176 Internet of Things

Module - V: IoT Analytics

Notes

e
Learning Objectives

in
At the end of this module, you will be able to:
●● Know about the applications of IoT

nl
●● Infer the smart health monitoring
●● Analyse the wearable IoT devices
●● Learn about the home automation

O
●● Examine the futures of automation using IoT
●● Discuss the industrial and commercial networks
●● State the common machine learning models for IoT

ity
●● Explain exploratory data analysis and predictive analysis
●● Summarise big data platforms - Hadoop, Storm and Spark
●● Evaluate usage of databases

Introduction

rs
ve
ni
U
ity

[Link]

The development and rise of the Internet of Things (IoT) promises to transform the
m

way humans live their lives by connecting all possible gadgets to a common network,
the Internet. Sensors, home appliances, health monitoring devices and any device that
generates measurable data are examples of these gadgets. This approach has given
rise to an infinite number of application fields that leverage interactions between humans
)A

and gadgets to draw conclusions and make predictions for increasing quality of life
and optimising the use of available resources. According to a recent McKinsey poll, the
number of installed IoT devices is predicted to exceed one trillion by 2025. Furthermore,
the economic influence of this technology on the global economy can be gauged by the
fact that it is predicted to take 11% of the market by the end of 2025. Having said that,
(c

the Internet of Things, as a technical paradigm, has a number of problems that must be
overcome before its vision can be realised.

Amity Directorate of Distance & Online Education

 Internet of Things 177

Identifying and developing architectures that can match the scalability requirements
Notes

e
of the IoT ecosystem is one of the obvious issues. In addition to several others, data
and resource management, as well as the requirement to preserve system security and
privacy, remain important. All gadgets that fit under the Internet of Things umbrella are

in
usually referred to as smart devices. As a result, the Internet of Things idea collects
and integrates data from various smart devices. This data is also saved and analysed
in order to create helpful insights. The Internet of Things envisions the world as a

nl
network of interconnected elements. The most obvious application of such a concept
is smart city, where the city’s connected nature is supposed to simplify resource
management. Healthcare and infrastructure management, among other IoT applications,

O
are frontrunners because to their need for real-time solutions in crucial scenarios.
Furthermore, there are numerous applications that can be created and commercialised
for human usage.

ity
5.1 IoT: Applications of IoT
The Internet of Things (IoT) refers to the operation of physical devices over the
internet, encompassing objects equipped with sensors, electronics, software and network
connectivity that enable them to gather and share data. Kelvin Ashton introduced

rs
the concept in 1998. IoT plays a pivotal role in establishing enhanced connectivity
among devices, systems and services, surpassing mere machine-to-machine (M2M)
communication. It finds applicability in various domains, including heart monitoring
implants, biochip transponders for livestock, smart buoys in coastal waters, sensor-
ve
equipped vehicles, devices for environmental/food/pathogen monitoring and tools aiding
firefighters in search and rescue operations.
In an evolving world where comfort is highly valued, IoT is especially embraced
for its applications in home automation, also referred to as smart home devices. This
ni

involves the control and automation of lighting, heating (such as smart thermostats),
ventilation, air conditioning (HVAC) systems and household appliances like washers,
dryers, robotic vacuums, air purifiers, ovens and refrigerators/freezers that can be
U

monitored remotely via Wi-Fi. In the future, a vast multitude of physical devices,
embedded with intelligent sensors, will seamlessly interact with each other on a Machine-
to-Machine basis, operating without the need for human intervention.
ity

5.1.1 Introduction
With the advent of IoT technology, one can communicate with his belongings such
as his home, car, fridge, AC, hoover cleaner, water pump and so on, from anywhere
in the globe. It increases the safety and security of people and their property. Future
m

cars are being developed to provide both drivers and passengers with a high level of
enjoyment during their journey. Patients can be regularly checked by their doctors
while performing their routine duties. Accidents caused by human mistake can be fully
prevented. With the above debate, it is determined that the implementation of IoT will
)A

undoubtedly set new standards for society worldwide.

A. Transportation
The Internet of Things (IoT) significantly facilitates communication, control and
information processing across various transportation modes. It extends to all facets of
(c

transportation systems, including vehicles and drivers. This dynamic interaction leads
to capabilities such as inter-vehicle communication, intelligent traffic management,
efficient parking solutions, electronic and logistic fleet management, vehicle control
Amity Directorate of Distance & Online Education
 178 Internet of Things

enhancements and safety measures and road assistance. In the automotive landscape,
Notes

e
vehicles are equipped with interconnected sensors linked to the internet, strategically
monitoring vehicle surroundings and performance. IoT technology plays a pivotal role
in ensuring road safety through systems like collision detection, lane change alerts,

in
smart traffic signal control and intelligent traffic scheduling protocols.
B. Environmental Monitoring

nl
IoT applications in environmental monitoring frequently employ sensors to contribute
to environmental protection by monitoring atmospheric conditions and various
ecological aspects. These applications can encompass activities like tracking the
migratory patterns of species and monitoring changes in their habitats. Through the

O
use of physical devices connected to the Internet, these systems can also function
as warning mechanisms, aiding emergency services in providing more efficient and
timely assistance in critical situations.

ity
C. Infrastructure Management
An essential application of IoT involves monitoring and managing the operations of
rural facilities such as bridges and railway tracks. This application holds significant
importance. The infrastructure of the Internet of Things can be harnessed to observe
any events or changes in the structural conditions of these facilities that might pose

rs
safety risks and threats. Furthermore, it can facilitate efficient scheduling of repair
and maintenance tasks by coordinating responsibilities among various service
providers and facility users. IoT devices can also take charge of controlling crucial
ve
infrastructure elements, like bridges that facilitate ship docking. The integration of IoT
devices for monitoring and controlling infrastructure is anticipated to enhance overall
management, emergency response coordination, service quality, operational uptime
and cost-effectiveness across all domains associated with infrastructure management.
ni

