Modern Mechanical Engineering, 2011, 1, 47-55
doi:10.4236/mme.2011.12007 Published Online November 2011 (http://www.SciRP.org/journal/mme)
Design and Development of a Competitive Low-Cost Robot
Arm with Four Degrees of Freedom
Ashraf Elfasakhany1,2, Eduardo Yanez2, Karen Baylon2, Ricardo Salgado2
1Department of Mechanical Engineering, Faculty of Engineering, Taif University, Al-Haweiah, Saudi Arabia
2Tecnológico de Monterrey, Campus Ciudad Juárez, Ciudad Juarez, Mexico
Received October 19, 2011; revised November 7, 2011; accepted November 15, 2011
Abstract
The main focus of this work was to design, develop and implementation of competitively robot arm with en-
hanced control and stumpy cost. The robot arm was designed with four degrees of freedom and talented to
accomplish accurately simple tasks, such as light material handling, which will be integrated into a mobile
platform that serves as an assistant for industrial workforce. The robot arm is equipped with several servo
motors which do links between arms and perform arm movements. The servo motors include encoder so that
no controller was implemented. To control the robot we used Labview, which performs inverse kinematic
calculations and communicates the proper angles serially to a microcontroller that drives the servo motors
with the capability of modifying position, speed and acceleration. Testing and validation of the robot arm
was carried out and results shows that it work properly.
Keywords: Robot Arm, Low-Cost, Design, Validation, Four Degrees of Freedom, Servo Motors, Arduino
Robot Control, Labview Robot Control
1. Introduction
The term robotics is practically defined as the study,
design and use of robot systems for manufacturing [1].
Robots are generally used to perform unsafe, hazardous,
highly repetitive, and unpleasant tasks. They have many
different functions such as material handling, assembly,
arc welding, resistance welding, machine tool load and
unload functions, painting, spraying, etc.
There are mainly two different kinds of robots: a ser-
vice robot and an industrial robotic. Service robot is a ro-
bot that operates semi or fully autonomously to perform
services useful to the well-being of humans and equipment,
excluding manufacturing operations [2]. Industrial robot,
on the other hand, is officially defined by ISO as an auto-
matically controlled and multipurpose manipulator pro-
grammable in three or more axis [1]. Industrial robots are
designed to move material, parts, tools, or specialized de-
vices through variable programmed motions to perform a
variety of tasks. An industrial robot system includes not
only industrial robots but also any devices and/or sensors
required for the robot to perform its tasks as well as se-
quencing or monitoring communication interfaces.
In 2007 the world market grew by 3% with approxi-
mately 114,000 new installed industrial robots. At the
end of 2007 there were around one million industrial ro-
bots in use, compared with an estimated 50,000 service
robots for industrial use [3].
Due to increase using of industrial robot arms, an evo-
lution to that topic began trying to imitate human move-
ments in a detail mode. For example a group of students
in Korea made a design of innovations that robotic arm
take account of dancing hand, weight lifting, Chinese cal-
ligraphy writing and color classification [4]. Another group
of engineers at USA develop eight degrees of freedom
robot arm. This robot is able to grasp many objects with
a lot of shapes from a pen to a ball and simulating also
the hand of human being [5]. In space, the Space Shuttle
Remote Manipulator System, known as SSRMS or Cana-
darm, and its successor is example of multi degree of
freedom robot arms that have been used to perform a va-
riety of tasks such as inspections of the space shuttle
using a specially deployed boom with cameras and sen-
sors attached at the end effector and satellite deployment
and retrieval manoeuvres from the cargo bay of the space
shuttle [6].
In Mexico, Scientists are on track to design and de-
velop many robot arms, and the Mexican government

