Research Paper By Gulak

23 September 2024
Artificial Intelligence

ABDULLAHI MOHAMMED GULAK

Department of Computer Engineering technology, Faculty of Engineering,
Federal Polytechnic Mubi,
Adamawa State Nigeria.
Email: [email protected]
www.fb.com/Drabdallahgulak
+2349023448275

BEING A SEMINR SUBMITTED TO THE DEPARTMENT OF COMPUTER ENGINEERING

TECHNOLOGY, SCHOOL OF ENGINEERING IN PARTIAL FULFILMENT FOR THE
REQUIREMET OF HIGHER NATIONAL DIPLOMA IN COMPUTER ENGINEERING
TECHNOLOGY. FEDERAL POLYTECHNIC MUBI.

1
 Abstract

This seminar serves as an insightful exploration into the critical role that Artificial Intelligence (AI) plays
in the domain of Robotics Engineering, offering a review of the intertwined evolution of these two fields.
It commences by retracing the origins of robotics, from ancient automations to the contemporary era of
advanced, multifunctional robots, elucidating how technology and engineering have converged to birth
these mechanical marvels. Concurrently, the seminar reviews more on a captivating journey through the
annals of AI, tracing its inception from theoretical frameworks to practical applications, underlining the
pivotal moments in its development. It then meticulously dissects how AI breathes life into robots,
endowing them with the capabilities to perceive, learn, adapt, and operate autonomously. This symbiotic
relationship between AI and robotics is underscored through a historical lens, showcasing the evolution of
robots from mechanical curiosities to indispensable tools that permeate industries like manufacturing,
healthcare, and space exploration. Furthermore, the seminar unveils how AI-powered robots have
cemented their position in the vanguard of contemporary technology. It aslso unravels the profound
transformative potential that AI in robotics engineering holds for society, providing invaluable insights
into the intersection of technology, innovation, and human progress.

KEYWORDS: Robotics. Artificial intelligence (AI), Machine Learning, Engineering, Language,

Processing, Robots, Recognition.

2
1.0 Introduction robots to plan and execute tasks efficiently.
Robotics Engineering based on Artificial Significance and the structured approach
Intelligence (AI) involves the integration of AI organizations adopt to harness its potential. In this
technologies, such as machine learning and discourse, the important of such interaction in
computer vision, into the development and between AI and Robots lead to the transformation
operation of robotic systems. Which enables and alternations of industries, ranging from
robots to perceive and interpret their manufacturing to healthcare. As well as by
surroundings, learn from their experiences, make enhancing efficiency, precision, and decision-
autonomous decisions, and interact effectively making.
with humans and their environments. This
1.1 Artificial Intelligence
interdisciplinary field encompasses various
Artificial intelligence is the general name
aspects such as perception, learning, decision-
of the technology for the development of
making, autonomy, and human-robot interaction,
machines, which are created entirely by artificial
with the ultimate goal of creating intelligent and
means and can exhibit behaviours and behave like
adaptive robots capable of performing a wide
human beings, without taking advantage of any
range of tasks in diverse settings, ranging from
living organism. Artificial intelligence products,
manufacturing and healthcare to exploration and
when approached as an idealist, are completely
everyday life. Artificial Intelligence (AI)in
human like and can perform ttasks such as feeling,
today’s modern technology of robotics
foreseeing, and making decisions, they are
Engineering represents the convergence of AI and
generally called robot. The artificial intelligence
robotics, aiming to imbue robotic systems with
where the first steps are being taken by the
advanced cognitive abilities and enhanced
question of Mathison Turing "Can machines be
autonomy which involves leveraging AI
considered?" Is one of the most important factors
techniques such as machine learning, deep
in the emergence of various military weapon
learning, and computer vision to empower robots
technologies and the development of computers
with the capacity to perceive their surroundings
during the period of World War II (Maad, 2015).
accurately. Furthermore, AI in Robotics
Engineering encompasses the aspect of learning, 1.2 Artificial Intelligence History
where robots can acquire new knowledge and
To be informed about the history of
skills autonomously. This learning process can be
artificial intelligence, it is necessary to go back to
supervised, unsupervised, or reinforcement-
previous dates in Milat. In the Ancient Greek era,
based, enabling robots to adapt to changing
it is proven that various ideas about humanoid
conditions and improve their performance over
robots have been carried out. An example of this
time. Decision-making is another critical facet,
is Daedelus, who is said to have ruled the
with AI algorithms used for path planning, motion
mythology of the wind, to try to create artificial
control, and higher-level reasoning, allowing
humans. Modern artificial intelligence begun to