D. Manufacturing
The Internet of Things facilitates the rapid development of new goods as well
as real-time optimisation of manufacturing production and supply by connecting
U

machines, sensors and control systems. The Internet of Things (IoT) assists digital
control systems in automating processes and optimising plant safety and security.
A vast number of networked sensors enable measurements, automated controls,
plant optimisation, health and safety monitoring and other activities. The National
ity

Science Foundation developed a collaborative industry/university research centre on

intelligent maintenance systems (IMS). The goal is to achieve near-zero breakdown
through IoT-based manufacturing. We should expect to see e-manufacturing plants
and e-maintenance activities in the future.
E. Medical and Health Care
m
)A
(c

[Link]

Amity Directorate of Distance & Online Education

 Internet of Things 179

Remote health monitoring and emergency notification systems can be enabled by IoT
Notes

e
devices. Some hospitals have begun to use smart beds that can detect when they are
occupied and when a patient attempts to get up.

in
F. Home Automation
Home automation, the residential counterpart of building automation, involves the
control and management of various aspects within a household. This includes the

nl
automation of lighting, heating, ventilation, air conditioning (HVAC) and security
systems. Additionally, home automation encompasses the connectivity of various
home appliances such as washers/dryers, ovens and refrigerators/freezers to the
Internet of Things (IoT), allowing for remote monitoring and control through Wi-Fi

O
connections.
G. Energy Management
Energy management can greatly benefit from the integration of sensing and actuation

ity
systems connected to the internet, leading to enhanced optimisation of energy
consumption. The Internet of Things (IoT) devices are anticipated to seamlessly
integrate with various energy-consuming equipment, enabling effective communication
with power generation sources. This integration holds the promise of achieving more
efficient energy utilisation across different sectors.
H. Media and Entertainment

rs
The usage of IoT in media leads data to be sent from one location to another via the
cloud; IoT enables good connection between people by transferring media from one
ve
location to another.
I. Agriculture
By transforming agricultural gear into smart devices, water pumps and sprayers may
ni

be operated from anywhere.

J. Security
There is a concern of robbers in modern life; by utilising IoT in home security systems,
U

the security equipment can be handled by a specific individual from anywhere via the
cloud.

Benefits
ity

1. Manufacturing of new items at manufacturing plants in a timely and accurate manner.

2. In hospitals, use for patient monitoring.
3. It can be used as a home security system.
m

4. It can aid in individual shipping tracking.

5. IoT systems provide faster and more accurately while using less energy, which
increases quality of life.
)A

6. The usage of IoT in transport reduces traffic congestion and collisions.

7. Transfer data from one person to another.

5.1.2 Smart Health Monitoring

(c

Enabled by the Internet of Things (IoT), smart health monitoring systems are
designed to monitor both mental and physical well-being. These systems play a crucial
role in tracking conditions such as stress, anxiety and hypertension, which are common

Amity Directorate of Distance & Online Education

 180 Internet of Things

contributors to both physical and mental health issues. Detecting these issues at an
Notes

e
early stage can significantly enhance the quality of life and lead to reduced healthcare
expenses. Through the utilisation of discrete wearable sensors and advanced machine
learning algorithms, an automated system was developed to identify and eliminate outlier

in
points caused by movement artefacts in blood pressure (BP) and photoplethysmogram
(PPG) signals. The study focused on establishing a relationship between diastolic blood
pressure (DBP) and systolic blood pressure using various features extracted from the

nl
oscillometric waveform envelope. This innovative technology has the potential to prevent
long-term health complications and contribute to the overall improvement of individual
well-being.

O
The Internet of Things (IoT) technology aims to bridge the global health inequality
gap through the implementation of remote patient health monitoring. Referred to as
the Internet of Medical Things (IoMT), these IoT technologies have the potential to
enhance healthcare and safety by establishing a connection between our bodies and the

ity
Internet. This connectivity allows healthcare providers to continuously monitor the health
of individuals, contributing to evidence-based solutions for disease prevention, early
diagnosis, treatment and injury prevention.
However, unlike mechanical or digital technologies, connecting the human body to

rs
the Internet presents challenges. While substantial sensing and measurement equipment
is often used to assess an individual’s health status, it’s constrained by factors like
controlled environments. This limitation restricts the ability to connect the human body to
the Internet in all settings and at all times.
ve
Symptoms of stress, such as headaches and dizziness, can be categorised into
distress and eustress, with distress being the more detrimental form. Stress can be
further divided into short-term, episodic and long-term stress, each characterised by its
duration and impact.
ni

Creating a stress-management strategy involving self-monitoring becomes

imperative in adapting to the advancements in the Internet of Medical Things (IoMT). By
incorporating these concepts into the IoMT framework, individuals can better manage
U

their stress levels and enhance their overall well-being.

ity
m
)A
(c

[Link]
56885789%2Ffigure%2Ffig1%2FAS%3A1099455319281664%401639141983235%2FLogical-analysis-for-
Amity Directorate of Distance & Online Education
 Internet of Things 181

[Link]&tbnid=HEI1pbH1wjNaIM&vet=1&imgrefurl=https%3A%2F%[Link].
net%2Ffigure%2FLogical-analysis-for-stress-detection_fig1_356885789&docid=lZR_y2UcwTROdM&w=6
Notes

e
00&h=842&source=sh%2Fx%2Fim%2F0

The role of stress in the development of anxiety and hypertension is significant, with

in
anxiety often associated with hypertension in various studies. Beyond this, anxiety can
lead to stress-related ailments such as cognitive decline and heart disease. Elevated
blood pressure also has the potential to contribute to heart disease.

nl
Effective monitoring and control of stress, along with its side effects like anxiety and
high blood pressure, can greatly reduce the risk of long-term and potentially irreversible
health issues. Tracking an individual’s stress, anxiety and blood pressure levels can also

