal16 -0.116 0.235 1.001 2.050420 0.0916879 0.3374000 0.156057999 1.03124 0.649038000
Table 5: Outcomes of the joint state values from the IK solver node in simulation for the user input coordinates using Joystick.

Goal Count X position Y position Z position DOF1 DOF2 DOF3 DOF4 DOF5 DOF6
(radians) (radians) (radians) (radians) (radians) (radians)
Goal 1 0.305 0.652 0.525 1.2704 -0.9540 0.7556 -1.0660 1.3424 -1.185
Goal 2 0.406 0.7304 0.376 1.0098 -1.230 0.9089 1.2786 0.476 1.4327
Goal 3 0.135 - 0.116 0.699 -0.5268 0.5871 -0.567 -0.6876 0.3067 0.0028
Goal 4 -0.394 0.530 0.825 2.2363 -0.6116 0.9548 0.043966 0.0151 0.6119
Goal 5 0.67 - 0.095 0.220 -0.0797 -1.0681 0.2312 -0.8217 0.4022 0.9090
Goal 6 - 0.015 0.242 1.020 1.4644 0.1755 0.3900 1.0424 0.5786 -1.297
Goal 7 0.209 0.645 0.529 1.3112 0.6558 0.3389 -0.9354 0.2892 1.14233
Goal 8 0.421 - 0.020 0.760 0 0 0 0 0 0
Goal 9 0.255 - 0.267 0.067 -0.89364 -1.23433 -0.70224 0.579329 0.924559 -0.41793
Goal 10 0.043 0.028 1.079 1.12537 0.424841 0.58048 -0.63930 0.2128 -0.99506
Goal11 - 0.017 - 0.110 0.759 1.396672 1.164312 -0.62972 0.222626 0.77854 0.505854
Goal 12 0.671 - 0.020 0.510 0.028259 -0.67949 0.223439 -0.42861 0.52486 0.095769
Goal 13 0.141 - 0.338 1.038 2.177904 0.639303 0.979159 1.05975 1.08808 -0.56990
Goal14 0.281 - 0.123 0.844 2.831651 1.213924 -0.16744 0.57619 0.93674 -1.19414
Goal 15 0.303 - 0.716 0.108 -1.25223 -1.48647 0.71624 1.17593 0.725 -0.75893
Goal16 -0.116 0.234 1.001 2.050420 0.0916879 0.3375 0.156057999 1.03124 0.649038000

www.astesj.com 314
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
Table 6: Outcomes of the joint state values from the IK solver node on hardware testing for the user input coordinates using GUI

Goal X Y Z DOF1 DOF2 DOF3 DOF4 DOF5 DOF6

Count position position position (radians) (radians) (radians) (radians) (radians) (radians)

Goal 1 0.301 0.651 0.525 1.2799 -0.9538 0.7600 -1.0659 1.3426 -1.184
Goal 2 0.402 0.730 0.376 1.0095 -1.222 0.9086 1.2786 0.480 1.4327
Goal 3 0.133 - 0.116 0.699 -0.5268 0.5868 -0.567 -0.6876 0.3067 0.0028
Goal 4 -0.392 0.532 0.825 2.2363 -0.611015 0.95467 0.0439666 0.01561 0.6119
Goal 5 0.675 - 0.095 0.220 -0.07978 -1.0676 0.2312 -0.82172 0.4022 0.9090
Goal 6 - 0.019 0.242 1.020 1.464454 0.1755 0.3900 1.0424 0.57862 -1.29711
Goal 7 0.204 0.644 0.529 1.3112 0.6556 0.3388 -0.9354 0.2892 1.14233
Goal 8 0.428 - 0.020 0.760 0 0. 0 0 0 0
Goal 9 0.255 - 0.267 0.067 -0.89364 -1.234335 -0.702246 0.579329 0.924559 -0.41793
Goal 10 0.043 0.027 1.079 1.12537 0.424841 0.5804405 -0.63930 0.21285 -0.99506
Goal11 - 0.017 - 0.110 0.759 1.3966725 1.1643121 -0.629728 0.2226262 0.77854 0.505854
Goal 12 0.671 - 0.020 0.510 0.028259 -0.67949 0.223439 -0.42861 0.52486 0.095769
Goal 13 0.141 - 0.338 1.038 2.177904 0.639303 0.97915954 1.0597566 1.088084 -0.56990
Goal14 0.281 - 0.123 0.844 2.831651 1.2139244 -0.16744 0.576198 0.93674 -1.19414584
Goal 15 0.303 - 0.716 0.108 -1.252234 -1.48647 0.7162445 1.175932 0.72512 -0.75893
Goal 16 -0.116 0.235 1.001 2.050420 0.0916879 0.3374000 0.156057999 1.03124 0.649038000