3
be seen in history with the aim of defining The first artificial intelligence applications were
philosophers' system of human thought. 1884 is introduced during this period. These applications
very important for artificial intelligence. Charles were based on logic theorems and chess game.
Babbage, on this date, has worked on a The programs developed during this period were
mechanical machine that will exhibit intelligent distinguished from the geometric forms used in
behaviour. However, as a result of these studies, the intelligence tests; which has led to the idea
he decided that he would not be able to produce a that intelligent computers can be created.
machine that would exhibit intelligent behaviours Milestones for Ai History in 1950, Alan Turing
as human being, and he suspended his work. In created a test to determine whether a machine is
1950, Claude Shannon introduced the idea that intelligent. This test shows the intelligence given
computers could play chess. Work on artificial to computers. The intelligence level of the
intelligence continued slowly until the early machines that passed the test at that time was
1960s. considered adequate. LISP (List Processing
Language), developed by John McCarthy in 1957,
and was a functional programming language
developed for artificial intelligence. One of the
rather old and powerful programming languages,
LISP is a language that allows you to create
flexible programs that represent basic operations
with list structure. 1965 to1970 could be called a
dark period for artificial intelligence.

Figure 1 Brief History of AI

The emergence of artificial intelligence officially Figure 2 Past Era AI

in history dates back to 1956 as shown in Figure The developments on artificial intelligence in this
11. In 1956, a conference artificial intelligence period are too few to be tested. The hasty and
session at Dartmouth College was introduced for optimistic attitude due to the unrealistic
the first time. Marvin Minsky (2016) stated in his expectations that have emerged has led to the idea
book "Stormed Search for Artificial Intelligence that it will be easy to uncover the machines with
“that "the problem of artificial intelligence intelligence. But this period was named as dark
modelling within a generation will be solved ". period for artificial intelligence because there was

4
no success with the idea of creating intelligent human beings except where such orders would
machines by simply uploading data .Between conflict with the First Law; and (3) a robot must
1970 and 1975, artificial intelligence gained protect its own existence as long as such
momentum. (Maad, 2015). protection does not conflict with the First or
Second Laws. Asimov (2013) in his work inspired
2.1. Definition of robotics
generations of scientists in the field of robotics,
According to Micheal (2020) Robotics is defined
AI, and computer science—among others the
as a field of engineering focused on the design and
American cognitive scientist Marvin Minsky
manufacturing of cyborgs, the so-called machine
(who later co-founded the MIT AI laboratory). At
man. They are used to perform tasks for human’s
roughly the same time, but over 3,000 miles away,
convenience or something too difficult or
the English mathematician Alan Turing worked
dangerous for human to perform and can operate
on much less fictional issues and developed a
without stopping such as in assembly lines. Figure
code breaking machine called The Bombe for the
3 shows a labelled Robot.
British government, with the purpose of
deciphering the Enigma code used by the German
army in the Second World War. The Bombe,
which was about 7 by 6 by 2 feet large and had a
weight of about a ton, is generally considered the
first working electro-mechanical computer. The
powerful way in which The Bombe was able to
break the Enigma code, a task previously
impossible to even the best human
mathematicians, made Turing wonder about the
Figure 3 Defined Label of Robot
intelligence of such machines. In 1950, he
published his semina article “Computing
2.2 Brief History of Robotics based on AI Machinery and Intelligence” where he described
how to create intelligent machines and in
The roots of AI can probably be traced back to the
particular how to test their intelligence. This
1940s, specifically 1942, when the American
Turing Test is still considered today as a
Science Fiction writer Isaac Asimov published his
benchmark to identify intelligence of an artificial
short story Runaround. The plot of Runaround—
system: if a human is interacting with another
a story about a robot developed by the engineers
human and a machine and unable to distinguish
Gregory Powell and Mike Donavan—revolves
the machine from the human, then the machine is
around the Three Laws of Robotics: (1) a robot
said to be intelligent (Michael & Andreas, 2019).
may not injure a human being or, through
inaction, allow a human being to come to harm;
(2) a robot must obey the orders given to it by

