2018 4th International Conference on Computing Communication and Automation (ICCCA)

Keyboard-Based Control and Simulation of 6-DOF

Robotic Arm Using ROS
Rajesh Kannan Megalingam, Nigam Katta, Raviteja Geesala, Prasant Kumar Yadav, Ruthvik Chanda Rangaiah
Dept. of Electronics and communication Engineering,
Amrita Vishwa Vidyapeetham,
Amritapuri, India
[email protected], [email protected], [email protected], [email protected], [email protected]

Abstract— In this paper, we are proposing the design and for the control of the 6DOF robotic arm using the keyboard is
simulation of a 6 DOF robotic arm to be used for the search and RVIZ.
rescue operations in the disaster-hit areas. An algorithm
developed for the keyboard based interaction with the 6 DOF II. RELATED WORKS
robotic arm which makes it more user-friendly is also discussed.
The 6- DOF articulated robotic arm was designed and visualized Paper [1], discusses about simulating a robot in the Gazebo
in RVIZ and, Moveit is being used as a control interface using with the help of ROS_control packages and how the robot
Robot Operating System (ROS). The manipulation of the arm configuration is generated in Gazebo using the URDF file. The
with end effector refrain us from using the joint by the joint paper describes the extensive use of simulation platform for a
control mechanism. The technique presented in this research real-life scenario. In paper [2] the design and inverse
work is inexpensive and easier to control. In the designed 6 DOF kinematics of a 3 DOF robotic arm is described. It discusses
robotic arm, the first three DOF is for the position of the arm and the practicality of the arm design in work and the calculations
the rest three are used for the manipulation of the gripper. of the joint angles using kinematics while performing some
tasks. Paper [3] describes how the robot’s motion is controlled
Keywords—Inverse Kinematics, URDF, Forward Kinematics,
Transfer Matrix, Moveit, 6-DOF, ROS.
using inverse kinematics and the use of depth camera to
capture the user's input. It also explains how to control the end
I. INTRODUCTION effector using some speech and gesture commands. Paper [4]
Basically a robotic arm is a mechanical system which reports the mobility of a 5 DOF articulated arm that has been
works similar to a human arm. It consists of an end effector tested using direct and indirect kinematic model. The robot is
usually designed to manipulate and interact with the simulated in MATLAB - Simulink and the trajectories and
surroundings. They are mostly used in industrial and service working domain is determined. Based on the known gripper
purposes like pick and place, exploration, manufacturing, position and orientation the desired rotation is calculated.
laboratory research, space exploration etc. The 6 Degrees of Paper[5] presents the concept of an in-parallel actuated
freedom is to pivot in 6 different ways which resembles a mechanism and also derives the basic kinematic equations to
human arm. The major issues concerned in an industrial manipulate the 6 DOF robotic arm. They also used Jacobian
robotic arm are its mechanical structure and the control matrix for simulation. Paper [6] explains about the most
mechanism. So we designed a lightweight robotic arm which commonly used simulations in various robotic fields. The
makes it easy to mount and carry on any robotic system. The mathematical and analytical approach to solve the problems of
keyboard based control proposed in this research work which forward and inverse kinematics to control the end effector is
makes the designed arm easier to control compared to the discussed in [7] and [14]. Underwater workspace and
other control mechanisms like the joystick control, slider kinematic analysis of a robotic arm is studied in paper [8]. In
based control etc. This keyboard based interface is compatible the paper [9] the equations of forward kinematics are derived
with any other robotic arm provided the dimensions and using D-H notation. The simulation is done in MATLAB by
URDF of the robotic arm. creating 3D graphics.. The simulation helped them to detect
Robot Operating System (ROS) provides an integrated the motion of the robotic arm. Paper [10] uses a simplified
platform to control robotic systems. ROS is a special kind of simulation to control the movement of a 4 axis arm. The
framework initially developed with the purpose of working on kinematics and inverse kinematics of the robotic arm based on
robots in the research fields. In order to understand how the the D-H parameters are calculated in paper [11]. Paper [12]
ROS framework works one should be clear about the concept and [13] gives a detailed description about multi-robot
of communication of messages through topic between nodes. simulator and the library for grasp synthesis. Paper [15] details
Simulation is one of the ways to optimize the design and us on tracking and position object using vision systems
improve the control of the robotic systems.ROS provides a 3- associated with the robotic arm using laser range method.
D visualizer (RVIZ) which helps us to visualize the pose or III. ROBOTIC ARM KINEMATICS AND DYNAMICS
estimates of the robot. With properly-set URDF file, one can
visualize robot model in RVIZ. The Simulation software used Kinematics deals with the spatial displacement of the robot
as a function of time. In general, we face two main problems

