Chapter 2

ARTIFICIAL INTELLIGENCE FOR IOT

Machine to Machine Communication in IoT

M2M/IoT devices are wireless networks that connect to the internet without human interaction.
In the era of the Internet revolution, they are the next generation of devices connected to the
Internet. These devices are also known as "active communication devices." Using sensors and
communication modules, M2M devices are able to send data wirelessly or via wired networks.

M2M connectivity provides quick, secure internet access from any location at any speed. It
functions both inside and outside. Numerous M2M gateways connect these devices to the
network domain. PAN (Personal Area Network) and LAN technologies are just two examples
of M2M area networks. IoT & M2M, collaboration offer extensive storage, resources, and
security. The following are only a few M2M communication characteristics:

• Numerous brief sessions

• Small payload
• It has a wide range of features (such as battery capacity) and
• A wide range of QoS requirements (e.g., delay).

Machine-to-Machine communication, or M2M, is the transmission of machine data between

mechanical or electronic devices using wired or wireless tools.

Depending on the reason for the transmission, it may be as simple as measurement or as

complex as if/then instructions. The exchange of data is done automatically without human
intervention.

M2M communication can occur through many formats and access methods, including serial
ports, traditional PLCs, or wireless connections through cellular, Bluetooth, Wi-Fi, RFID, or
fiber. These transmission formats allow data from sensors, tools, and other devices to be
transmitted to another device or an edge computing device to report status, log a transaction,
or give or request instructions for the device itself.

1|Page
 Fig 2.1. A basic diagram of an M2M network

In general, these two technologies differ in the following ways:

Basis M2M IoT

Connection Type Simple device-to-device IP networks are used by devices
Used communication is often done at the to communicate.
client site with embedded software.
Communication Direct interaction between machines Automation of IoT sensors.
Used both conventional protocols and modern Internet protocols include
Communication communication methods. Telnet, HTTP, and FTP.
Protocol
Intelligence Observation of intelligence to some Objects are responsible for
extent. decision-making.
Technology Hardware-based. hardware and software-based.
Data Sharing exchange of data exclusively between Data sharing among different
communicating parties. applications to enhance the user
experience.
Scope deployed in a closed system. It connects to a larger network.
Open API Support Open APIs are not supported. supports open API integration.
Internet Internet connectivity is not required for It is necessary for you to have
devices. an active internet connection.

2|Page
IoT reference model

Fig. 2.2 IoT reference model

Device or Sensing Layer:

The device/ sensing layer is the first layer of the IoT architecture and is responsible for
collecting data from different sources. This layer includes sensors and actuators that are placed
in the environment to gather information about temperature, humidity, light, sound, and other
physical parameters. These devices are connected to the network layer through wired or
wireless communication protocols.

Network Layer:

The network layer of an IoT architecture is responsible for providing communication and
connectivity between devices in the IoT system. It includes protocols and technologies that
enable devices to connect and communicate with each other and with the wider internet.
Examples of network technologies that are commonly used in IoT include WiFi, Bluetooth,

3|Page
Zigbee, and cellular networks such as 4G and 5G. Additionally, the network layer may include
gateways and routers that act as intermediaries between devices and the wider internet, and
may also include security features such as encryption and authentication to protect against
unauthorized access.

Data processing Layer/ Service:

The data processing layer/ Service layer of IoT architecture refers to the software and hardware
components that are responsible for collecting, analysing, and interpreting data from IoT
devices. This layer is responsible for receiving raw data from the devices, processing it, and
making it available for further analysis or action. The data processing layer includes a variety
of technologies and tools, such as data management systems, analytics platforms, and machine
learning algorithms. These tools are used to extract meaningful insights from the data and make
decisions based on that data. Example of a technology used in the data processing layer is a
data lake, which is a centralized repository for storing raw data from IoT devices.

