Design and Experimental Tests of a Pick and Place Robot: Theoretical and

Experimental Approaches
Hasan Ghorabi*. Yaser Maddahi**. Seyyed Mohammad Hosseini MonseI* and Ali Maddahi***
* School oI Mechanical Engineering . Islamic Azad University (Saveh Branch). Saveh. IRAN
**Department oI Mechanical and ManuIacturing Engineering. University oI Manitoba. Winnipeg. MB. Canada
***Faculty oI Mechanical Engineering. Khaie -Nasir University oI Technology. No. 17. Pardis St.. Mollasarda
Ave.. Vanak Sq.. Tehran. Iran.
Email: ymaddahi(umanitoba.ca
Abstract. This article covers the Design process oI a Iour DOF pick and place robot considered and manuIactured
Ior research aims. Here. the modelling and simulation oI this robot have been done using analytical methods and
Visual Nastran soItware. Also. some pre-deIined experiments are perIormed based on ISO 9283 regulations and
requirements. Finally. an algorithm is presented to calculate the maximum allowable load carrying capacity
(MALCC) as the main contribution oI this research study. As shown in this paper. the derived data Irom MALCC
tests speciIy a close relationship between analytical. simulation and experimental approaches.
Kev-Words. - Pick and Place Robot. Robotic Arm. Design. ISO 9283. Modeling. Maximum Allowable Load
Carrying Capacity.
1 Introduction
Recently. interests in design. manuIacturing and
control oI robotic stations have increased with a
various theoretical and practical contributions being
made. The industrial robots have been implemented in
many Iields oI human liIe. Particular advantages oI
these robotic systems include:
a) Easily controlled/ programmed movements.
b) High accuracy.
c) Accuracy. speed and pay load capacity constant
over entire working range.
d) Control system simplicity.
e) Inherently stiII structure.
I) Large area coverage.
g) Large payload capacity.
h) Structural simplicity. oIIering good reliability.
Korayem et al. presented the design oI a robot Ior
industiral aims in the assembly line. and the proposed
robot is applied Ior one oI the important stations in the
assembly line. Kinematic and dynamic modeling.
Iinite element analysis (FEA). quality Iunction
deployment (QFD). and Iailure mode and eIIect
analysis (FMEA) are used in this paper |1|. Ghayoumi
has presented the experimental test on 3P robot.
Simulation results have demonstrated improvement in
comparison with neural network method |2|. Azhdari
has presented a dynamic model Ior a two degree-oI-
Ireedom planar robot arm. The links oI the arm.
connected to prismatic and revolute ioints. are
considered to be Ilexible. They are assumed to be
Iabricated Irom either aluminum or laminated
composite materials. The model is derived based on
the Timoshenkov beam theory in order to account Ior
the rotary inertia and shear deIormation |3|.
Wondergem developed an adaptation mechanism Ior
linear Iriction parameters in order to achieve mutual
synchronization Ior one direction oI the CFT robot
system. Also in this research. he tested the adaptive
mutual synchronization scheme on the CFT robot
system in the DCT laboratory at the Eindhoven
Uni\versity oI Technology |4|. Callegari et al.
presented a case oI the CO2 high-speed Cartesian
robot produced by Campetella Robotic Center. This
robot is characterized by good dynamic perIormances
but has been chosen by the producer Ior re-
engineering that should allow it to meet even more
advance targets |5|.
Wang and Ravani were developed this technique to