A. ELFASAKHANY ET AL.
48
estimates that in Mexico there are about 11,000 robotic
arms used in different industrial applications. However,
the experts think that the apogee of the robot arms is not
only of higher quality, but also accurately, repeatability,
and stumpy cost.
Most robots are set up for an operation by the teach-
and-repeat technique. In this mode, a trained operator (pro-
grammer) typically uses a portable control device (a teach
pendant) to teach a robot its task manually. Robot speeds
during these programming sessions are slow.
The present work is part of a two-phase project, which
requires a mobile robot to be able to transport the tools
from the storage room to the industrial cell. In this phase
in the project, which carried out at Monterrey University
of Technology, Mexico, the main focus was to design,
development and implementation of an industrial robotic
arm with stumpy cost, accurate and superior control. This
robot arm was designed with four degrees of freedom and
talented to accomplish simple tasks, such as light mate-
rial handling, which will be integrated into a mobile plat-
form that serves as an assistant for industrial workforce.
2. Mechanical Design
The mechanical design of the robot arm is based on a
robot manipulator with similar functions to a human arm
[6-8]. The links of such a manipulator are connected by
joints allowing rotational motion and the links of the ma-
nipulator is considered to form a kinematic chain. The
business end of the kinematic chain of the manipulator is
called the end effector or end-of-arm-tooling and it is
analogous to the human hand. Figure 1 shows the Free
Body Diagram for mechanical design of the robotic arm.
As shown, the end effector is not included in the design
because a commercially available gripper is used. This is
because that the end effector is one of the most complex
Figure 1. Free body diagram of the robot arm.
parts of the system and, in turn, it is much easier and
economical to use a commercial one than build it.
Figure 2 shows the work region of the robotic arm.
This is the typical workspace of a robot arm with four
degree of freedom (4 DOF). The mechanical design was
limited to 4 DOF mainly because that such a design al-
lows most of the necessary movements and keeps the
costs and the complexity of the robot competitively. Ac-
cordingly, rotational motion of the joints is restricted where
rotation is done around two axis in the shoulder and
around only one in the elbow and the wrist, see Figure 1.
The robot arm joints are typically actuated by electri-
cal motors. The servo motors were chosen, since they in-
clude encoders which automatically provide feedback to
the motors and adjust the position accordingly. However,
the disadvantage of these motors is that rotation range is
less than 180˚ span, which greatly decreases the region
reached by the arm and the possible positions [9]. The
qualifications of servo motors were selected based on the
maximum torque required by the structure and possible
loads. In the current study, the material used for the struc-
ture was acrylic.
Figure 3 shows the force diagram used for load calcu-
lations. The calculations were carried out only for the joints
that have the largest loads, since the other joints would
have the same motor, i.e. the motor can move the links
without problems. The calculations considered the weight
of the motors, about 50 grams, except for the weight of
motor at joint B, since it is carried out by link BA. Fig-
ure 4 shows the force diagram on link CB, which con-
tains the joints (B and C) with the highest load (carry the
links DC and ED) and the calculations are carried out as
follows.
Figure 2. Work region of the robotic arm.

49 A. ELFASAKHANY ET AL.
Figure 3. Force diagram of robot arm.
Figure 4. Force diagram of link CB.
The values used for the torque calculations:
Wd = 0.011 kg (weight of link DE)
Wc = 0.030 kg (weight of link CD)
Wb = 0.030 kg (weight of link CB)
L = 1 kg (load)
Cm = Dm = 0.050 kg (weight of motor)
LBC = 0.14 m (length of link BC)
LCD = 0.14 m (length of link CD)
LDE = 0.05 m (length of link DE)
Performing the sum of forces in the Y axis, using the
loads as shown in Figure 4, and solving for CY and CB,
see Equations (1)-(4). Similarly, performing the sum of
moments around point C, Equation (5), and point B, Equa-
tion (6), to obtain the torque in C and B, Equations (7)
and (8), respectively.

g
ydmcmY
FLWDWC C  
0 (1)

2
1.141kg9.8ms11.18N
Y
C (2)

0
ydmcmBB
FLWDWCWgC 
 (3)

2
1.171 kg9.8ms11.4758N
B
C (4)

22
0
cCD DE
cDCD
CDDEm CDc
WL L
MWL
LLLD LM
 
 
 
 
 
 (5)



