Detecting Object Position and Size Using Artificial Intelligence

This document discusses the use of artificial intelligence, MATLAB, and ultrasonic sensors for detecting object position and size. It explores the motivation behind the research, key literature and theoretical framework, and the significance of the study. The document also presents hypotheses and research questions related to the topic.

18392124412819022887.docx
Document Owner : Head of University College
Page 1 of 14
INITIAL RESEARCH PROPOSAL FORM
(also referred to as ‘Statement of Intent Form’)To be submitted by the researcher to the Institute Research Sub-Committee (IRC)
Research Title:
Detecting oObject position and size detection using AIArtificial Intelligence to detect size of an object
Institute name
Institute of Eengineering and Ttransport (Electrical and Electronics)
Course / Programme:
Bachelor of Engineering in Electronics Engineering
Level and year of study
Level 6
Main area of study being proposed:
1. Power Electronics
2. Digital Electronics
3. C++ Programming
4. Simulation Using MATLAB
5. Digital Electronics
6. Microcontrollers i.e. Arduino
Digital ElectronicsArtificial Intelligence (Deep Learning)
Simulation Using MATLAB
Microcontrollers i.e. Arduino
Name of Researcher:
Rubin El-ellaghi
Researcher’s I.D. Number:
0114050A
Signature of Researcher Date of submission of Form
14/03/2019
Name of Tutor (or Recommended Tutor):
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 2 of 14
Dr. Ing Joseph A. Zammit and Jesmond Gatt
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 3 of 14
Personal Motivation for the Choice of Research Theme.
As human beings we have been able to identify, recognize and approximate sizes and shapes of objects
just by looking at them. And over time we have evolved to using cameras and then using computation
techniques to calculate the estimate size of the object.
Using artificial intelligencedeep learning with MATLAB for object recognition in their natural setting and
then measuring sizes of the objects is considerably a new form of artificial intelligence in science
advancements. This is possible if we combine the Matlab software fed up with data by from the ultrasonic
sensor and Arduino as thea microcontroller as a medium to convert the signals between the ultrasonic and
the Matlab.
Ultrasonic sensor enables one to measure the sizes of objects by non-contact method.
I am motivated by various inventions and advancements that have been made in this field of artificial
intelligence, simulation and digital electronics. Some of the projects that have become a success include
object recognition using AI and simulation to detect colors of objects. If this project becomes a success it
will considerably reduce computations and process this information on a real time basis.
Outline of Key Literature and Theoretical Framework or Propositions.
Artificial Intelligence
Artificial intelligence sometimes referred to as machine intelligence is the intelligence demonstrated by
machines. Artificial intelligence, studied in a broader topic of robotics, is being used in making of tools
which have the capabilities to imitate humans in performing many tasks. Artificial intelligence software is
fed into robots or computers or other related tools. It works by collecting large swathes of data with
complex functions and then allowing the software to learn skills from patterns of the inserted data,
AI is the one behind today’s driverless cars, and apple’s SIRI talking program. Artificial intelligence has
helped in development of tools which are useful in solving numerous tasks and increasing the efficiency
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 4 of 14
and speed of solving the tasks.
Artificial intelligence has automated repetitive learning and discovery through data. This has helped to a
great deal in performing large volume, frequent and computerized tasks without fatigue unlike automation
from hardware.
Al is also easy to adapt to repetitive learning algorithms. Artificial intelligence makes the algorithm
acquire a skill through repetitive and progressive entry of data and this leads to the point whereby the
algorithm can be a predictor or a classifier.
Applications of artificial intelligence in today’s life is everywhere to see. For instance, AI has made the
internet a marketplace for business whereby patterns of data can be learned to help companies plan. In
healthcare AI applications are providing information such as x-ray readings.
Some of the advantages of using AI include improvement of existing tools and their abilities in what tasks
they do, give humans better understanding and improve our memory and much more.