978-1-5386-6947-1/18/$31.00 ©2018 IEEE 1

in controlling the robotic arm: first is forward/ direct
kinematics and the second is inverse kinematics. We majorly
control the robotic arm using inverse kinematics because it
makes our job simpler and we also have IK solvers to achieve
our goal. Denavit and Hartenberg (DH parameters) [1955]
proposed a generalized and simpler approach of using the
matrix algebra to describe the spatial geometry of the links of
a robot arm with respect to a fixed reference frame. This
method deals with homogeneous (4x4) matrices. In any
robotic system, to understand its dynamics we must be aware
of transformation matrices. Even inverse kinematics deals
with transform matrices. These kinematics and inverse
kinematics equations help us to understand the motion of the
joints. We calculate the position and orientation of joints with Fig. 1. Frame A and B
respect to a reference joint. The general equations of transform
matrices in the 3D-coordinate system are explained here. As Steps to move the end-effector from the actual position to the
shown in Fig. 1 consider two frames ‘A’ and ‘B’ and let P be a new Cartesian goal position:
1. We know our actual position (x, y, z - coordinates and
given point in frame B which is represented by the matrix orientation) with forward kinematics.
in equation (1). In order to understand the position and 2. We define our new goal position (x, y, z - coordinates and
orientation of point ‘P’ with respect to frame ‘A’, we apply orientation).
3D- transformation by multiplying the matrix with the 3. Solve the inverse kinematics to get the angles for the new
transformation matrix in equation (2). In the goal position.
transformation matrix, we deal with both rotational and 4. Send angles to the motors/motor controllers to define speed
translational matrices. Rotational matrix is a 3x3 matrix and and acceleration and implement a control loop for that.
must be orthonormal whereas translational matrix is 3x1. The 5. Robot arm should move to a new position.
rotational matrices in describe the orientation of the frame.
IV. DESIGN AND IMPLEMENTATION
A 6 DOF articulated robotic arm is designed in Solidworks
with dimensions mentioned in Table 1. It is converted it into
URDF using the same software with maximum reach length of
120cm. Fig. 2 represents the design of the model 6 DOF
(1) robotic arm in Solidworks. URDF (Unified Robot Description
Format) is the basic/conventional format for describing robots
in ROS. We have used this format to build our virtual arm.
The joint positions, link lengths and global origin are defined
according to our need. This is further used in RVIZ to
(2) visualize the motion of the arm accordingly. Any robot with
revolute joints between all its base members falls in the
category of the articulated robot. In an articulated arm
numbers of joints are from two to ten. The axes of the revolute
joints can be parallel or orthogonal to each other with some
pairs of joints parallel and others orthogonal to each other.
Arm base called as waist which is vertical to the ground and
the upper body of the robot base is connected to the waist
through a revolute joint which rotates along the axis of the
waist and to which we have our second DOF (shoulder). Third
DOF which resembles the elbow of a human arm. The fourth
DOF gives yaw movement, fifth gives the pitch and the last
DOF gives the roll movement. The reason behind our choice
for this kind of arm is it has more work envelope. This kind of
design can be used in industrial and research purpose as well.
The first 3 DOF’s give positional (i.e x, y, z) coordinates to
the end-effector and last 3 DOF’s gives the orientation of the
(3)
end-effector.

2
 The User interface block has two different interfaces
associated in controlling of 6 DOF robotic arm. Through slider
based GUI user can simulate the model robotic arm using
forward kinematics. Fig. 4 represents the slider based GUI.
This GUI consists of 6 sliders which represents the current
joint states of each DOF. Every joint can be controlled
manually by changing the joint values in GUI. The inputted
joint values are published to Move group. Move group
monitors the current transform information and joint states to
know the current pose of the model arm. With reference to the
current pose, it produces the corresponding joint trajectory
path to the required goal pose. Once the path is produced,
RVIZ will simulate the model arm.