2
2
0
2
DE
BBCCDDEDBCCD
CD
mBCCDc BC
BC
mBCBB
L
MLLLL WLL
L
DL LWL
L
CL WM

  







 



(6)
1.968Nm278.6 ozin
c
M

 (7)
3.554Nm503.38 ozin
B
M

 (8)
The servo motor that was selected, based on the cal-
culations, is the Hextronik HX12K, which has a torque of
280 oz/in. This motor was recommended because it is
much cheaper than any other motor with same specifica-
tions. Since we need more torque at joint B, see Equation
(8), we used two motors at point B to comply with the
torque requirements; however, one motor is enough for
the other joints. Using two motors at joint B is much
cheaper than using one big motor with 560 oz/in. Other
relevant characteristics of the motors, which can be shown
in Figure 5, are that they can turn 60 degrees in 130 mil-
liseconds and they have a weight of 47.9 grams each.
Once the initial dimensions for the robot arm and the
motor were defined, the design were carried out using
the SolidWorks platform; design should carefully take
into account the thickness of the acrylic sheet and the
way that the pieces would be attached to each other. The
acrylic sheet used to make the robot is 1/8 thickness and
Figure 5. Servo motor.

A. ELFASAKHANY ET AL.
50
that thin sheet was chosen because it easier for machining
and less weight with a good resistance.
During design, we faced some difficulties due to the way
of joining thin acrylic parts strongly. It was needed tools
to burn and join the acrylic parts and that weren’t avail-
able and the team considered that a mechanical junction
based on screws and nuts would be much strong than other
alternatives, such as glue for example. In order to accom-
plish this, a small feature was designed which allowed to
fasten the bolts with the nuts without having to screw in
the thin acrylic layer. The result of this process was the
tridimensional design shown in Fi gu re 6.
By end of design, each part was printed in full scale in
cardboard paper and then we verified all the dimensions
and the interfaces of the assembly. In turn, we built the
first prototype of the robot arm. Next, parts of the robot
arm were machined from the acrylic sheet using a circu-
lar saw and Dermal tools. The detailing on the parts was
done in a professional workshop since the parts of robot
arm were too small and it is not an easy for accomplish-
ing such small and accurate cuts.
During assembling the robot parts with the motors, few
problems pop up. There were critical points that did not
resist the fastening and, in turn, may break down; hence,
reinforcements in these points were considered. The final
result of the robot arm is shown in Figure 7.
3. Robot Arm Inverse Kinematics
To validate the right positioning of the robotic arm, in-
verse kinematics calculations are carried out. Such cal-
culations are used to obtain the angle of each motor from
Figure 6. Robot arm 3D model.
Figure 7. Robot arm complete assembly.
a position given by using the Cartesian coordinate sys-
tem, as shown in Figure 8. Each motor will have a spe-
cific function: the motor located in the A union positions
the final element in the y axis, the motors B and C posi-
tions the final element in the x and z axis.
The problem was simplified by using the xz plane, as
shown in Figure 9. In which the following known values
were defined [9]:
LAB: the forearm length.
LBC: the arm length.
z: the position in the z axis.
x: the position in the x axis.
y: the position in the y axis.
Using trigonometry relations, as shown in Figure 9,
the motor angles θ2 and θ1 are obtained, as seen in Equa-
tions (9) and (10).
222
2180 arcCos2
LAB LBCx z
LAB LBC



 



2
(9)
222
122
arcTan arcCos
2
zLABLBCx
xLABx z







 

2
z
(10)
0arcTan y
x




(11)
The motor B is going to use θ1 and the motor C is go-
ing to use θ2. The angle for the motor A is calculated as