O
aid caregivers in making informed diagnoses and initiating timely interventions to prevent
long-term damage.
Several wearable devices positioned on different parts of the body, such as the

ity
chest, wrist, earlobe, fingertip, etc., utilise technologies like PPG, pressure and radio
frequency to measure pulse rate. These wearables serve various purposes, from
detecting heart rate variations and providing medication reminders through vibrations
to collecting data for early detection of cardiac arrest using multimodal systems that
leverage signal processing and machine learning.

rs
Further innovations include wearable systems that monitor heart activity parameters
using ECG devices and fabric-embedded sensors, measure respiratory rate through
thermistor nasal sensors and offer IoT solutions for asthma patients’ respiratory rate
ve
measurement, transmitting and analysing data securely in the cloud. Implantable
electrodes can detect respiratory vibrations, with the acquired data wirelessly transmitted
using ultra-wideband transmitters.
IoT-enabled wearable devices are also advancing healthcare. A portable IoT device,
ni

embedded in a smart vest, constantly monitors the respiratory rate of Chronic Obstructive
Pulmonary Disease (COPD) patients during rest periods. Wearable solutions for real-time
personal health monitoring, including body temperature assessment, are cloud-based,
offering data visibility on both a cloud dashboard and an embedded display.
U

Innovations like ingestible sensors for core-body temperature assessment through

stomach acid power generation and wearable devices for monitoring blood pressure,
blood oxygen levels and glucose levels are emerging. The latter even facilitates
ity

emergency notifications and user-triggered alarm systems.

IoT’s application extends to mental health monitoring, with wearable platforms
that combine behaviour monitoring, privacy-protected audio features and environment
sensing. This holistic approach allows for the evaluation of physiological signals and
m

behavioural features to assess mental well-being without storing raw audio data.

5.1.3 Wearable IoT Devices

)A

IoT-enabled wearables encompass a wide array of devices, ranging from external

accessories and clothing integrations to implanted or bonded gadgets on the skin.
These wearables connect to the Internet to collect, transmit and receive data, enabling
informed decision-making. Over time, wearables have evolved from basic accessories to
specialised tools, becoming a pivotal element in IoT technology.
(c

Smart wearables possess the ability to interact with various devices, notably
smartphones, for computation and communication. Their significance has grown due
to the mobility of humans and animals. These devices can capture and exchange data
Amity Directorate of Distance & Online Education
 182 Internet of Things

while in motion and access information from the Internet, contributing to enhanced
Notes

e
decision-making. The integration of smart wearables can yield benefits such as improved
application performance, better quality of life, heightened productivity and enhanced
safety measures.

in
The rapid evolution of wearable technology has been propelled by advancements
in energy-efficient mobile networks, miniaturisation of electronic devices and sensors
and the advantageous capabilities offered by smart wearables. In recent years, we have

nl
witnessed swift progress in the creation of smart wearable products tailored for diverse
applications. This range of wearables encompasses smartwatches, wristbands, eyewear,
headsets, earbuds, body straps, hand and foot worn devices and intelligent jewellery,

O
each designed to serve specific purposes.

ity
rs
ve
ni

[Link]
U

Figure: Different wearables developed for various applications.

The evolution of wearable technology has seen distinct waves in the market,
beginning with fitness activity trackers, followed by Bluetooth headphones, smartwatches
ity

and web-connected eyewear. The gaming industry also integrated wearables with the
introduction of virtual reality and augmented reality headsets. Nonetheless, the most
transformative and impactful applications of wearables are evident in health monitoring
and medical contexts.
m

Numerous research papers have delved into different facets of wearable devices.
The initial comprehensive analysis of IoT-enabled wearables explores their applications,
opportunities and challenges, categorising them into various clusters and highlighting the
unique algorithmic aspects associated with each cluster. A comprehensive exploration
)A

of cellular IoT is warranted, given its potential to revolutionise the landscape of IoT-
based wearable technology, encompassing applications, potential advancements and
challenges.
Wearable IoT devices are inherently reliant on consistent power consumption to
(c

perform intricate processing tasks and transmit data. Nevertheless, the versatility of these
devices extends to diverse applications and situations, ranging from fall detection, even
during sleep, to tracking the trajectory of a bird over a year. Enhancing the longevity of

Amity Directorate of Distance & Online Education

 Internet of Things 183

battery life cycles calls for strategic charging practices, where wearables can be worn
Notes

e
during periods of rest and recharged following active periods.
Furthermore, the integration of energy harvesting solutions offers an effective

in
means to extend the duration of battery life cycles. By curbing power consumption
and embracing energy harvesting, the endurance of wearable IoT devices can be
significantly prolonged. Collectively, the judicious management of power resources and
the incorporation of energy harvesting techniques stand as crucial factors in safeguarding

nl
the sustained functionality of wearable IoT devices.
The boundless potential of wearable IoT technology is poised to be fully harnessed
through integrated IoT systems. Wearable IoT devices predominantly connect to the

O
Internet using two approaches: data transmission to the cloud for offline processing or
offloading computation to the Internet for real-time analysis. The complete realisation
of wearable IoT’s potential hinges on overcoming challenges such as data ownership,
sharing policies, privacy and security concerns. While most wearable IoT devices

ity
rely on short-range communication like Wi-Fi and Bluetooth, the advent of cellular IoT
technologies like LTE-M and NB-IoT in 3GPP Release 13 is set to introduce novel
applications and connectivity methods, moving beyond current limitations.