5
2.3 How AI Makes Robotic Play a Vital Roles

The Robots learn from machine learning and

artificial intelligent platform which is given and
there is much concern about these robots that
these machines will replace the humans and
humans will be washed-out from their jobs. New Figure 5 Technological Weapons
opportunities can be made to increase the An Artificial intelligent Robot namely “Daksh”
productivity by different companies using developed by “Defense Research and
Development Organization (DRDO) and its
Artificial intelligence in Robots (DatabasTown, functionality is to remove the life threatening
2023). These Robots can make work safer and objects in danger zone.
valuable time can be saved through these
technological applications. Some main fields
where the Robots are playing more vital role are:

Industries: Robots are used in industries as an

important individual organ as shown in Figure 4,
they plays a vital role in productivity to perform Figure 6 DRDO Military Robot
an assigned task and whole process is being Medicine: There was an ample need of artificial
managed by these artificial intelligence machines. intelligent technology in medical field. Numbers
of clinical tests are being carried out through the
mechanical artificial intelligence Robots.

Figure 4 Industrial Robotics

Military: The World has entered in to

technological zone. During war, there are number Figure 7 Robot in Medicine Field
of technological weapons as shown in Figures 5 Many complicated operations like brain tumor
and 6 which are used by developed countries for removal etc. are being performed by the modern
their defence. Robots are very important in this artificial intelligence Robots which is part of the
regards for the development of artificial advancement in medical fields.
intelligence weapons having capability to remove
2.5 Role of AI in working of Robots
life threatening objects in war field.

6
Mostly Robots are used commercially in different learning for both human and machine learning. In
fields, where these machines are designed to machine learning it requires that knowledge is
perform specific automated tasks and functions. shown without being changed in the system and
Robots use number of different applications no further process is desired (Firebaugh, 2017).
according to the nature of the task. A robot
consists of moving parts and sensors but the
artificial intelligence helps these parts and sensors
in processing program environment through AI
which enables a robot to understand and do things
on its own (DatabasTown, 2023).

Sense and respond: Robots have the ability to

sense and respond by using Artificial intelligence Figure 9 AI as Machine Learner
technology. The functions are pre-programmed in It is obvious that rote learning is a comprehensive
Robots to identify objects. There are sensors, method for artificial intelligence in robotics as it
which are installed in Robots to detect rapidly and involves basic steps of learning and it can be
perform the task as programmed. applied easily. Additionally, decision tree
learning is an extensive method that is widely
used as deductive learning. In many scenarios
such as the cases where model cognoscibility is
needed one of the most applied method is decision
tree learning algorithm that is functional for
Figure 8 Sense and respond of AI
knowledge finding and data mining (Barros,
(DatabasTown, 2023). 2015). Decision tree algorithms have significant

3.1 The Basic Role of AI in Robotics role in robotics because they provide model
comprehensibility and can be used for both
The basic application of AI in robotic is
knowledge discovery and data mining. All in all it
categorised based on machine learning, pattern
can be safely said that deductive learning plays an
recognition and knowledge representation.
essential role in strengthening one of the most
3.2 Roles of AI in Robotics Machine vital application of artificial intelligence, machine
Learning learning, with the powerful techniques it has.