Application Layer:

The application layer of IoT architecture is the topmost layer that interacts directly with the
end-user. It is responsible for providing user-friendly interfaces and functionalities that enable
users to access and control IoT devices. This layer includes various software and applications
such as mobile apps, web portals, and other user interfaces that are designed to interact with
the underlying IoT infrastructure. It also includes middleware services that allow different IoT
devices and systems to communicate and share data seamlessly. The application layer also
includes analytics and processing capabilities that allow data to be analysed and transformed
into meaningful insights. This can include machine learning algorithms, data visualization
tools, and other advanced analytics capabilities.

Security Layer:

The Security Model for IoT consists of communication security that focuses mostly on the
confidentiality and integrity protection of interacting entities and functional components such
as Identity Management, Authentication, Authorisation, and Trust & Reputation.

4|Page
Management layer:

The Management FG includes the necessary functions for enabling fault and performance
monitoring of the system, configuration for enabling the system to be flexible to changing User
demands, and accounting for enabling subsequent billing for the usage of the system. Support
functions such as management of ownership, administrative domain, rules and rights of
functional components, and information stores are also included in the Management FG.

Horizontal and vertical application in IoT

With the advancement of IoT technologies, two business models have emerged: vertical and
horizontal.
The vertical market is closely focused on meeting the needs of one specific industry that is
provided and controlled by the same company. The horizontal market can sell its products and
services in more than one industry and therefore targets a wider range of business segments.
Vertical application in IoT
Smart Cities

Smart city solutions examples such as smart parking, noise mapping, smart lighting control,
etc. will be extremely useful. These applications are under development, but their
implementation is inevitable, because the prospects for public safety, air pollution monitoring,
and optimization of city traffic, especially in metropolitan areas, are enormous. Therefore the
Smart City IoT applications are a fast-growing and very promising industry.

Intelligent buildings

IoT applications allow for the control of Intelligent Buildings and Home systems: facility-
automation and monitoring for building systems (HVAC, lighting, elevators, smoke alarms,
fire extinguishers), building utilization, and security (room use, access, surveillance). This
enables rational resource usage and saves valuable resources for the future.
Smart Homes
Today, IoT applications are helping us make our life more comfortable and safer. There are
many products built into the Smart Home system. Among them: smart thermostats and air

5|Page
conditioners, speakers, smart sockets, smoke and motion detectors, smart intercom system, pet
feeders with a timer, light bulbs that turn on at dusk, etc. IoT Smart Home technologies are one
of the most fast-growing segments and are expected to have around 15.12% CAGR over the
forecast period 2019-2026.

The global internet of medical things (IoMT) market is expected to reach a $158 billion
valuation in 2022. IoT use cases in medicine are changing the treatment of patients, tracking
their condition and collecting vital information. Such solutions provide data on the current state
of the patient, monitor their changes, and automatically warn of potential problems before they
become critical. Data analysis using devices is aimed at improving the diagnosis of patients'
conditions. Constant monitoring of patients allows doctors to have a holistic and relevant
picture of a patient’s wellbeing.

Consumer Goods
Technological solutions in everyday goods allow for continuous monitoring of the main
indicators and they make life easier and more pleasant. "Smart" assistants in the form
of wearable devices on the hand have gained immense popularity. Smart bracelets are full-
fledged gadgets for health monitoring that also count steps, analyze activity, and provide health
reports to the user. Moreover, IoT use cases in consumer goods can help to care for children or
older persons, or pets, by giving 24/7monitoring, care, and communication.

Industrial
Among the best examples of IoT applications in the industry are all kinds of sensors, software
systems, and big data analytics for operations management and control in the shop-floor or
remote production processes. Smart machines improve productivity and correct common
human errors, especially those related to quality control and sustainability.

Automotive
Already, we can use smart cars equipped with countless sensors and an internet connection to
guarantee us maximum comfort. Other IoT use cases include remote control of doors and
garage temperatures. Moreover, smart cars can independently assess the situation on a road and
correct a car’s route without human intervention. They can also be used to update maps and
assess traffic situations.

6|Page
Agriculture
IoT technology in the form of sensors records various indicators of soil conditions: moisture,
plant nutrition, temperature, etc. Now there is no need to go around entire fields to control
crops as the work can be done with smart systems and products. IoT use cases in agriculture
include precision farming, livestock monitoring, irrigation control and automated drones for
farm surveys, field mapping, crop spraying, and more.