6. Conclusion [1] Megalingam R.K, Sivanantham V, Kumar K.S, Ghanta S, Teja P.S,
Gangireddy R, Sakti Prasad K.M, Gedela V.V, “Design and development of
inverse kinematic based 6 DOF robotic arm using ROS”, International Journal
In this research work, the authors proposed and evaluated of Pure and Applied Mathematics, 118, pp. 2597-2603, ISSN: 1314-3395
reliable methods for controlling a robotic arm by testing it in both [2] Xinquan Liang, Haris Cheong, Yi Sun, Jin Guo, Chee Kong Chui, Chen-Hua
hardware and simulation using RVIZ. The authors in [17] used Yeo, “Design, Characterization, and Implementation of a Two-DOF Fabric-
the MoveIt to build a kinematics library for the IK of the robotic Based Soft Robotic Arm”, IEEE Robotics and Automation
letters(Volume:3,Issue:3,July2018),https://doi.org/10.1109/LRA.2018.28317
arm, but as an extension of that, this research designed and 23
developed a kinematics-based solution using DH parameters [3] Asghar Khan, Wang Li Quan. “Structure design and workspace calculation of
method. Using the derived equations, successfully tested the 6-DOF underwater manipulator”. 2017 14th International Bhurban
Conference on Applied Sciences and Technology (IBCAST), 10-14 Jan 2014,
designed robotic arm in Rviz using the proposed control https://doi.org/ 10.1109/IBCAST.2017.7868119
mechanisms. A survey is conducted in evaluating the best control [4] Elkin Yesid Veslin; Max Suell Dutra; Omar Lengerke; Edith Alejandra
methodology where Figure [12] and Figure [13] depict the results. Carreño; Magda Judith Morales, “A Hybrid Solution for the Inverse
Kinematic on a Seven DOF Robotic Manipulator”, IEEE Latin America
Based on the survey, the research validates that the joystick failed Transactions ( Volume: 12, Issue: 2, March 2014 ), https://doi.org/
in achieving the desired input coordinate positions due to 10.1109/TLA.2014.6749540
signaling and hardware issues. But GUI and Keyboard showed [5] Pietro Falco; Ciro Natal, “On the Stability of Closed-Loop Inverse
Kinematics Algorithms for Redundant Robots”, 05 May 2011, IEEE
better results in controlling the arm. Transactions on Robotics (Volume: 27, Issue: 4, Aug. 2011 ),https://doi.org/
As a rule, simulations do not reproduce the exact real-time 10.1109/TRO.2011.2135210
behavior of an entity or a system. PID based control can reduce [6] Seo-Wook Park, Jun-Ho Oh, “Hardware Realization of Inverse Kinematics
for Robot Manipulators”, IEEE Transactions on Industrial Electronics Vol.41,
the error. The research can be enhanced by testing the proposed No 1, February 1994, https://doi.org/ 10.1109/41.281607
design in the Gazebo simulation software. Gazebo provides the [7] Rajesh Kannan Megalingam, Nigam Katta, Raviteja Geesala, Prasant Kumar
Yadav, “Keyboard Based Control and Simulation of 6 DOF Robotic Arm
flexibility to use a PID-based controller, which helps in smooth using ROS”, 2018 4th International Conference on Computing
and exact mimicking of the real-time robotic arm as per Communication and Automation(ICCCA),
simulation in RVIZ. The proposed testing gives the developer https://doi.org/10.1109/CCAA.2018.8777568
[8] Wei Qian, Ziyang Xia, Jing Xiong, Yangzhou Gan, Yangchow Guo, Shaokui
good results. Design of the end effector can be improved for Weng, Hao Deng, Ying Hu, Jiawei Zhang, “Manipulation Task simulation
performance of multiple, divergent tasks in a real time using ROS and Gazebo”, Robotics and Biomimetics (ROBIO), 2014 IEEE
International Conference”, https://doi.org/10.1109/ROBIO.2014.7090732
environment. The singularity issues can be reduced for better [9] Žlajpah Leon, Simulation in robotics (2008) Mathematics and Computers in
performance and enhancement of the task. Simulation, 79 (4), 15 December 2008, pp. 879-897,
https://doi.org/10.1016/j.matcom.2008.02.017
Acknowledgment [10] Kichang Lee ; Jiyoung Lee ; Bungchul Woo ; Jeongwook Lee ; Young-Jin
Lee, Kimhae-si ; Syungkwon Ra “Modeling and Control of an Articulated
Robot Arm with Embedded Joint Actuators ”,2018 International Conference
We are grateful for Amrita Vishwa Vidyapeetham and on Information and Communication Technology Robotics (ICT-ROBOT),
Humanitarian and Technology Labs for providing us with all the https://doi.org/10.1109/ICT-ROBOT.2018.8549903
[11] Alla N Barakat, Khaled A. Gouda, Kenza.Bozed, “ Kinematics analysis and
sophisticated requirements to develop and complete this paper. simulation of a robotic arm using MATLAB”,2016 4th International
Conference on Control Engineering & Information Technology (CEIT),16-18
References Dec. 2016, Hammamet, Tunisia, https://doi.org/
10.1109/CEIT.2016.7929032

