Presentation - College of Engineering

copygrouperMechanics

Nov 13, 2013 (3 years and 8 months ago)

83 views

Automatic Separation of Objects Using
Intelligent Mechanism

(InteMech)

Omer Mohammed Ali

980211585

Masoud Abdullah Moh’d

980715175

Sohail Khalid AL
-
Awadhi

980710590

Hamad Abdullah Rashid

970710100



United

Arab Emirates

University

College of

Engineering

Advisor: Dr. A.M.M Sharif Ullah

Introduction


InteMech main system.



The sub
-
systems:


Control system.


Vision ( Robot eye).


ANN (Robot Brain).



Plan for next semester.




InteMech System




Object

Web Camera

Image Processing + ANN

Inputs

Action

Robot Arm

Mechanic

System


In this project the robot is
already designed


(
by Dr. Khalifa H. Harib
), but
there is a need to modify
the design to satisfy the
desired objectives of the
project.

Arm Robot

Robot Gripper


Transitions Wire

Gripper

Grippers are the most
common end
-
effectors. They
provide the equivalent of a
thumb and an opposing
finger, allowing the robot to
grasp small parts


Spring

Robot Gripper

Spring

Open

Closed

Robot Kinematics




Forward Kinematics.



Inverse Kinematics.



3
D schematic diagram

Robot Kinematics

2
D schematic diagram

X

Y

Z

Φ


1


2


3


4

Forward Kinematics

Inverse Kinematics

Robot Control



Controlling the robot arm
would be possible after the
inverse kinematics
relations were found.


C language program will
control robot Actuators
(stepper motors).


Why the stepper motor was
selected to be the robot
actuators?




This part is very important because its deal
with the robot vision and how it can interact
with the environment.




Robot Vision

Software:



Vision For MatLab (VFM) Toolbox



Image Processing Toolbox

Robot Vision

Hardware:


Web Camera

Web Camera

A normal web camera was selected
to capture a picture from the
environment, because there are no
need for high quality pictures

Robot vision

Example for Robot Vision

Hand written letter (E)


The picture was captured
from the environment using
web camera and VFM
Toolbox

Binary image

Example for Robot Vision


The picture was Converted
to binary using the Image
Processing Toolbox.



Problems faced using this
technique:



Huge number of inputs.

Line
1

Line
2

Line
3

…….

…….

…….

Example for Robot Vision

Inputs Lines

Proposed solution:

Modified binary image

Example for Robot Vision

Artificial Neural Network


What is ANN?






Artificial Neural Network


The Transfer Functions.

log
-
sigmoid transfer function

Linear transfer function

X
Y
linear

X
sigmoid
e
Y



1
1
Artificial Neural Network


Learning Method (Back
-
propagation)



Imply a backward pass of error to each
internal node within the network.


Calculate weight gradients for that node.


propagating forward the activations and
propagating backward the errors.


The Back
-
propagation Equation:



j
i
ij
ij
x
w
w




Line
1

Output
1

Line
2

Output
2

Line
3

Output
3

Artificial Neural Network

Artificial Neural Network

Output
1

Output
3

Output
2

Letter
1

Letter
2

Letter
3

Cost Estimation

Cost

#
Item

Cost/Item

Item Name

80

1

80

Web Camera

900

3

300

Stepper Motor type
1

240

2

120

Stepper Motor type
2

700

2

350

Gearbox

1920





Total cost

Impact


This project is using low cost equipment to
do sophisticated technology.



Applicable in many industrial applications.



The robot could be used in many courses
like control and dynamics system or any
courses related to robotics subject.



Next semester plan


Building the Robot Arm.


Improving the ANN.


Connecting The Three Sub
-
Systems.


Testing and modifying InteMech System.

Conclusion

THANK YOU