The first significant application of

artificial intelligence is machine learning. The
main approach for machine learning is deductive
learning. Initially, the basic technique of
deductive learning is rote learning. As Firebaugh
states, the basic level of learning known as rote
7
 3.3 Roles of AI in Enabling Robot
Process Pattern Recognition

The second remarkable application of

artificial intelligence in robotics is in the process
of pattern recognition. Firstly, speech recognition
is a nonignorable way of identifying patterns in
Figure 10 AI as Inductive learner
data in robotics.
Inductive learning is the other important approach
for machine learning. For instance, active learning
is an advanced learning system that is effectively
applied as inductive learning as shown above in
Figure 11. Settles stated that, a machine learning
algorithm that works with higher efficiency with
less guidance labels when it is free to select the
data from the source is the main concept of active
learning (Settles, 2010). Active learning is an Figure 11 Pattern Recognition Process
impressive and beneficial technique that can be
First of all, hidden markov model is essential for
used in numerous scenarios with less effort.
providing auditory figure in robotics. Blunsom
Furthermore, rule-based learning is another
stated that, hidden markov model which is a
powerful inductive learning process. Rule- based
capable statistical gizmo, is used for modeling
learning algorithms are well-performing and easy
generative sequence. These sequences is
to compact machine learning algorithms that
identified by an underlying process producing an
approximate the values of a real valued functions
observable sequence (Blunsom, 2016). Lee and
from multiple input variables using ordered
Hon stated that, statistical learning is widely used
samples of decision rules (Weiss & Indurkhya,
method in speech recognition and maximum
2014). It is possible to compact rule-based
performance may be gained by using a learning
learning algorithms to every system with high
method which is hidden markov model. Powerful
performance by defining correct decision rule
and curt modeling of speech can be made by using
samples so they can be considered as the most
hidden markov models (Lee & Hon 2020).
useful learning algorithms. It can be concluded
Processor of some machines senses the speech
that inductive learning has very critical role for
sources like human's sense organ perceives,
machine learning as it has many advanced
artificial intelligence can be observed in this
methods to reinforce in different circumstances.
hearing and construing event. In addition to
hidden markov model, dynamic time warping is
an important method for obtaining audial pattern.
Müller stated that, dynamic time warping which

8
is used to find optimum linage between two given statistical independent basis with all potential. So,
sequences in particular boundary, is known as an independent component analysis may be called a
important technique, and warping the sequences general version of principal component analysis.
is used to provide pair to each other. Basically, While principal component analysis focuses on to
distinct speech patterns in autonomy speech obtain representation of the inputs based on
recognition have been compared by using uncorrelated variables, independent component
dynamic time warping (Müller, 2007). Clearly, analysis seeks to obtain representation of ones
dynamic time warping is identifying the speech based on statistical independent variables (Déniz,
patterns by comparing them with one in data & Hernandez, 2022).
which is known by system, they will be
3.4 Roles of Ai In Process Of Knowledge
alternative ears for deaf. If a clever robot is
Representation
wanted, it must hear its environment and
The third remarkable application of artificial
response, therefore speech recognition is one of
intelligence in robotics is in the process of
the indispensable technique of artificial
knowledge representation.
intelligence.

Another way of identifying patterns in data in

robotics is visual recognition. Firstly, principal
component analysis is a significant way to
identify patterns in data. Smith (2014) stated that,
patterns in data may be identified by using
principal component analysis which shows
resemblances and dissimilarities between
patterns. It is difficult that patterns are found in
data of high dimension. The number of dimension
may be reduced by using principal component
analysis without much loss of information. So
principal component analysis makes easier the
finding patterns in data (Smith, 2014). This
technique compares patterns with patterns which Figure 12 AI in Knowledge Representation

are known by processor. Then, it sorts and To begin with, declarative knowledge
matches them one by one for recognition. representation is an essential process in robotics.
Additionally, independent component analysis is Primarily, logic representation is the foremost of
a technique to define patterns in data. Déniz declarative knowledge representation as it is of
&Hernandez (2022) stated that, a linear transform high preference in robotics. First Order Logic
for the input data is found by independent (FOL) is known as the most frequently reviewed
component analysis technique with using a and applied logic. It is comprised of about twelve