Supply chains
Top IoT applications track and shape supply chains in real time to help businesses make
informed decisions and prevent shortages. Accurately tracking the availability of goods at a
remote warehouse, based on the order or current demand, IoT applications automatically
generate the best delivery or exchange operation scheme that is convenient for all interested
parties. At the same time, the minimum need for labor is achieved and, as a result, the best
economic effect of logistics in general.

Energy systems
The energy sector, which will grow by 40% in the next 25 years, requires a particularly
painstaking approach to the design and organization of electrical networks. Thus, by collecting
data on the time and nature of consumption, IoT applications analyze the processes of
electricity circulation between suppliers and consumers, automatically optimize the use of
resources, and improve their delivery parameters.

Horizontal IoT applications

A horizontal application is any software application that connects numerous users with
different knowledge and skill sets by allowing multiple providers to work with a common
framework. They can be off-the-shelf solutions or integrated and proprietary vendor solutions.
The horizontal architecture makes it possible to combine and match hardware devices from
many producers with a single application. It enables open connectivity, improved privacy and
security, scalability and flexibility, higher quality, and lower cost.
The horizontal enablers comprise all those technology and service elements required to make
a smart connected system. Every IoT application requires all these enablers:

7|Page
Sensors and devices
They gather information about the physical world, and can even interact with it. Raw data such
as temperature and pressure are recorded and then sent to a gateway or any other device for
local analysis or further processing. These can be various thermometers, microphones,
cameras, and dozens of other less common devices.

IoT gateways and communication protocols

These are devices that are usually assigned the logic of superficial analysis of information
coming from sensors connected to them. In certain situations, data analysis may require a few
computing resources, so that the gates are quite capable of making some decisions on their
own.

Security Infrastructure
Security for operational efficiency and protection of physical and digital assets is a critical
element that needs to be built into each component of a system, including the devices, apps,
gateways, hubs, communications, servers, and cloud platforms. With the necessary tools, you
can monitor, detect and defend against attacks on informational and operational resources.

8|Page
IoT Platforms
The IoT platform is a complex software environment. It collects information from sensors,
devices, networks, and software and links them all together. Together, they provide valuable
and effective data. Horizontal platforms enable companies to select best-in-class products
tailored for their work needs in a heterogeneous environment.

Data acquisition and analytics

Big data, predictive analytics, stream processing, and other methods are used to design “smart”
IoT devices.

IT infrastructure

All previous elements must be supported by a cloud, hosted, or hybrid IT infrastructure that
integrates into legacy IT and business systems.

Big data and IoT

Today, the Internet of Things, commonly referred to simply as IoT, and Big Data is the hot
terms used on a daily basis in Information Technology circles. It’s almost difficult to talk
about one without mentioning the other. Despite their intimate connection, they are not
interrelated. Both are the future of data and by data, we mean massive volumes of data. We
live in a digital era where new objects are connected to the Internet with the aim of improving
people’s lives.

Internet of Things, or IoT, refers to a global network of internet-connected devices or

machines able to collect and exchange data. In simple terms, IoT simply is devices that collect
data and send it to the Internet. IoT is a large heterogeneous collection of things, which differ
from each other. IoT aims to integrate and collect information from and offer services to a
diverse spectrum of physical things used in different domains.
Every day trillions and trillion bytes of data are generated by billions of devices that are
connected to the Internet, so much that 90 percent of the data in the world has been generated
in the past couple of years. This massive volume of data comes from various sources; from
posts on social media sites, digital pictures and videos, transaction records, payment histories,
sensors used to gather climate information, or GPS signals, to name just a few. Collectively,

9|Page
this data is called Big Data. Analyzing and processing such a massive volume of both
structured and unstructured data is almost impossible using traditional software techniques.

SR.NO. IOT BIG DATA

Big Data is described as large sets of

IoT is a global system of interrelated
data generated from a variety of
1 computing devices that are able to sense,
sources that are so large to process
collect, and exchange data over the Internet.
using traditional techniques.