51 A. ELFASAKHANY ET AL.
Figure 8. Coordinate system.
Figure 9. xz Plane.
seen in Equation (11). With these calculations, the angles
of servomotors are obtained and in turn they take the ac-
tion to move the whole structure to the specific position.
4. End-Effector Selection
The end effector is probably one of the most important and
most complex parts of the system. Wisely, it is much ea-
sier and economical to use a commercial one than build
it. The end effector varies mainly according to the appli-
cation and the task that the robot arm accomplishes for; it
can be pneumatic, electric or hydraulic. Since our robot arm
is based in an electric system, we may choose electric ba-
sis of end effector. Besides, the main application of our
system is handling, accordingly, the recommended type
of our end effector is a gripper, as shown in Figure 10.
Please note that the end effector is controlled by a servo
motor and, in turn, the total servo motors used for our
robot arm will be 5 motors that move the structure.
Figure 10. Grip per with ser v o .
5. Robot Arm Control
The robot arms can be autonomous or controlled manually.
In manual mode, a trained operator (programmer) typi-
cally uses a portable control device (a teach pendant) to
teach a robot to do its task manually. Robot speeds during
these programming sessions are slow. In the current work
we enclosed the both modes.
The control for the presented robot arm consists basi-
cally of three levels: a microcontroller, a driver, and a com-
puter-based user interface. This system has unique char-
acteristics that allow flexibility in programming and con-
trolling method, which was implemented using inverse
kinematics; besides it could also be implemented in a full
manual mode. The electronic design of control is shown
in Figure 11.
The microcontroller used is an Atmega 368 which
comes with a development/programming board named
“Arduino”, as shown in Figure 12. The programming
language is very similar to C but includes several library-
ies that help in the control of the I/O ports, timers, and
serial communication. This microcontroller was chosen
because it has a low price, it is very easy to reprogram,
the programming language is simple, and interrupts are
available for this particular chip.
The driver used is a six-channel Micro Maestro servo
controller board. It supports three control methods: USB
for direct connection to a computer, TTL serial for use
with embedded systems, such as the Arduino microcon-
troller, and internal scripting for self-contained and host
controller-free applications. This controller, as shown in
Figure 13, includes a 0.25 μs resolution for position and
built-in speed and acceleration control.

A. ELFASAKHANY ET AL.
52
Figure 11. Electronic scheme of control.
Figure 12. Arduino microcontroller board.
Figure 13. Servo controller driver.
The user interface depends on the control method used,
i.e., inverse kinematics or a full manual mode. In the fol-
lowing, each interface is described:
5.1. Inverse Kinematics Control
In this control method, the user inputs the coordinate sys-
tem position where the gripper should be. As consequence,
interface is generated with Labview through a visual user,
as shown in Figure 14. The program automatically per-
forms the inverse kinematics calculations to obtain the
angles that each motor should have and then sends a
command either to the microcontroller or directly to the
driver that will move the robot to the specified position.
Communication is performed with the RS-232 protocol.
In the following, you may see the Labview user interface
inputs and output.
The Labview user interface inputs are:
x axis position.
y axis position.
z axis position.
Gripper opening.
Gripper attack angle.
Serial port.
The Labview user interface outputs are:
Motor A angle.
Motor B1 angle.
Motor B2 angle.
Motor C angle.
Attack angle.
Gripper angle.
Such output variables are treated and sent by an appro-
priate way, so that information can be interpreted in a
correct manner. The outputs are sent via the serial port
which is communicated with the controller. When the but-
ton “Move” is clicked, a process will take place, as shown
in Figure 15. With this action, the robotic arm will change
its position according to the input values. In addition, it has
a standby button that stops the communication controller.
Figure 14. Labview user interface.