Fig. 2. Design of the 6 DOF robotic arm in Solidworks.

Using the above-designed URDF we configured MoveIt

required files using MoveIt setup assistant. In this section, we
first loaded the URDF file and created the collision matrix.
Then the virtual joints are defined which are used to attach the
arm base to the environment. Moveit divides the arm joints
and manipulator into groups. This helps to have a safer control
of the robotic arm by resolving errors due to singularity issues.
TABLE I. LINKS LENGTHS OF THE ARM.
Fig. 4. Slider-Based GUI
S.NO JOINTS LENGTHS
(mm) Keyboard is another interface through which the user can
1. WAIST (1st DOF) 100
simulate the model robotic arm. In this interface, the end
effector of the model robotic arm is manipulated by using
2. SHOULDER (2nd DOF) 400 inverse kinematics plugin in Moveit. This plugin calculates the
required joint values of the corresponding DOF when a
3. ELBOW (3rd DOF) 100 coordinate goal position is given to the end effector. Through
a custom python node, the user can input the desired x, y, z
4. YAW ORIENTATION (4th DOF) 350 coordinates to manipulate the model arm manipulator. When
the keys a, s, and d on the keyboard are pressed, accordingly
5. PITCH ORIENTATION (5th DOF) 100
the coordinate variables x, y, z increases respectively. When
6. ROLL ORIENTATION (6th DOF) 150 the keys z, x, and c are pressed the coordinate variables x, y, z
decreases respectively. These x, y, z variable values are
V. ARCHITECTURE published to the move group through Move commander. Move
commander acts as a bridge between keyboard and the Move
The authors described the detailed description of how to group. Move group subscribes the x, y, z variable values. IK
use Robotic operating system (ROS) as a platform to control plugin in Moveit returns the corresponding joint values for the
and visualize a robotic arm in RVIZ. Fig. 3 represents the given x, y, z values. Move group gets the transforms and
detailed architectural diagram of the Robotic operating system. current joint states through the robot state publisher and joint
state publisher. With reference to current pose, the Open
Motion Planning library (O+MPL) plans the collision-free
path for the manipulator to reach the required goal pose. The
produced plan is executed by the Move group. RVIZ
visualizes the entire execution of the model robotic arm.
VI. EXPERIMENT AND RESULTS
The Moveit configured files were used to launch the demo
version of our URDF in RVIZ. We added Axes plugin to
visualize a coordinate axis at the base of the model robotic
arm. Fig. 5 represents the initial pose and a coordinate axis at
the base of the model arm in RVIZ. Through Moveit
Fig. 3. Architectural diagram. commander node we built a connection between keyboard and

3
move group. We published different x, y, and z coordinate
values to the Move group to manipulate the pose of the model
arm. In figures 6, 7, 8 and 9 the transparent orange colored
arm represents the previous pose of the model arm. The silver
colored arm represents the goal pose of the arm for the user
inputted values. The blue line indicates the z-axis, the red line
indicates the x-axis and green line indicates the y-axis.
The algorithm is tested by simulating robotic arm for
different user inputs. Figures 6, 7, 8 and 9 are some of the
sample results of the robotic arm visualization of different
coordinates given by the user.

Fig. 8. Model arm at goal pose X = 0.29668090586, Y= 0.29668090586, Z =

0.451957497039.

Fig. 5. Simulation of a model arm in RVIZ.

Fig. 9. Model arm at goal pose X = 0.48730810225, Y = -0.0859955474569, Z

= 0.251662956214.

VII. FUTURE WORKS

The 6 DOF robotic arm can be simulated in Gazebo using the
ROS package controller and explore a more realistic situation
to deploy the functionality of the controls. The end effector
design can be improved for performing multiple divergent
tasks in disaster-hit areas. The singularity issues can be
Fig. 6. Model arm at goal pose X = 0.347021457694, Y =0.201272445463, Z reduced for better performance and enhancement of the task.
= 0.478438453617.
VIII. CONCLUSION
This paper proposed a method to control the 6 DOF robotic
arm using simulation software i.e RVIZ using keyboard and
sliders based GUI in ROS. The simulation helped us to know
the robot parameters like the orientation and position of the
robot. The main control used the concepts and algorithms of
forward kinematics and inverse kinematics which helped us to
control it in an easier way. The design of the 6 DOF arm is
done using Solidworks CAD software and simulation is done
in the RVIZ. The testing of the control system is done in RVIZ
and successfully obtained the results which are required.