9
primary symbols with reference to the selected pattern of the signs. Procedural indicates the
predicates, profoundly articulated knowledge entrance scheme which further describes how
representations are formed with these elements. questions are formed, modified and replied, while
The illustrative and decision axiom techniques are semantic creates the method of associating the
the basic reasoning types in First Order Logic meaning (Sharma & Kelkar, 2015). Obviously,
(Trentelman, 2021). Clearly, the high preference associative networks are very efficient in
of logic representation can be attributed to its illustrating complex data and knowledge in a
simplicity and high expressivity. Another clear, precise and well-structured manner. Also,
important declarative knowledge representation frames is another significant paradigm of
technique is production rules. Trentelman procedural knowledge representation. As
explains that it is optimal for illustrating Tanimoto explains, a frame is known as a set of
procedural knowledge, he further states that knowledge attributed to a single entity, scenario
production rule is an intelligent process that or theory. Each frame is usually comprised of
applies affirmation and a specific order for several pieces of data and there are many frames
knowledge representation. The order which is in a knowledge base. In the process of problem
called production is comprised of the precedent solving, some frames are formed and annihilated,
set of conditions and subsequent set of actions. It while others remain everlasting. A frame gives the
is expressed in the form “if (conditions)” and illustration for an entity, scenario or rank in the
“then (actions)” (hath, 2019). Apparently this form of a collection of trait nomenclatures and
technique follows an explicit rule of satisfied values for the traits. A Frame is made up of two
conditions and actions. Therefore, it is very entities, a nomenclature and a list of trait value
obvious that one of the most important techniques pairs. The traits may be known as “slot
applied in illustrating knowledge in robotics is nomenclatures” and the values as “fillers”.
declarative knowledge representation. Therefore, a frame is an assigned collection of
slots and the fillers connected with the slots
According to Sharma and Kelkar (2015),
(Tanimoto, 2018). Evidently, frames as a
semantic networks are used as substitute to
procedural knowledge representation technique
predicate logic as a means of knowledge
simplify and exemplify illustrated knowledge. As
representation. The knowledge can be saved in a
a result, it can be concluded that procedural
chart pattern with junctions depicting objects in
knowledge representation technique is an
the world and bows illustrating links between
inevitable process in the science of robotics.
objects. Semantic network is also referred to as
Associative Network. Semantic representation is In an article obtained from Onlindegrees
made up of four entities which include lexical, (2023) which states how AI is used in Robotics
structural, procedural and semantic. Lexical explained that; the combination of AI and
describes signs that can be used in illustrations robotics holds tremendous potential, leading to
language. Structural depicts restrictions on the increased productivity and efficiency, improved

10
safety and greater flexibility for workers in a pattern recognition, speech and visual recognition
variety of professions. One of the key ways in are very important in the programming of robots.
which AI is used in robotics is through machine Finally in knowledge representation, declarative
learning (onlinedegres, 2023). This technique and procedural knowledge representations are
enables robots to learn and perform specific tasks very key processes in robotics. The advancement
through observing and mimicking human actions. of artificial intelligence and its applications have
AI gives robots a computer vision that enables gone a long way in transforming robotics. It has
them to navigate, detect and determine their further led to the creation of more intelligent and
reactions accordingly. This helps them go beyond task efficient robots which have various uses. The
simply performing repetitive tasks to become true application of artificial intelligence will pave way
“cognitive collaborators.” Another way that AI is for more outstanding breakthroughs in robotics
used in robotics is through edge computing. This and additionally create robots that will reflect the
approach provides machines with real-time processing capacity of the human brain. The
awareness, enabling robots to act on decisions at importance of artificial intelligence cannot be
a rate much quicker than human capabilities overemphasized and as such it should be taught as
allow. part of computer science programs in high schools
and emphasis should be given to it in universities,
AI also helps robots learn to perform specific
further researches should be encouraged through
tasks through the use of various sensors, which
government and corporate funding.
may include:
Reference
• Time-of-flight optical sensors,
Temperature and humidity sensors, Barros, R.C. (2015). A bottom-up oblique
Ultrasonic sensors, Vibration sensors, decision tree induction algorithm. Paper
Millimeter-wave sensors Presented at the Intelligent Systems
These sensors help robots to learn and adapt, Design and Applications 2015 11th
International Conference on, Cordoba,
making them more intelligent and better equipped
Spain. Retrieved from
to act and react in different scenarios. These are http://ieeexplore.ieee.org/xpl/articleDetail
just a few of the ways that artificial intelligence is s.jsp?arnumber=6121697
used in conjunction with robotics (onlinedegree, Blunsom, P. (2016).”Hidden markov models”.
2022). Retrieved from Iowa State University web
site: digital.cs.usu.edu/~cyan/
4.0 Conclusion CS7960/hmm-tutorial.pdf