53 A. ELFASAKHANY ET AL.
Figure 15. Program process.
The main advantages of this approach are that it uses
an efficient way of moving and offers further capabilities
that could be implemented, such as position and sequence
learning. A disadvantage, on the other hand, is that the
possible positions that have valid angles after the inverse
kinematics calculations are very limited because the servo
motors have a restraint of 180˚.
5.2. Manual Control
This type of control is an extra option for our system that
useful in specific positions. In case of mandatory posi-
tions that the inverse kinematics mode cannot calculate
their valid angles, we may use the manual control instead.
Basically, manual control consists of a series of analog
inputs, such as potentiometers, that are connected with
the microcontroller which will interpret the values and
send a command to the servo driver. In order to imple-
ment this, a control board, as shown in Figure 16, should
be built to work as an interface with the user. Possible
implementation includes a teaching feature where the mi-
crocontroller stores positions in memory and by a keypad
or a series of switches we may recall these positions.
6. Testing and Validation
Several tests were carried out to validate the robot arm and
its components. The testes covered both the particular ele-
ments and the overall system, as shown in Figure 17.
For the microcontroller, the tests are occurred by sending
different commands by the software to the microcontrol-
ler and check changes on the output which was connected
to a servo motor that turned on or off depending on the
command.
The servo motors were tested afterwards by sending
different direct pulses to each servomotor and verifying
the response of moving to the right position. We used a
mark to know where the initial position was and the final
Figure 16. Potentiometer board.
Figure 17. Robot arm tests.
position of the motors is determined by sending a signal
with the microcontroller and, in turn, it is interpreted by
the servo and compared to the signal provided by the
encoder, resulting in the rotation to the desired position.
During this test, the servo motor was inconsistence with
the robot arm system because of an incorrect polarization.
The servo motor driver was also tested using the Lab-
view software to send commands to the microcontroller
which sent the specific commands to the driver which
had one motor connected to change the position accord-
ing to the commend. It is important to notice that at the
beginning of the project a different servo motor driver
was selected but several problems related to the commu-
nication between them and the microcontroller were pre-
sent. So we choose a driver that allows the data to be sent
directly from the computer to it with only a USB wire, so
the microcontroller would only be used in case of the
implementation of manual control.
Other tests were performed to verify the functionality
of the whole system, as shown in Figure 18. Those tests

A. ELFASAKHANY ET AL.
54
Figure 18. Robot arm in action.
were occurred by introducing a specific position in the
Labview interface and measuring the distance between a
reference point and the final point in order to verify: the
correct transformation from inverse to direct kinematics,
the relationship between the specified angles and the ro-
tation of the motors.
Testing and validation of the robot arm is one of the
tasks that require elongated time because several iterations
are needed. During our tests, many problems arise as:
wrong angle calculations, wrong calibration of the mo-
tors, problems with the physical angle and position mea-
surements, and one of the servo motors burned because
of an overload that wasn’t expected.
7. Results and Discussions
Results from the robot arm at different operating condi-
tions are presented as follows:
7.1. Servo Motors Movement Range
The limits of the servo motors were obtained since speci-
fication of this type of motors contains that it has less
than a 180 degree span. The real range for all motors was
found to be in the range 125 - 142 degrees, as shown in
Table 1. This clearly demonstrated that real operation of
robot arm is different from the stander case.
Table 1. Motor angle ranges.
Motor Angle Range
Motor A 130˚
Motor B1 135˚
Motor B2 140˚
Motor C 142˚
Motor Attack Angle 125˚
7.2. Current Consumption
The current consumption depends on the load and the type
of motion of the robotic arm. In the current study, there
are 4 levels of current consumptions:
 Low (from 0 to 200 mA). This consumption takes
place when the robot is at rest (not motion case).
• Normal (from 200 to 500 mA). This happened when
the robot arm is moving with capability to go to the tar-
get without needs of great torque.
• High (from 500 mA to 900 mA). This range is rea-
ched at the beginning of carrying loads. By overcoming
the initial moment of inertia for loads, the normal range
takes a place.
Over current (more than 900 mA). The load is too hea-
vy and the motor cannot move at all. For being under this
condition for more than one minute, the motor will burn,
i.e. it is not possible to be used any more.
7.3. Maximum Load
These results were obtained using different weights; a bag
of corn was used with a scale to determine bag’ weight.
Results carried out by using the robot arm to pick up the
bag and move it to specific positions. Table 2 presents
the current consumption at different weights of bag of corn.
From Table 2, it can be seen that the robot can move
without problems at loads lower than 50 grams. At loads
60 grams, the robot arm start having difficulties and after
passing 80 grams severe condition occurred where ire-
versible damage could be happened in motors.
7.4. Final Position
Results show the precision of the robot arm to move dif-
ferent weight (<50 grams) is presented in Table 3. As
shown, the robot arm is able to perform the movement to
the position specified. However, this movement is not
smooth and sometimes the motors do not have enough
force, especially when the load is heavy. In addition,
some problems may appear due to synchronizing the two
bottom motors. The steps of the two motors were not
coincidental and that causes tension in the acrylic parts,
which in case of being too much will break the parts.