Fig. 7. Model arm at goal pose X = 0.374935093033, Y = 0.315734815186, Z

= 0.0813328702545.

4
 ACKNOWLEDGMENT [8] Asghar Khan, Wang Li Quan, “Structure design and workspace
calculation of 6-DOF underwater manipulator,” 2017 14th International
We thank Amrita University and Humanitarian Technology Bhurban Conference on Applied Sciences and Technology (IBCAST),
Lab of ECE Dept for providing us a wonderful opportunity to 10-14 January 2014.
test and complete our project successfully. [9] Alla N Barakat, Khaled A. Gouda, Kenza.Bozed, “Kinematics analysis
and simulation of a robotic arm using MATLAB,” 2016 4th
International Conference on Control Engineering & Information
REFERENCES Technology (CEIT), 16-18 December 2016, Hammamet, Tunisia.
[1] Wei Qian, Zeyang Xia, Jing Xiong, Yangzhou Gan, Yangchao Guo, [10] Teerawat Thepmanee, Jettiya Sripituk, Prapart Ukakimapurn, “A simple
Shaokui Weng, Hao Deng, Ying Hu, Jianwei Zhang, “Manipulation technique to modeling and simulation four-axe robot-arm control,” 17-
Task Simulation using ROS and Gazebo,” Robotics and Biomimetics 20 October 2007, Coex, Seoul, Korea.
(ROBIO), 2014 IEEE International Conference, pp. 2594-2598, 5 [11] Weimin Shen, Jason Gu, Yide Ma, “3D Kinematic Simulation for PA10-
December 2014. 7C Robot Arm Based on VRML, 18-21 August 2007, Jinan, China.
[2] Ramish, S. B. Hussain and F. Kanwal, “Design of a 3 DoF robotic arm,” [12] Stefano Carpin, Mike Lewis, Jijun Wang, “USARSim: a robot simulator
2016 Sixth International Conference on Innovative Computing for research and education” Proceedings 2007 IEEE International
Technology (INTECH), Dublin, 2016, pp. 145-149. Conference on Robotics and Automation, 10-14 April 2007, Roma Italy.
[3] Akhilesh Kumar Mishra, Oscar Meruvia-Pastor, “ Robot arm [13] Jongwoo Park, Chan-Hun Park, Dong-li Park, “The library for grasp
manipulation using depth-sensing cameras and inverse kinematics,” synthesis and robot simulation,” 2017 14th International Conference on
2014 Oceans-St John’s, 14-19 September, 2014 IEEE International Ubiquitous Robots and Ambient Intelligence (URAI), 28 June - 1 July
Conference, St. John's, NL, Canada. 2017, Jeju South Korea.
[4] Leba, Monica & Pop, Emil, “Articulated Robotic Arm Simulation And [14] Megalingam R.K, Sivanantham V, Kumar K.S, Ghanta S, Teja P.S,
Control”, 2018. Gangireddy R, Sakthiprasand K.M, Gedela V.V, “Design and
[5] Annamareddy Srikanth, Y. Ravitej, V.Sivaraviteja, V.Sreechand. development of inverse kinematic based 6 dof robotic arm using ROS,”
“Kinematic Analysis and Simulation of 6 D.O.F. Of Robot for Industrial International Journal of Pure and Applied Mathematics, 118, pp. 2597-
Application”, International Journal Of Engineering And Science Vol.3, 2603.
Issue 8 (September 2013), PP 01-04. [15] Megalingam R.K, Rajesh Gangireddy, Gone Sriteja, Ashwin Kashyap,
[6] Žlajpah Leon, “Simulation in robotics,” Mathematics and Computers in Apuroop Sai Ganesh, “Adding intelligence to the robotic coconut tree
Simulation, 79 (4), 15 December 2008, pp. 879-897. climber,” 2017 International Conference on Inventive Computing and
[7] Alireza Khatamian, “Solving Kinematics Problems of a 6-DOF Robot Informatics (ICICI), 23-24 November 2017, Coimbatore India.
Manipulator,” Int'l Conf. Scientific Computing CSC'15.