Wearable IoT Classification

rs
In order to classify wearable IoT applications, we conducted a comprehensive survey
of recent research from sources like IEEE Xplore and ACM digital libraries. Some clusters
exhibit a broader range of applications due to their significance, while additional wearable
ve
IoT devices are employed in fields like gaming, payments and education. However, these
additional use cases fall outside the scope of this survey.
ni
U
ity

[Link]
05c10d99923f0a400f39d5a0bf4b458%[Link]&tbnid=XpLnMaBfx6beWM&vet=1&imgrefurl=htt
m

ps%3A%2F%[Link]%2Fpaper%2FWearables-and-the-Internet-of-Things-(IoT)%2C-
and-A-Dian-Vahidnia%2F02342fc133862753156f5ed3a266f9ee324dd2c1%2Ffigure%2F3&docid=88mYtW2
DISon1M&w=660&h=296&source=sh%2Fx%2Fim%2F0
Figure: Taxonomy of wearable health care IoT sensors.
)A

Each category is covered in the following sections in the most efficient manner.
A. Health
IoT health wearables are crucial for remote patient monitoring and treatment. These
devices use sensors to capture health data, which is then processed and sent to
(c

the Internet for analysis. Wearables can connect to smartphones, allowing data
visualisation and analysis through cloud-based platforms like Microsoft Azure or AWS.

Amity Directorate of Distance & Online Education

 184 Internet of Things

In therapeutic cases, analysed data can prompt wearable actions such as heating or
Notes

e
shock for personalised treatment.
1) Health Treatment and Rehabilitation Wearable Systems

in
IoT-based rehabilitation devices aid individuals with impairments in enhancing their
physical and mental abilities. A physiotherapy system using a walker continuously
monitors movement parameters, sending data to the cloud for analysis and displaying

nl
results on a mobile app. Another device, a wristband for stroke rehabilitation, detects
biopotential signals, transmitting data to a robot hand for real-time feedback and posture
adjustment. A smart wheelchair with smartphone interaction allows remote monitoring

O
by carers. Stroke patients benefit from a rehabilitation approach that merges gaming
and wearables. These innovations showcase IoT’s role in personalised and interactive
rehabilitation techniques for improved patient outcomes.
2) Health Monitoring Wearable Systems

ity
Health monitoring wearables are categorised into four major groups based on the type
of sensors used:
™™ Electroencephalography (EEG), electrocardiography (ECG), electromyography
(EMG), photoplesmography (PPG) and other bio-potential sensors are
examples.

rs
™™ Accelerometer and gyroscope are motion sensors.
™™ Environmental sensors include ultrasonic, pressure, temperature and other
ve
variables.
™™ Transdermal glucose is a biochemical sensor.
™™ Analysing the aforementioned signals provides substantial information on the
ni

patient’s/user’s health status.

B. Activity Recognition and Sports
This category includes apps in which wearables are worn during sports activities
U

to record various parameters of the user/athlete’s activity in order to improve his/

her performance. This cluster’s applications also address data collection for the
recognition of human and animal everyday activities. Although activity recognition has
certain uses in medical diagnostics and out-of-hospital health-care, the applications in
ity

this cluster involve use cases that are not related to health. In this section, we will look
at how wearable IoT can be used for two different activities.
™™ The first group of activities is concerned with recognising daily physical activity;
and
m

™™ The second is concerned with activities associated with specific sports.

Tracking daily physical activities and body movements, including actions like walking,
running and sitting, is a primary application of IoT in sports. This technology offers
)A

immense potential for enhancing athlete performance, training methods, injury

prevention and fan engagement. Wearable IoT devices have revolutionised sports
by enabling better training insights, injury analysis and personalised fan experiences.
Despite challenges posed by league regulations and obstacles, the adoption of
wearable IoT devices in sports is growing as organisation s recognise the benefits they
(c

offer to players, coaches and fans. While wearable IoT presents limitless possibilities

Amity Directorate of Distance & Online Education

 Internet of Things 185

in sports, this section focuses on activity recognition for specific sports such as skiing,
Notes

e
martial arts, tennis and more.
The detection of human physical activity has garnered significant attention. Wearable

in
devices with sensors attached to the body capture signals that indicate various
activities, which are subsequently analysed. This process involves several stages.
Pre-processing eliminates noise and prepares signals for further analysis. Feature

nl
extraction identifies distinctive properties of the signals, categorised as structural or
statistical, which help differentiate activities. Feature reduction eliminates redundant
data to retain essential information for the final stage.

O
The final stage involves classification, wherein algorithms identify patterns among
activities using the reduced set of features. Classification algorithms calculate activity
probabilities based on features, use decision trees, neural networks, or machine
learning algorithms to map features to activities, providing real-time feedback to

ity
users. This approach aids in improving precision, performance and consistency in
sports-related movements.
Real-time feedback is a key aspect of wearable IoT devices in sports. Gamers receive
performance insights and guidance to enhance their skills. Messages are conveyed

rs
through wearable screens, displays, audio, or haptic indicators, fostering better
understanding of performance quality and improvement strategies. Various sports
have been studied, including badminton, basketball, rowing, swimming, hockey, skiing,
martial arts, weightlifting, tennis, baseball and golf. Despite focusing on specific sports
ve
in this context, the broader potential of wearable IoT in sports remains significant.
C. Tracking and Localisation
This category primarily focuses on utilising wearable IoT devices to track the positions
ni

of humans and animals in real time. Applications range from studying bird flight
trajectories to monitoring the location of seniors in care facilities, analysing visitor
movements at exhibitions and tracking pets. While offers a comprehensive study
on IoT-based localisation, it does not specifically concentrate on wearable IoT
U

applications.
Broadly, localisation methods in the literature can be divided into two types: those
dependent on offline training and those that are not. Training-dependent approaches
ity

fall into three groups: fingerprinting, stochastic-oriented models and machine learning-
based schemes.
Fingerprinting extensively explores various signals, such as Received Signal Strength
Indicator (RSSI) values from wireless signals, audio, video and motion. The RSSI
m

values are collected during a site survey and form a fingerprint database used by online
localisation algorithms to estimate positions. Collaborative localisation techniques
based on mutual location measurement and relative distance among IoT devices can
also be employed.
)A

The category of non-training dependent localisation methods can also be subdivided.