A. ELFASAKHANY ET AL.
55
Table 2. Load vs. current consumption.
Load Current Consumption
20 grams Low
40 grams Normal
50 grams Normal
60 grams High
80 grams Overcurrent
100 grams Overcurrent
Table 3. Precision on all axis.
Axis Precision (+/-)
x 1 cm
y 2 cm
z 1 cm
8. Conclusions
This paper presents the design, development and imple-
mentation of robot arm, which has the talent to accom-
plish simple tasks, such as light material handling. The
robot arm was designed and built from acrylic material
where servo motors were used to perform links between
arms and execute arm movements. The servo motors in-
clude encoder so that no controller was implemented; how-
ever, the rotation range of the motor is less than 180º
span, which greatly decreases the region reached by the
arm and the possible positions. The design of the robot
arm was limited to four degrees of freedom since this
design allows most of the necessary movements and keeps
the costs and the complexity of the robot competitively.
The end effector is not included in the design because a
commercially available gripper is used since it is much ea-
sier and economical to use a commercial one than build it.
During design, we faced some difficulties due to the way
of joining thin acrylic parts strongly. A mechanical junc-
tion based on screws and nuts is used and in order to ac-
complish that, a small feature was designed which allowed
fastening the bolts with the nuts without having to screw
in the thin acrylic layer.
To control the robot arm, three approaches are imple-
mented: a microcontroller, a driver, and a computer-based
user interface. This system has unique characteristics that
allow flexibility in programming and controlling method,
which was implemented using inverse kinematics; be-
sides it could also be implemented in a full manual mode.
This robotic arm is contrast with others as being much
cheaper than available robot arms, also it can be controlled
all of its movements from a computer, using a Labview
interface.
Several tests were carried out to validate the robot arm
where the testes covered both the particular elements and
the overall system; results at different operating conditions
show trustful of the robot arm presented.
9. References
[1] Manipulating Industrial Robots—Vocabulary, Interna-
tional Organization for Standardization Standard 8373,
1994.
[2] Industrial and Service Robots, IFR International Federa-
tion of Robotics, 2010. http://www.ifr.org/home
[3] Case Studies and Profitability of Robot Investment, The
IFR Statistical Department, 2008.
http://www.ifrstat.org/downloads/2008_Pressinfo_english
.pdf
[4] R. J. Wang, J. W. Zhang, et al., “The Multiple-Function
Intelligent Robotic Arms,” FUZZ-IEEE Journal, Korea,
20-24 August 2009, pp. 1995-2000.
[5] L. B. Duc, M. Syaifuddin, et al., “Designing 8 Degrees of
Freedom Humanoid Robotic Arm,” International Confer-
ence on Intelligent and Advanced Systems, Kuala Lumpur,
25-28 November 2007, pp. 1069-1074.
[6] C. R. Carignan, G. G. Gefke and B. J. Roberts, “Intro to
Space Mission Design: Space Robotics,” Seminar of Space
Robotics, University of Maryland, Baltimore, 26
March 2002.
[7] Occupational Safety and Health Administration Technical
Manual, OSHA 3167, United States Department of Labor,
1970.
[8] B. Siciliano, L. Sciavicco, L. Villani and G. Oriolo, “Ro-
botics, Modelling, Planning and Control,” Springer, Lon-
don, 2009.
[9] M. P. Groover and M. Weiss, “Robotica Industrial, Tech-
nologia, Programacion y Aplicaciones,” Mc-Graw Hills,
Mexico D.F., 1989.