The challenges that AI is facing include; AI cannot function like humans in multitasking such as tools
developed to play poker cannot play chess. This means that AI only works in defined data. In addition, AI
is going to change every industry and so the conflict that will arise between humans and robotics will be
unimaginable and ways have not been devised on how to control the conflicts such as loss of jobs. [1]
Using Ultrasonic Sensor to Detect Size of an Object
Ultrasonic sensors have been in use all over the world for some time now. Ultrasonic sensors work by
sending sound pulses with a frequency high above human hearing frequency to the object and then
measuring the time the sound signal takes to return to the transducer. This data then is computated to
determine the distance and the size of the target (object)
The ultrasonic sensors can be use in measuring proximity and ranging. In proximity detection, the sensor
has pre-recorded range within which if an object is passing within will be detected and sound signal
reflected back to the transducer.
Advantages of Ultrasonic sensors is that since they measure the sizes and presence of objects using non-
contract method, they are preferably used in environmentally difficult places. And the sensor is also not
affected by optical factors such as color, reflectivity, transparency and opaqueness.
Nowadays ultrasonic sensors such as senix ultrasonic sensors are used in various machines in
determination of presence and sizes of objects.
Limits of using ultrasonic sensor to detect size of an object is that it is difficult to measure object at longer
distances. In addition, they cannot be used to measure sizes and proximity of curved or sound absorbing
objects such as fabrics. [4]
Artificial Intelligence and MatLab
MatLab is a tool that uses data entered into it to produce different output by forming algorithms and__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 5 of 14
complex functions. It works with all software and hardware that are dependent on data-dependent systems.
Matlab and artificial intelligence works in situations whereby traditional programming method fail.
Normally this would not give desired output in situations where there are progressive data. With AI and
MatLab one only feeds the software data and the desired output and the software will write the program
without sequence of the traditional program. Basically meaning they can be used as predictors and
classifiers.
For instance, in today’s world AI is combining MatLab and neural networks in detection and
determination of the sizes of objects in various environment After thorough studies and research of the
above projects I realized that artificial intelligence with simulation by MatLab is of paramount importance
in enhancing clarity, accuracy and speed of tools and other hardware developed for the main purpose of
object detection, diameters and sizes measurements and even their colors. My projects will aim in
detection of real time objects, computation of their sizes and relaying of data immediately. [2]
Various studies have been done in the past in this field of study. The projects have been different
approaches and have different objectives.
1.MATLAB to measure diameter of an Object using Ultrasonic Sensor and Arduino
This project that was done in 2014, used artificial intelligence the sizes of objects using MATLAB and
image processing toolbox application to measure the estimate sizes of objects using their images.
This project was done by using MATLAB app where the object being detected by the Ultrasonic sensor.
The pulse signal generated is sent to the Arduino microcontroller. The Arduino then sends data to the
Matlab which generates algorithmsdoes the analysis [1]
This study does not use artificial intelligence or advanced coding language. In addition, the study only the
images to estimate the sizes of objects which may not give the exact size of the object. This is because it is
the image that is used in size estimation and not real objects and thus cannot be used to measure real time
objects such as moving people.
My project will be using artificial intelligence to measure the sizes of objects, and an advantage will be
done by relaying real time data to the user.
2. Tunnel Excavation using Ultrasonic Sensor
This project was developed by a Japanese construction company. It works by training a deep learning
network which automatically recognize metric based images of the tunnel face. This helps in measuring
the fractures in tunnels and the spacing between fractures. This helps in assessing the suitability of the site
where the tunnel is being constructed.[2]
The only limitation of this project is that the project needs millions of images to recognize and distinguish
different fractures. In addition, it only uses images so it can take a lot of time in processing the data before
a conclusive assessment is made.
My project works on real time computation where it uses AI with MATLAB to form algorithms and
functions to compute sizes of different metric objects of a site and thus show the suitability of a site for
construction of a tunnel.