maximize the DLCC Ior an entire traiectory rather
than in the neighborhood oI the robot conIiguration.
In these works Iixed base. rigid links and ioints are
assumed |6|. Seraii presented an online approach Ior
motion control oI mobile manipulators using
augmented Jacobian matrices. The approach is
kinematic and requires additional constraints to be
met Ior the manipulator conIiguration |7|. Maddahi
has proposed an algorithm Ior determination oI
maximum allowable load carrying capacity oI a
redundant mobile robot designed Ior obiect
Proceedings of the 9th WSEAS International Conference on APPLICATIONS of ELECTRICAL ENGINEERING
ISSN: 1790-2769 144 ISBN: 978-960-474-171-7
transIering. Also he determined and corrected the
motion errors oI a mobile manipulator |8|.
In this paper. design. modeling and simulation process
to solve direct and inverse kinematics oI a pick and
place robot are prescribed. AIter discussing on
simulation soItware results oI pick and place robot.
we simulated control and perIormance tests oI robot
and at last the results oI tests according to ISO 9283
standard requirements are analyzed. Moreover. an
algorithm Ior determination oI load capacity oI the
manuIactured robot is proposed and Iinally. in order
to validate the the suggested algorithm. consequences
oI analytical. simulation and experimental approaches
are done on this robot.
2 Design of Pick and Place Robot (PaP)
This PaP robot has been designed and constructed in
IAU Robotic and Automation Research Lab. (Fig. 1).
The robot is designed on basis oI an assumption in
which each ioint has an independent actuator with
gear reduction and measuring angular ioint position
sensor. In keeping with this. mechanical design and
manuIacturing process oI the robot were done using
Mechanical Desktop. In addition. the design oI PaP
is carried out using ISO14971 requirements which
discuss about the risk assessment oI product in design
and production stages. Using this international
standard. the risks and hazards are recognized and
reduced during the design and manuIaturing steps.
The schematic mechanical plan modeled in
Mechanical Desktop soItware is shown is Fig 1-a.
Also. Fig. 1-b depicts the top view oI the mechanical
structure oI the manuIactured robot. The combination
oI all Iour motors creates the Iinal motion oI PaP.
3. Kinematics Equations
Within kinematics analysis. the study oI the position
and velocity variables are done. The kinematics oI
manipulators involves the study oI the geometric and
time based properties oI the motion. and in particular
how the various links move with respect to each other
over the time. Figure 2 depicts the kinematic diagram
oI this robotic manipulator. Also Table 1 shows the
Denavit-Hertenberg variables oI this robot.
Table 1. D-H parameters oI PaP
Joint 0
i
q
i
a
i
d
i
1
1
-90 0 0
2
2
0 a
2
0
3
3
90 a
3
0
4
4
0 a
4
0
Fig. 1. Modelled (a) and manuIactured (b) PaP robot
Fig. 2. Kinematic diagram oI PaP associated
coordinate system
a
b
Proceedings of the 9th WSEAS International Conference on APPLICATIONS of ELECTRICAL ENGINEERING
ISSN: 1790-2769 145 ISBN: 978-960-474-171-7
3.1 Direct Kinematics of Robot
For an n-axes rigid-link manipulator. the direct
kinematics solution gives the coordinate Irame or pose
oI the last link. For PaP robot. direct kinematics
equations will be as Iollows:
( ) ( ) ( ) ( )
4
3
4 3
2
3 2
1
2 1
0
1
u u u u A A A A T = (1) (2)
Where.
3 2 1
. . u u u and
4
u are generalized coordinates.
AIter expanding the right side oI the Eq. (2). we have:

=
1 0 0 0
z z z z
v v v v
x x x x
p a s n
p a s n
p a s n
T
(2)
(3)
where the martix indexes are:
4 1 3 2 3 2 4 1
) ( S S S S C C C C n
x
=
) (
3 2 3 2 4
S C C S C n
z
+ =
4 1 4 3 2 1 4 3 2 1
C S S S S C S C C C S
x
+ =
4 1 4 3 2 1 4 3 2 1
C C S S S S C C C S S
v
+ + =
4 3 2 4 3 2
S S C S C S S
z
+ =
(3)
3 2 1 3 2 1
C S C S C C a
x
+ =
3 2 1 3 2 1
C S S S C S a
v
+ =
3 2 3 2
C C S S a
z
+ =
2 1 2 3 4 3 3 1 2 3 4 2 1
) ( S C a C a S a C S S d C C P
x
+ =
2 1 2 3 4 3 3 1 2 3 4 2 1
) ( S S a C a S a S S S d C S P
v
+ =
where
2
a .
3
a and
4
a are the lengths oI links as
shown in Table 1.
3.2 Inverse Kinematics of PaP
Given a desired position and orientation Ior the end
eIIector oI robot. Ior Iinding values oI the ioint
parameters which satisIy the direct kinematics
equations. we need to solve inverse kinematics oI
robot.
4. Software Simulation Using Working
Model
In order to validate the derived mathematical
Iormulations. a consequence oI simulation was carried
out using the model designed in Mechanical
Desktop and simulated using Working Model.
As shown in Fig. 3. the torques derived Irom
simulation study cane be Iitted as the Iunctions oI the
robot parameters and time.
Fig. 3. Torques oI Motor 3 derived Irom theoretical
and soItware results
5. Performance Tests
The manuIactured robot should accomplish the given
commands accurately and smoothly. This is possible
in the case that the motion oI the end eIIector oI the
robot is accurate enough relative to the target-obiect
that is the point that the end eIIector oI the
manipulator has reached to. The accuracy oI actual
robot is under the eIIect oI the Iollowing Iactors such
as the accuracy oI manuIacturing mechanical parts oI
the robot. the accuracy oI assembling the constituting
part oI robot. accuracy during the robot operation that
is inIluenced by external Iorces. electronics system
accuracy and motors operations. the clearance existing
in the system. wear behaviours. that is change in
accuracy oI the robot in long duration. change in
accuracy oI system aIter assembling the disassembled
parts due to repair and change in the system accuracy
during the PM (Preventive Maintanace) periodic
programs.
So in simulator program these errors are Iiguratively
inserted. Despite the recent international eIIorts by
many oI the standard committees. research and
industrial labs. many oI the robots users still sustain a
loss. as a result oI the lack oI the standard. technical
approaches and necessary determinations oI the robot
examinations.
This matter is caused by complexity oI the most robot
designations and their vast limitations. In this research
we try to do some oI these approaches by using
camera and visual system according to the standards
such as ISO9283.
Proceedings of the 9th WSEAS International Conference on APPLICATIONS of ELECTRICAL ENGINEERING
ISSN: 1790-2769 146 ISBN: 978-960-474-171-7
Fig. 4. The designed soItware oI robot using Visual Basic soItware
The aim oI these standards is providing technical
inIormation to help users to select the most convenient
robot Ior their purposes. These standards deIine
important principles based on the path. and then
diIIerent appearances will be seen to evaluate them.
These principles are approximate accuracy oI the path.
absolute accuracy oI the path repetition ability oI the
path rapidness speciIications. and corner variable.
Evaluation oI the mentioned principles is one oI the
most convenient ways Ior evaluation oI the whole
activity based on industrial robots path. Also. the
measurement oI these principles brings the
opportunity oI comparing similar robots operations.
To make tests more applicable. statistic analyzes
according to ISO9283 standards are perIormed. The
interIace page and some Ieatures oI designed soItware
is shown in Fig. 4.
The results oI experimental test Ior straight Iorward
traiectory are shown in Fig. 5. As shown. the position
oI the robot end-eIIector has some errors in this path
which originated Irom some resources such as Iriction
Iorce misalignment in ioint.
6. Load Capacity of P aP
In order to determine dynamic load carrying capacity oI the
mobile manipulator. proper modelling oI robot is a
prerequisite. Here. the robot dynamic computations are
executed in three methods consists oI the amount derived Irom
analytical Iormulations. soItware approach (Working Model)
and experimental tests. The assumed traiectory in all method is
the path presented in Fig. 6 and named as 'load and
end-eIIector traiectory.
0
1
2
3
4
5
6
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Desired Path
Actual Path
Fig. 5. Actual and desired traiectories oI robot motion
in Iorward straight path
6.1. Software Simulation Method
In next method the dynamic load capacity is calculated
using soItware simulation study. In this manner. Ior
determining the maximum load capacity. the model is
moved in given traiectory in Working Model soItware
and the torques oI motors are derived Irom soItware
output. In this estimation. the amount oI plus load mass
is increased in each trying and the torques is checked.
When one oI motors reaches to its critical torque limit
in torque. it meaning that robot is carrying its maximum
load in manipulator.
Proceedings of the 9th WSEAS International Conference on APPLICATIONS of ELECTRICAL ENGINEERING
ISSN: 1790-2769 147 ISBN: 978-960-474-171-7
Fig. 6. Flow chart oI maximum allowable load capacity calculation |8|
For calculation oI maximum allowable load. PaP is
moved in given traiectory and the initial load is taken to
zero. then the load is increased by steps oI 0.1 kg.. As a
results according to this experimental tests results. Ior
g m
load
600 = . motor 2 (Y axis motor) has a maximum
value mm N. 200 = t . Then. the maximum allowable
load Ior given traiectory is determined
as kg m
allowable
6 . 0 = .
For i1...n and For i1...p
( ) ( ) ( )
T
k Z i Y i X k i i X | ) ( | . . =
Start
Input the given trajectory and [Ic]
Discretize the path in to m points in plane and then n and p points in the X and Y directions
For R1..m
and I1.2.3
Find ioints vector space
( ) R q