www.astesj.com 315
 R.K. Megalingam et al. / Advances in Science, Technology and Engineering Systems Journal Vol. 5, No. 3, 306-316 (2020)
[12] Stefano Carpin, Mike Lewis, Jijun Wang, “USARSim: a robot simulator for
research and education” Proceedings 2007 IEEE International Conference on
Robotics and Automation, 10-14 April 2007, Roma Italy, https://doi.org/
10.1109/ROBOT.2007.363180
[13] ] Jun-Di Sun; Guang-Zhong Cao; Wen-Bo Li; Yu-Xin Liang; Su-Dan Huang,
“Analytical inverse kinematic solution using the D-H method for a 6-DOF
robot”, 2017 14th International Conference on Ubiquitous Robots and
Ambient Intelligence (URAI), https://doi.org/10.1109/URAI.2017.7992807
[14] Akhilesh Kumar Mishra; Oscar Meruvia-Pastor; “ Robot arm manipulation
using depth-sensing cameras and inverse kinematics ”2014 Oceans-St
John’s,14-19 September 2014 IEEE International Conference, St. John's, NL,
Canada, https://doi.org/ 10.1109/OCEANS.2014.7003029
[15] C. A. G. Gutiérrez; J. R. Reséndiz; J. D. M. Santibáñez; G. M. Bobadilla, “A
Model and Simulation of a Five-Degree-of-Freedom Robotic Arm for
Mechatronic Courses”, IEEE Latin America Transactions ( Volume: 12,
Issue: 2, March 2014 ) ,https://doi.org/ 10.1109/TLA.2014.6749521
[16] Zenghua Bian ; Zhengmao Ye ; Weilei Mu,” Kinematic analysis and
simulation of 6-DOF industrial robot capable of picking up die-casting
products”, 2016 IEEE International Conference on Aircraft Utility Systems
(AUS), https://doi.org/ 10.1109/AUS.2016.7748017
[17] S. Hernandez-Mendez, C. Maldonado-Mendez, A. Marin-Hernandez, H. V.
Rios-Figureueroa, H. Vazquez-Leal and E. R. Palacios-Hernandez, “Design
and implementation of a robotic arm using ROS and MoveIt!”, 2017 IEEE
International Autumn Meeting on Power, Electronics and Computing
(ROPEC), Ixtapa, 2017, pp. 1-6, http://doi.org/
10.1109/ROPEC.2017.8261666
[18] Weimin Shen, Jason Gu, Yide Ma, “3D Kinematic Simulation for PA10-7C
Robot Arm Based on VRML”, 18-21 Aug. 2007, Jinan, China.
https://doi.org/10.1109/ICAL.2007.4338637
[19] José de Jesús Rubio, “Sliding mode control of robotic arms with dead zone”,
IET Control Theory & Applications (Volume: 11, Issue: 8, 5 12 2017),
https://doi.org/ 10.1049/iet-cta.2016.0306
[20] Megalingam R.K, Rajesh Gangireddy, Gone Sriteja, Ashwin Kashyap,