In conclusion, application of artificial intelligence DataBaseTownc (2023). Artificial Intelligence.

Application of AI in Robotics. Role of AI
is quite prevalent and essential in the various
in Working of Robots. Retrieved From
stages of robot design. Firstly, in the process of https://databasetown.com/application-of-
machine learning, artificial intelligence plays a ai-in-robotics/
very vital role in the deductive and inductive
learning process of robot design. Secondly,

11
Firebaugh, M.W. (2017). Artificial intelligence: Smith, L.I. (2002).A tutorial on principal
A knowledge-based approach. Boston: components analysis. Retrieved from
PWS University of Otago web site:
https://onlinedegrees.sandiego.edu/ Tanimoto, S.L (2018). The elements of artificial
intelligence: An introduction using LİSP.
Lee, K. & Hon, H. (2020).Speaker-independent
Maryland: Computer Science Press.
phone recognition using hidden markov
models. Abstract retrieved from Trentelman, K (2021) Survey of knowledge
http://ieeexplore.ieee.org/stamp/stamp.jsp representation and reasoning systems.
?tp=&arnumber=46546 Edinburgh, South Australia: Defence
Science and Technology Organisation
Maad M. Mijwil. History of Artificial
(DSTO).
Intelligence. 2015. DOI:
10.13140/RG.2.2.16418.15046. n at: Weiss, S. M. & Indurkhya, N. (2014). Ruled-
https://www.researchgate.net/publication/ based machine learning methods for
322234922 functional prediction. Journal of Artificial
Intelligence Research, 3, 383-403,
Michael Haenlein1 and Andreas Kaplan2. 2019.
doi:10.1613/jair.199
A Brief History of Artificial Intelligence:
On the Past, Present, and Future of www.cs.otago.ac.nz/cosc453/student_tutorials/pr
Artificial Intelligence. DOI: incipal_components.pdfDéniz, O.,
10.1177/0008125619864925 Castrillon, M. & Hernandez, M.
(2022).Face recognition using
Müller,M.(2007).Information retrieval for music
independent component analysis and
and motion. Retrieved from
support vector machines. Pattern
link.springer.com/chapter/10.1007/978-3-
recognition letters. Retrieved from
540-74048-3_4
http://www.sciencedirect.com/science/art
OnlideDegree (2023). Introduction to AI icle/pii/S0167865503000813#
Applications in Robotics. Retrieved from
https://onlinedegrees.sandiego.edu/applicatio
n-of-ai-in-robotics/

Settles, B. (2010). Active learning literature

survey (Technical Report No: 1648).
Retrieved From
http://csis.bitsilani.ac.in/faculty/goel/cour
se_material/Machine%20Learning/2013/
Reading%20Material/settles.activelearnin
g.pdf
Sharma, T., & Kelkar, D. (2015). A tour towards
knowledge representation technique.
International Journal of Computer
Technology and Electronics Engineering,
2(2), 131-135. Retrieved from
http://www.ijctee.org./files/VOLUME2IS
SUE2/IJCTEE_0412_22.pdf

12