The concept is to provide interconnection

The concept is to find insights in new
between devices to create a smart
and emerging types of data and
2 environment thereby making machines
content that lead to better decisions
smart enough to bypass human
and strategic business moves.
intermediaries.

IoT collects, analyzes, and processes data The data streams are not subjected to
streams in real-time without any delay to processing real-time and there is a
3
make control decisions in an effective delay between when the data is
manner. collected and when it is processed.

IoT involves analyzing machine-generated Big Data deals with human-generated

4 data such as sensors in home appliances and data such as social media usage,
so on. photos, and videos, etc

Big data is more into collecting and

IoT is about simultaneously collecting and
5 accumulating huge data for analysis
processing data to make real-time decisions.
afterward.

With big data, you can analyze all the

Using IoT you can track and monitor assets
6 information you have about failures
like trucks, engines, HVAC systems, and
and start to uncover the root causes.

10 | P a g e
SR.NO. IOT BIG DATA

pumps. You can correct problems as you

detect them.

Introduction TensorFlow
TensorFlow is a software library or framework, designed by the Google team to implement
machine learning and deep learning concepts in the easiest manner. It combines the computational
algebra of optimization techniques for easy calculation of many mathematical expressions.
Let us now consider the following important features of TensorFlow:
• It includes a feature of that defines, optimizes and calculates mathematical expressions
easily with the help of multi-dimensional arrays called tensors.
• It includes a programming support of deep neural networks and machine learning
techniques.
• It includes a high scalable feature of computation with various data sets.
• TensorFlow uses GPU computing, automating management.
It also includes a unique feature of optimization of same memory and the data used.
Why is TensorFlow So Popular?
TensorFlow is well-documented and includes plenty of machine learning libraries. It offers a few
important functionalities and methods for the same. TensorFlow is also called a “Google” product.
It includes a variety of machine learning and deep learning algorithms. TensorFlow can train and
run deep neural networks for TensorFlow handwritten digit classification, image recognition,
word embedding and creation of various sequence models.

Keras – Introduction

Deep learning is one of the major subfields of machine learning framework. Machine learning
is the study of design of algorithms, inspired from the model of human brain. Deep learning is
becoming more popular in data science fields like robotics, artificial intelligence (AI), audio &
video recognition and image recognition. Artificial neural network is the core of deep learning
methodologies. Deep learning is supported by various libraries such as Theano, TensorFlow,
Caffe, Mxnet etc., Keras is one of the most powerful and easy to use python library, which is

11 | P a g e
built on top of popular deep learning libraries like TensorFlow, Theano, etc., for creating deep
learning models.

Overview of Keras

Keras runs on top of open-source machine libraries like TensorFlow, Theano or Cognitive
Toolkit (CNTK). Theano is a python library used for fast numerical computation tasks.
TensorFlow is the most famous symbolic math library used for creating neural networks and
deep learning models. TensorFlow is very flexible and the primary benefit is distributed
computing. CNTK is deep learning framework developed by Microsoft. It uses libraries such
as Python, C#, C++ or standalone machine learning toolkits. Theano and TensorFlow are very
powerful libraries but difficult to understand for creating neural networks.

Keras is based on minimal structure that provides a clean and easy way to create deep learning
models based on TensorFlow or Theano. Keras is designed to quickly define deep learning
models. Well, Keras is an optimal choice for deep learning applications.

Features

Keras leverages various optimization techniques to make high level neural network API easier
and more performant. It supports the following features −

• Consistent, simple and extensible API.

• Minimal structure - easy to achieve the result without any frills.
• It supports multiple platforms and backends.
• It is user friendly framework which runs on both CPU and GPU.
• Highly scalability of computation.
Benefits

Keras is highly powerful and dynamic framework and comes up with the following advantages

• Larger community support.

• Easy to test.
• Keras neural networks are written in Python which makes things simpler.
• Keras supports both convolution and recurrent networks.
• Deep learning models are discrete components, so that, you can combine into
many ways.

12 | P a g e