3.Object Detection using Raspbery Pi
This is a tiny hardware that was designed by using complex deep learning models. It is basically in
detection of real time images such as passing cars in road over a certain period of time, number of
signatures in a document and people like in malls.[3]
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 6 of 14
The tool works by using a multiple of data being fed into system like the millions of facial images of a
person after which it can then establish like detection of criminals with criminal record. It is basically the
tool that is used in making drones which are in identification of people such as terrorists by putting the
images in segments and then draw bounding boxes.
The limitation of this tool is that it cannot be used to detect sizes and diameters of objects or bodies under
detection, and suffers from the difficult of creating millions of training data.
My project seeks to solve this problem by using AI with Matlab and Ultrasonic sensor with Arduino to be
able to detect precise sizes and diameters of objects and bodies.
AI with MATLAB and Ultrasonic Sensor and Arduino
After thorough studies and researchreviews of the above projects I realized that artificial intelligence with
simulation by Matlab is of paramount importance in enhancing clarity, accuracy and speed of tools and
other hardware developed for the main purpose of object detection, diameters and sizes measurements and
even their colors. My projects will aim in detection of real time objects, computation of their sizes and
relaying of data immediately.
In using Artificial intelligence with Matlab and Ultrasonic Sensor and Arduino it will be easier to measure
sizes and location of objects (within an enclosed space) by non-contact method.
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 7 of 14
Significance of the Study.
This project can be very useful in a number of everyday life applications. For instance, it can be very
useful in manufacture of driverless cars. This is possible because by developing the simulation and
artificial intelligence a driverless car is able to measure to distance between it and the car in front and thus
determine when to slow down, speed up or stop and thus avoid causing accidents by non-contact method.
Also using artificial intelligence with Matlab and Ultrasonic sensor and Arduino measuring in object sizes
detection helps in improving the workability of drones used in detection, and prevention of high level
crime such as detection terrorists in areas which are too risky for security personnel to access and fight
terrorism.
Hypotheses and/or Research Questions
The major hypothesis is whether it will be costly to build a tool utilizing ultrasonic sensors to relay
impulses to the microcontroller Arduino and the signals transmitted to algorithms and complex functions
in Matlab to measure exact sizes and diameters of objects.
Matlab software together with AI and Ultrasonic sensors and Arduino microcontroller should be able to
build the tool for recognition and detection of objects and their sizes easily and less time when even used
by people with little or no knowledge on their area such as data scientists.
Matlab together with AI and Ultrasonic sensor and Arduino should be able to unify different work
environments and domains into a single work flow. This can be helpful in feeding complex algorithms into
advanced levelled machines such as Robots.
Can AI be adopted to read accurately the location of an object enclosed within an array of ultrasonic
sensors? Matlab software together with AI and Ultrasonic sensors and Arduino microcontroller should be
able to build the tool for recognition and detection of objects and their sizes easily and less time when
even used by people with little or no knowledge on their area such as data scientists.
Can the size (and maybe shape) of an object be determined by ultrasonic signals fed to an AI system built
by Matlab? The hypothesis is
Ultrasonic sensor and Arduino should be able to unify different work environments and domains into a
single work flow. This will be in trying to determine the size of the object by sending signals into AI built
in MatLab. [3]
In conclusion, major hypothesis is whether it will be costly to build a tool utilizing ultrasonic sensors to
relay impulses to the microcontroller Arduino and the signals transmitted to algorithms and complex
functions in Matlab to measure exact sizes and diameters of objects.
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 8 of 14
Target Participants and Research Methods for Data Collection and Analysis
N/A
All data will be collected through electronic sensors/equipment
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 9 of 14
Ethical Considerations.