Apuroop Sai Ganesh “Adding intelligence to the robotic coconut tree climber”
2017 International Conference on Inventive Computing and Informatics
(ICICI), 23-24 Nov. 2017, Coimbatore India, https://doi.org/
10.1109/ICICI.2017.8365206
[21] Teerawat Thepmanee, Jettiya Sripituk, Prapart Ukakimapurn, “A simple
technique to modeling and simulation four-axe robot-arm control” 17-20 Oct.
2007, Coex, Seoul, Korea, https://doi.org/10.1109/ICCAS.2007.4406694
[22] Tingting Meng; Wei He, “Iterative Learning Control of a Robotic Arm
Experiment Platform with Input Constraint”, IEEE Transactions on Industrial
Electronics (Volume: 65, Issue: 1, Jan. 2018), https://doi.org/
10.1109/TIE.2017.2719598
[23] Gourab Sen Gupta, Subhas Chandra Mukhopadhyay, Christopher H.
Messom, Serge N. Demidenko, “Master-Slave Control of a Teleoperated
Anthropomorphic Robotic Arm With Gripping Force Sensing ”, IEEE
Transactions on Instrumentation and Measurement ( Volume: 55, Issue: 6,
Dec. 2006 ),https://doi.org/ 10.1109/TIM.2006.884393
[24] P. Lucibello,“ Repositioning control of robotic arms by learning ”, IEEE
Transactions on Automatic Control ( Volume: 39, Issue: 8, Aug 1994 ),
https://doi.org/ 10.1109/9.310053
[25] Qingyuan Sun; Lingcheng Kong; Zhihua Zhang; Tao Mei, “Design of
wireless sensor network node monitoring interface based on Qt”, 2010
International Conference on Future Information Technology and
Management Engineering, Changzhou, 2010, pp. 127-130
[26] R. Takuma; Y. Asahara ; H. Kajimoto ; N. Kawakami ; S. Tachi,
“Development of anthropomorphic multi-D.O.F master-slave arm for mutual
telexistence ”, IEEE Transactions on Visualization and Computer Graphics (
Volume: 11 , Issue: 6 , Nov.-Dec. 2005 ), https://doi.org/
10.1109/TVCG.2005.99
[27] Ramesh, S. B. Hussain and F. Kangal, "Design of a 3 DOF robotic arm," 2016
Sixth International Conference on Innovative Computing Technology
(INTECH), Dublin, 2016, pp. 145-149, https://doi.org/
10.1109/INTECH.2016.7845007
[28] Jose de Jesus Rubio; Adrian Gustavo Bravo; Jaime Pacheco; Carlos
Aguilar,“Passivity analysis and modeling of robotic arms”, IEEE Latin
America Transactions ( Volume: 12 , Issue: 8 , Dec. 2014 ) ,https://doi.org/
10.1109/TLA.2014.7014505

www.astesj.com 316