The primary cluster includes techniques like multilateration, triangulation and region
overlapping, which leverage geometric information for position estimation. Mobility,
path planning and statistics comprise the other non-training categories.
(c

The cellular IoT 3GPP Release employs the Observed Time Difference of Arrival
(OTDOA) method for position estimation based on geometric principles. OTDOA

Amity Directorate of Distance & Online Education

 186 Internet of Things

calculates the time difference of arrival of Positioning Reference Signals (PRSs) from
Notes

e
multiple nodes relative to a reference node’s transmission. This data is converted into
a geographical hyperbola, with the device’s position approximated at the intersection
point of hyperbolas. The horizontal error position distribution is measured, indicating

in
that LTE-M network usage provides a horizontal resolution of around 50 meters,
suitable for various wearable IoT applications.
Wearable IoT devices play a pivotal role in real-time position tracking for various

nl
purposes. While training-dependent methods rely on fingerprinting, stochastic models
and machine learning, non-training dependent approaches encompass geometric
techniques, mobility, path planning and statistics. The OTDOA method, specifically

O
in cellular IoT, aids in accurate position estimation by exploiting geometric properties.
The use of LTE-M networks further enhances the resolution of position data, making
wearable IoT localisation applicable to a wide array of scenarios.

ity
D. Safety
This category encompasses wearable IoT devices designed to enhance user safety.
For instance, a fatigue monitoring system can alert drowsy drivers and inform
employers, promoting road safety. Similarly, wearable gadgets gather air quality data
in mines to safeguard worker well-being, mitigate risks and reduce employer costs.

rs
Fall prevention and detection, especially among the elderly, is a critical concern, with
wearable devices serving various applications in this context.
Wearable IoT safety sensors predominantly cater to three vital applications:
ve
1) Fall Detection and Prevention
The elderly population faces a significant risk of accidental falls, which can greatly
impact their quality of life. Wearable IoT devices prove advantageous for continuous
ni

fall monitoring compared to permanent infrastructure solutions like fixed cameras or

smart tiles, due to their location flexibility and lack of scene obstruction. This section
delves into the literature regarding wearable IoT applications in fall detection and
U

prediction.
Incorporating inertial sensors like accelerometers or gyroscopes is common for
fall detection. Rapid detection of falls is essential, requiring the system to discern
between falls and routine activities while minimising false positives. Factors like body
ity

position, falling speed and angular velocity contribute to this differentiation and these
characteristics vary not only among individuals but also with age. Consequently,
accurate fall detection mandates complex calculations.
Most fall detection algorithms leverage motion data to differentiate falls from everyday
m

actions. A fall is identified as an involuntary, abrupt motion not executed in a controlled

manner. Threshold-based techniques are prevalent, where motion data (e.g., vertical
speed profile) is compared to threshold values to ascertain stability, fall occurrence, or
)A

likelihood of falling. Thresholds can be either fixed or adaptive, with the latter adjusting
based on historical data or user characteristics like gender, height, weight and age.
For fall detection, predefined reference templates specific to different fall types are
utilised. These templates involve comparing angles and angular velocities of thigh
segments between fall events and regular activities. Some studies propose machine
(c

learning methods such as Support Vector Machines (SVM) employing feature selection

Amity Directorate of Distance & Online Education

 Internet of Things 187

techniques on raw signals from gyroscopes and accelerometers. This facilitates

Notes

e
accurate fall detection and prevention.
2) Drowsiness-Fatigue-Detection

in
Drowsiness detection and alert mechanisms can be categorised into five groups: image
processing-based, EEG-based, artificial neural network-based, vehicular-based and
subjective measures. These techniques demand robust processing capabilities to

nl
perform intricate computations for identifying drowsiness or fatigue. Alternatively, data
processing can be offloaded to powerful cloud servers by transmitting raw data.
A drowsiness-fatigue-detection system, cantered around wearable smart glasses,

O
enhances road safety. Comprising smart glasses, an in-car telematics platform, an
automobile diagnostic bridge, an active vehicle rear light alert mechanism and a cloud-
based management platform, this setup detects driver drowsiness or fatigue in real
time. Upon detection, other drivers are alerted via the vehicle’s rear light mechanism,

ity
while data is simultaneously transmitted to a cloud platform for further analysis.
For worker drowsiness, a cost-effective EEG-based system is proposed. A modified
safety smart helmet captures data and transmits it to a local server, where a random
forest classification algorithm determines the severity of the sleep situation. Should a

rs
worker fall, an Inertial Measurement Unit (IMU) sensor detects it. A dedicated mobile
application allows supervisors to track the worker’s status consistently.
3) Environmental Condition Monitoring
ve
Environmental hazards can pose significant health risks, necessitating swift response
mechanisms. Wearable sensors offer valuable insights into the impact of environmental
factors on health. They can also be integrated into mobile sensor networks on vehicles
or animals to predict dangerous conditions. For instance, data gathered from animals
ni

in diverse locations can aid in predicting earthquakes.

Research has yielded safety wearables featuring self-powered sensor networks,
employing solar energy harvesting with Maximum Power Point Control (MPPT), low-
U

power environmental sensing nodes and long-range wireless modules like LoRa.
These wearables measure temperature, humidity and UV index, transmitting data
via LoRa to gateways and cloud servers. Another application employs wearables to
detect hazardous conditions like CO2 levels, temperature and humidity, utilising XBee
ity

DigiMesh modules for communication.