Refer to guidance points below. You are also additionally required to read MCAST Document 074
‘Research Ethics Policy and Procedure’ that is available on the College website via link
http://www.mcast.edu.mt/MainMenu/Full-TimeCourses/Rules,PoliciesandRegulations.aspx
1. Research shall be conducted in such a manner so as to avoid any psychological and physical harm
to humans and animals and financial damage to organizations
2. Only the supervisor and examiners will have access to any data gathered.
3. Participants will remain free to withdraw from the study at any time without having to provide
any reason. In the case of withdrawal, all the records and information collection will be
deleted.
4. The participant, who is the sole proprietor of the data provided, is granting that such data
would be processed for this study purposes only.
5. The data collection process will be a transparent process.
6. All transcriptions and/or electronic recordings reflecting the data collected, once exhausted,
are to be deleted
7. Confidentiality, anonymity and data protection procedures are to be ethically abided by.
8. The researcher would provide a soft copy of the study to the participant, if required.
Enter details here regarding possibility of issues regarding confidential personal data:
N/A
This project does not involve confidential data involving 3rd parties/entities
Enter details here regarding possibility of physical harm:
N/A
No possible harm can occur when this project is being implemented and tested
Enter details here regarding possibility of moral harm:
N/A
This project does not involve issues which could cause moral harm
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 10 of 14
Enter details here regarding possibility of business harm:
N/A
This project does not involve issues which could cause business harm
Anticipated Contributions of the Study.
My project if successful, I believe, has the potential to greatly change to the better how AI machines work
in various disciplines such as healthcare, medicine and transport.
Firstly, AI with MatLab in combination with ultrasonic sensors and Arduino software can be used in
monitoring cows and their health. This can be applicable in large scale cow rearing for commercial
purposes. Using AI facial features, shapes and sizes of cows are detected and recorded overtime such that
any little change in those variables will automatically trigger the tool to recommend health checkup of the
said cow(s).
In addition,This AI with Matlab can be used to in manufacturing of small devices such as mobile phones
for instance it can be used in real time intendances such as detection of images of people photo bombing
while taking selfies. Developing this algorithm can make work easy in the fields like statistics because this
algorithm can be used in giving perfect data concerning a given sample for a given case study which can
use this program of data analysis to make predictions.
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 11 of 14
Dissertation Project Plan.
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 12 of 14
References
[1] H. Zarrinoup, Understanding LTE with Matlab, Toronto: John Wiley and Sons, 2014.
[2] Engineering Development Group, "Mathworks," Mathworks, 11 March 2019. [Online]. Available:
http://www.mathworks.com. [Accessed 11 March 2019].
[3] A. B. Biran, Matlab and Simulink, London: CRC Press, 2010.
[4] R. C. Asher, Ultrasonic Sensors, Michigan: Institute of Physics Pub, 2007.
References
1. https://www.mathworks.com/videous/image-processing-made-easy-81718.html
2. https://mathworks.com/campaigns/offers/ai-with-matlab.html
3. http://medium.com/nanonets/how-to-easily-detect-objects-with-deep-learning
4.
5.
6.
7.
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 13 of 14
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited

18392124412819022887.docx
Document Owner : Head of University College
Page 14 of 14
This section is to be filled in by the representative of the Institute Research Sub-Committee prior to forwarding of
this Form to the ‘MCAST Research Ethics Committee’ for final ethics approval:
Nature of ethical
consideration
Outcome (*) Comments
Consideration of possibility of
issues regarding confidential
personal data:
Consideration of possibility of
physical harm
Consideration of possibility of
moral harm
Consideration of possibility of
business harm
(*) Legend to record outcome by Institute Research Sub Committee:
A – Ethical considerations have been addressed appropriately by Researcher;
B – No (Nil) relevant ethical considerations are applicable under purpose of study as described by Researcher.
C – Ethical consideration have not been addressed appropriately by Researcher;
D – Applicable ethical consideration have not been considered by Researcher.
Details of Representative to the ‘Institute Research Sub-Committee.
Name Signature
Designation Date
__________________________________________________________________________________________________________________
MCAST Controlled and approved document
Unauthorized copying or communication strictly prohibited