II
( )
max min
q R q q
f

< <
.
( )
max min
q R q q
f

< <
Find Jacobian matrix
Det|J(k)|0
Find Motion Equations
Find Torques oI Actuators
Optimum base location and maximum load
max . max . max .
.
load b b
m and Y X
Compute
) 1
nl
q
e

.
) 1 (
) (
nl
s
q
T T
e

+ =

( ) ) ( ) (
) ( ) (
k T k T k T
+
< <
Compute
( ) i i m
load
.
Finish
Proceedings of the 9th WSEAS International Conference on APPLICATIONS of ELECTRICAL ENGINEERING
ISSN: 1790-2769 148 ISBN: 978-960-474-171-7
6.2. Computational Method
In the second step. the computational procedure Ior
determining the maximum allowable load capacity is
outlined and also Ilow charted in Fig. 6. It can be seen.
aIter discretizing the given traiectory into m points.
ioint motion constraints. Jacobian singularity conditions
and ioint velocity constraints are checked |8|.
In cases in which each oI these constraints is violated.
the given traiectory is unrealizable and a new traiectory
should be selected. At next step. ioint accelerations are
Iound and then the lagrangian Iormulation is used to
compute load and end eIIector dynamic eIIects. When
program reaches the dashed line loop beIore the Iinal
step. it iumps backward and selects a new discretized
point along the given traiectory until the conditions
km is satisIied.
To investigate the proposed algorithm. a simulation
study is presented. In this study. a speciIied traiectory
Ior the load is assumed. In most cases the base motion
due to much inertia most be planned much smoother
and slower than the manipulator motion. ThereIore.
only actuators constraints are considered to Iind
maximum allowable load oI the mobile manipulator.
Considering the modeling equations. proposed
algorithm and type oI actuators the task space traiectory
is descritized into equally spaced m9 points. The
allowable load carrying capacity Ior the mobile
manipulator at every point oI the traiectory is
determined and maximum allowable load is Iound
kg m
allowable
45 . 0 = at (X(t)6.8 cm. Y(t)0 cm.
Z(t)4.5cm) as shown in Fig. 7.
6.3. Experimental Approach
In this approach. the amounts oI torques are derived
using laboratory tests applied on this robot. The
obtained data are perIectly experimentally and all
existing errors sources such as Iriction eIIects have
been considered. The amount oI load capacity Ior this
path is obtained about 0.39 kg. Based on a logic
evaluation. it should be obtained less than this value
Ior soItware and theoretical tryings.
As shown in this section. there are some diIIerences in
the results oI two methods in calculation oI maximum
allowable load carrying capacity ( kg m
allowable
6 . 0 = Ior
the soItware study. kg m
allowable
45 . 0 = Ior the
computational calculation and kg m
allowable
39 . 0 = in
the experimental approach). The eIIects oI Iriction
Iorce. load and inertia distribution types are the maior
reasons oI this diIIerences.
7. Conclusions
In this paper. design. modeling and manuIacturing oI
a 4 DOF pick and place robot were investigated.
Then. three diIIerent approaches were done on a
Iabricated robot in order to obtain the amount oI
maximum load capacity oI robot. The obtained data
were perIectly experimentally and all available
structural errors sources such as Iriction eIIects.
misalignment in ioints and backlash. The amount oI
derived load capacity in soItware. analytical and
experimental studies conIirmed the theoretical side oI
MALCC Iormulations.
Fig. 7. The variation oI the load along the load traiectory and MALCC
Proceedings of the 9th WSEAS International Conference on APPLICATIONS of ELECTRICAL ENGINEERING
ISSN: 1790-2769 149 ISBN: 978-960-474-171-7
8. References
1. M.H. Korayem. N. Shiehbeiki and T. Khanali.
'Design. manuIacturing. and experimental tests oI a
prismatic robot Ior assembly line. The International
Journal oI Advanced ManuIacturing Technology. pp.
379-388. 2005.
2. M. Ghayoumi. P. Porkar Rezayeyeh and
M. H. Korayem. 'Correlation Error Reduction oI
Images in Stereo Vision with Fuzzy Method and Its
Application on Cartesian Robot. Lecture Notes in
Computer Science. Volume 4304. pp. 1271-1275. 2006.
3. A. Azhdari. N. G. Chalhoub and F. Gordanineiad.
'Dynamic Modeling oI a Revolute-Prismatic Flexible
Robot Arm Fabricated Irom Advanced Composite
Materials. J oI Nonlinear Dynamics. pp. 171-186. 1991
4. M. Wondergem. 'Adaptive mutual synchronization
oI a CFT robot system. Traineeship report. Eindhoven
University oI Technology. Department oI Mechanical
Engineering. Dynamics and Control group. Eindhoven.
March 8. 2004.
5. Massimo Callegari. Ferdinando Cannella. Sergio
Monti. Claudio Santolini and Paolo Pagnanelli.
'Dynamic Model Ior a Re-Engineering oI a High-Speed
Cartesian Robot. IEEE/ASME Int. ConI. on Advanced
Engineering Mechatronic Proceedings. Italy. 2001.
6. L. T. Wang and B. Ravani. 'Dynamic load carrying
capacity oI mechanical manipulators-Part 1: Problem
Iormulation. J. oI Dyn. Sys. Meas. And Control. vol.
110. pp. 46-52. 1988.
7. H. Seraii. 'A uniIied approach to motion control oI
mobile manipulators. Int. J. oI Rob. Res.. Vol. 17. No.
12. PP. 107-118. 1998.
8. Y. Maadahi. "Calculation oI Load Carrying Capacity
on a Redundant Manipulator". 2nd WSEAS Int. ConI.
on Circuits. Systems. Signal and Telecommunications
(CISST'08). Acapulco. Mexico. January 25-27. 2008.
Proceedings of the 9th WSEAS International Conference on APPLICATIONS of ELECTRICAL ENGINEERING
ISSN: 1790-2769 150 ISBN: 978-960-474-171-7