Furthermore, an urban environmental monitoring network integrates seven sensors,
including infrared temperature, pressure, accelerometer, ambient light and IMU,
using Wi-Fi for wireless communication with minimal power consumption. Similarly,
m

a low-power wearable network is under development, assessing CO2 concentration,

magnetic field, temperature and humidity, communicating through Bluetooth
technology.
)A

E. Others
In addition to the four major clusters covered thus far, wearable IoT technology is being
developed for use in various applications such as education, virtual and augmented
reality, law enforcement and surveillance.
(c

Amity Directorate of Distance & Online Education

 188 Internet of Things

5.1.4 Home Automation

Notes

e
in
nl
O
[Link]

ity
Home automation, often known as a smart home, is the application of IoT technology
to control and manage various household devices and systems. This includes devices
like lights, fans, TVs, as well as features related to lighting, climate, entertainment and
security. By connecting these devices to the internet, they become part of the Internet of
Things (IoT), allowing remote control and monitoring.

rs
A typical home automation setup involves linking devices to a central hub, which
can be accessed through various means such as wall-mounted terminals, smartphones,
tablets, or online interfaces. This integration enables the management and automation of
ve
services within a residence, like heating, cooling, lighting and security. The advent of IoT
and Machine-to-Machine (M2M) communication has further expanded the capabilities of
home automation, enabled remote control and monitored over networks.
The key advantage of IoT-enabled home automation lies in its ability to control a
ni

wide array of devices from a centralised interface, enhancing convenience and efficiency.
This encompasses everything from adjusting temperatures and lighting to managing
security cameras and alarms. Another significant benefit is the capability to monitor
and control devices remotely, even when away from home, contributing to energy
U

conservation and overall safety.

Incorporating smart devices like thermostats, light bulbs and security cameras into
IoT-enabled home automation is commonplace. These devices utilise wireless protocols
ity

like Zigbee, Z-Wave and Bluetooth to communicate with each other and the central hub.
Moreover, these systems can seamlessly interact with voice assistants such as Alexa and
Google Home, augmenting their functionality.
IoT-enabled home automation brings a host of advantages to homeowners,
m

including enhanced convenience, energy efficiency and security. Nevertheless, ensuring

the security of these systems is crucial due to potential vulnerabilities that may lead to
hacking and cybersecurity risks.
)A

Home automation relies on three fundamental components to create a seamless

smart home experience:
1. Hardware
2. Software/Apps
(c

3. Communication Protocols
In addition to these components, firmware also plays a pivotal role. Firmware is
the software embedded within hardware devices, responsible for managing data, data
Amity Directorate of Distance & Online Education
 Internet of Things 189

transfer, over-the-air (OTA) updates and other critical processes necessary for effective
Notes

e
communication between devices. Firmware ensures that devices can communicate with
each other and with central systems smoothly and reliably.

in
Collectively, these components work in harmony to create a comprehensive home
automation ecosystem that offers convenience, efficiency and enhanced control over
various aspects of daily life.

nl
Applications of Home Automation
Home automation is poised to revolutionise various aspects of the modern
consumer’s life, reshaping their expectations and delivering a host of innovative

O
applications. The integration of IoT-enabled connectivity in home automation is
anticipated to bring transformative changes to the following domains:
™™ Lighting control

ity
™™ HVAC
™™ Lawn/Gardening management
™™ Smart Home Appliances
™™ Improved Home safety and security
™™ Home air quality and water quality monitoring
™™ Natural Language-based voice assistants
™™ Better Infotainment delivery
rs
ve
™™ AI-driven digital experiences
™™ Smart Switches
™™ Smart Locks
ni

™™ Smart Energy Meters

This list is by no means exhaustive and will evolve over time to accommodate
emerging IoT applications. A more detailed exploration of the components required to
U

develop a fundamental home automation prototype is warranted

Home Automation Components

These have already been discussed, but let’s break them down into components
ity

to assist develop a realistic model of what major components are involved in building a
smart house. The primary components are as follows:
™™ IoT Sensors
™™ IoT Gateways
m

™™ IoT Protocols
™™ IoT Firmware
™™ IoT Cloud and Databases
)A

™™ IoT Middleware (if required)

There are thousands of IoT sensors used in home automation, as well as hundreds
of home automation gateways. The majority of the firmware is written in C, Python, Node.
Js, or some other programming language.
(c

Platform as a service (PaaS) and infrastructure as a service (LaaS) are the two
largest providers in IoT cloud. The following are major IoT platforms as a service provider:

Amity Directorate of Distance & Online Education

 190 Internet of Things

™™ AWS IoT
Notes

e
™™ Azure IoT
™™ Thingworx

in
™™ Ubidots
™™ Thingspeak
™™ Carriots

nl
™™ Konekt
™™ TempoIQ

O
™™ Xively
™™ IBM Bluemix

Characteristics of IoT Platforms

ity
These platforms exhibit a significant range of diversity with regard to their IoT
applications and security offerings. Many of these platforms are available as open-source
solutions, some of which are even free. Here’s a glimpse into the typical attributes one
can anticipate from a standard IoT platform:

rs
™™ Robust Device Security and Authentication
™™ Efficient Message Brokers and Message Queuing Mechanisms
™™ Comprehensive Device Administration Features
ve
™™ Support for Essential Protocols such as CoAP, MQTT and HTTP
™™ Capability for Data Collection, Visualisation and Basic Analysis
™™ Seamless Integration with Other Web Services
™™ Scalability in Both Horizontal and Vertical Dimensions
ni

™™ WebSocket APIs Facilitating Real-Time Information Flow

Platform developers are progressively expanding access to their libraries for
U

developers. For instance, the Dallas Temperature Library for DS18B20 on Arduino
underwent rapid enhancement through open-source collaboration, enabling seamless
integration of DS18B20 with Linkit One. Recognising these dynamics is pivotal as the IoT
landscape advances swiftly and possessing an equally responsive platform is imperative
ity

for confident business progress

Let’s take a closer look at each of these components, beginning with IoT sensors.
A multitude of additional sensors are available that could potentially be
encompassed in this compilation. To maintain conciseness, we’ll provide a succinct
m

overview as an introduction to smart home technologies. The following roster organises

IoT sensors for home automation according to their distinct sensing capabilities:
1. Temperature Sensors
)A

2. Lux Sensors
3. Water Level Sensors
4. Air Composition Sensors
5. Surveillance Video Cameras
(c

6. Voice/Sound Sensors
7. Pressure Sensors

Amity Directorate of Distance & Online Education

 Internet of Things 191

8. Humidity Sensors
Notes

e
9. Accelerometers
10. Infrared Sensors

in
11. Vibration Sensors
12. Ultrasonic Sensors

nl
Depending on specific requirements, a combination of these sensors can be
harnessed to craft an authentically intelligent home IoT product. We will now delve into
some of the most prevalent and widely used sensors in the realm of home automation.

O
Temperature Sensors
The market is flooded with them, but the most well-known temperature sensors
include TI’s DHT11/22, DS18B20, LM35 and MSP430 family. The MSP430 series is more
accurate than the others, but it is also one of the most expensive for prototyping or first

ity
product testing. MSP430 temperature sensors outperform all others in terms of precision
and battery consumption.
●● MSP430 temperature sensors outperform all others in terms of precision and battery
consumption.
●●

●● rs
DHT11 has a relatively limited temperature range and has accuracy difficulties.
DHT22, on the other hand, is slightly more accurate, although it is still not preferred.
DS18B20s, on the other hand, are more accurate. Unlike digital temperature
ve
sensors such as the DHT22 and 11, Dallas temperature sensors are analogue and
can be exceedingly accurate to within 0.5 degrees.

Home-automation-sensors
ni

Please be aware that the temperature readings you directly obtain from these
sensors may not always be accurate and at times, you might even observe values
exceeding 1000°F regardless of the actual conditions.
U

Lux Sensors
Lux sensors play a pivotal role in gauging luminance and hold the potential to initiate
an array of actions, spanning from corroborating motions to activating lights in response
to diminishing light levels. Among the extensively employed light sensors, TSL2591 and
ity

BH1750 stand out. Recent research involving these sensors in low-power IoT devices
has demonstrated their commendable performance across a spectrum of practical
scenarios.

Water Level Sensors

m

While developing your prototype, consider using a solid state eTape liquid
level sensor, or just an HC-SR04 ultrasonic sensor, as others have done. In other
circumstances, where those two are insufficient, something with far higher performance
)A

must be used. Float level sensors and other integrated circuits, such as the LM1830,
provide IoT developers with more precise measuring capabilities. They are, however, far
more expensive than others.

Air Composition Sensors

(c

Certain sensors are selectively utilised by developers to effectively monitor diverse

air constituents:

Amity Directorate of Distance & Online Education

 192 Internet of Things

™™ MiCS-5525 for CO monitoring

Notes

e
™™ MQ-8 for assessing Hydrogen gas levels
™™ MiCS-2714 for gauging nitrogen oxide

in
™™ MQ135 for detecting hazardous gas levels (NH3, NOx, Alcohol, Benzene,
smoke, CO2)
Many of these sensors necessitate a warm-up period before providing accurate

nl
measurements due to their heating requirement.

Gas-sensor-before-and-after-heating

O
These sensors detect gas components primarily through their surface. When they
first begin sensing, there is always something on their surface, some form of deposition
that requires some heating to remove. As a result, once the surface is sufficiently heated,
genuine values begin to emerge.

ity
Video Cameras for Surveillance and Analytics
In such cases, a variety of webcams and cameras specific to Hardware development
kits are typically employed. Hardware with USB connections is available for integration,
as is a camera module for building functionalities. However, using USB connections

processing.
rs
is inefficient, especially when it comes to real-time video transfer or any type of video

Consider the Raspberry Pi, which includes a camera module (Pi cam) that attaches
ve
to the board through a flex connector rather than the USB port. As a result, the Pi cam is
incredibly efficient.

Sound Detection
ni

Sound detection holds significance in various applications, from monitoring infants

to triggering light switches based on ambient noise and recognising pets’ vocal cues. For
quick prototyping, widely utilised sound detection sensors encompass SEN-12462 and
EasyVR Shield.
U

However, it’s worth noting that while these sensors are effective for rapid
development, they may not match the precision of industrial-grade counterparts like
those offered by 3DSignals. Industrial sensors such as those from 3DSignals excel at
ity

detecting even minute noise levels and finely distinguishing patterns to identify machinery
anomalies.

Humidity Sensors
These sensors enable smart houses to detect humidity/RH levels in the air.
m

The accuracy and sensing precision are heavily influenced by a variety of elements,
including the overall sensor design and placement. However, sensors designed for rapid
prototyping, such as the DHT22 and 11, usually perform badly when compared to high-
)A

quality sensors, such as the HIH6100 and Dig RH.

When developing a product to detect humidity levels, be certain that there is no
localised layer of humidity that obscures the true results. Also, keep in mind that in some
tiny rooms, the humidity may be too high at one end relative to the others. When looking
at free and open places where the air components can move freely, the distribution
(c

surrounding the sensor can be assumed to be uniform, requiring very few corrective
steps for proper calibration.

Amity Directorate of Distance & Online Education

 Internet of Things 193

Home Automation Protocols

Notes

e
One of the most crucial aspects of developing a home automation product is
considering protocols, which your device will use to communicate with gateways, servers

in
and sensors. Until recently, the only option to do so was to use Bluetooth, Wi-Fi, or GSM.
However, most such solutions failed due to increased costs for cellular sim cards and
poor Wi-Fi performance.

nl
Until a short time ago, the available alternatives for connectivity included Bluetooth,
Wi-Fi, or GSM. Nevertheless, many of these solutions encountered challenges
such as elevated expenses for cellular SIM cards and subpar Wi-Fi performance.
Bluetooth persevered and evolved into Bluetooth Smart or Bluetooth Low Energy. This

O
advancement played a crucial role in fostering connectivity within the “mobile server
powered economy,” wherein smartphones function as intermediaries to collect data from
BLE-enabled sensors and transmit it to the internet.

ity
When it comes to significant home automation protocols, the following are at the top
of the list:
™™ Bluetooth low energy, often known as Bluetooth Smart, is a wireless protocol
that includes mesh capabilities, security, data encryption methods and other

rs
features. Ideal for IoT-based smart home gadgets.
™™ Zigbee: A low-cost, mesh-networked, radio frequency-based protocol for IoT.
Different Zigbee versions do not communicate with one another.
ve
™™ X10: A vintage protocol for signalling and control that makes use of powerline
wiring.
™™ Insteon: Communicates with devices wirelessly as well as through wires.
™™ Z-wave: A home automation technology that focuses on security.
ni

™™ Wi-Fi is self-explanatory.
™™ UPB: Reduces costs by utilising existing electricity lines installed in a home.
™™ Thread: 6lowpan is a royalty-free protocol for smart home automation.
U

™™ ANT: An ultra-low-power protocol that enables developers to create low-power

sensors with mesh distribution capabilities.
™™ 6lowpan
ity

Home Automation: Which protocol is the best?

When it comes to selecting a protocol for smart home development, it’s crucial to
begin by identifying the specific needs of the project and then refining the options. The
commonly utilised protocols include Bluetooth Low Energy, Z-Wave, Zigbee and Thread.
m

To streamline the selection process, consider the following criteria:

™™ Capability to perform identity verification
™™ Eminence of sensor networks
)A

™™ Data transfer rate

™™ Security level
™™ Network topology needed
™™ Density of objects around
(c

™™ Effective Distance to be covered

Amity Directorate of Distance & Online Education

 194 Internet of Things

5.1.5 Futures of Automation using IoT

Notes

e
The Internet of Things (IoT) has enormous potential for the future of automation,
especially when paired with the capabilities of Artificial Intelligence (AI). This combination

in
of IoT and AI is projected to result in profound changes across industries, revolutionising
how businesses function and interact with the physical environment. Here’s a glimpse
into the future of automation via IoT and AI integration:

nl
1. Intelligent Decision-Making: IoT devices create massive amounts of data from their
physical surroundings. These devices can analyse real-time data and make intelligent
judgements on their own by incorporating AI algorithms. This enables proactive actions
based on findings, such as optimising manufacturing processes or automatically

O
altering traffic signals to reduce congestion.
2. Predictive Maintenance: Sensors with IoT capabilities collect data from machines and
equipment. This data can be processed by AI algorithms to predict when maintenance

ity
is required, minimising costly breakdowns and downtime. AI-powered predictive
maintenance, for example, may identify when a machine is going to fail and send out
a repair warning before it happens.
3. Cognitive Automation: Cognitive automation refers to systems that can understand,
learn and reason in the same way as humans do. When IoT and AI are combined,

rs
devices can learn from patterns and previous data, enhancing their performance over
time. This has the potential to result in self-learning systems that adapt to changing
situations, such as optimising energy use in smart buildings.
ve
4. Enhanced Customer Experiences: Businesses may build highly personalised and
context-aware customer experiences by merging IoT-generated data with AI-driven
insights. Retailers, for example, can employ IoT devices to collect data on customer
behaviour in-store and then use AI to adjust product recommendations and offers in
ni

real-time.
5. Autonomous Vehicles and Transportation: The integration of IoT and AI is critical
for the development of self-driving cars. IoT sensors gather information about road
U

conditions, traffic patterns and vehicle performance, which AI algorithms use to make
split-second judgements for safe navigation and collision avoidance.
6. Healthcare and Remote Monitoring: IoT devices can continuously monitor vital
signs and health parameters in patients. AI-powered analytics can detect anomalies,
ity

anticipate health deterioration and notify medical personnel or carers ahead of time,
improving patient care and lowering hospitalisations.
7. Energy Efficiency and Sustainability: Smart buildings’ IoT devices can collect data
on energy usage, occupancy patterns and environmental factors. This data can be
m

analysed by AI systems to optimise energy use, lowering costs and carbon footprints.
8. Supply Chain Optimisation: AI-powered analytics can optimise supply chain logistics
by leveraging data from IoT sensors installed on goods and vehicles. This involves
)A

tracking shipments in real time, forecasting delivery delays and modifying routes for
optimal transportation.
9. Agriculture and Precision Farming: In agriculture, IoT-enabled sensors collect data
on soil conditions, weather patterns and crop health. This data is analysed by AI
algorithms to provide insights on optimal planting times, irrigation schedules and pest
(c

control tactics.
10. Smart Cities and Urban Planning: The integration of IoT and AI is critical for the

Amity Directorate of Distance & Online Education

 Internet of Things 195

development of smart cities. Sensors integrated in urban infrastructure can collect

Notes

e
information about traffic flow, trash management, energy use and public services.
Analytics enabled by AI can assist municipal planners in making educated decisions
to increase efficiency and quality of life.

in
The integration of IoT and AI has enormous promise for the future of automation.
It enables organisations and industries to improve productivity, sustainability and overall
human experiences by leveraging real-time data, intelligent decision-making and

nl
autonomous systems. As these technologies progress, their combined potential will fuel
innovation and influence how we interact with our surroundings.

O
Machine-to-Machine Communication and Autonomous Decision-Making
The Internet of Things (IoT) has enormous promise for altering industries and daily
lives in the future. Machine-to-machine (M2M) connectivity and autonomous decision-
making are two fundamental factors driving this transition. Let’s take a look at how these

ity
ideas are shaping the future of automation:

Machine-to-Machine Communication (M2M)

rs
ve
ni
U
ity

[Link]

Machine-to-machine communication (M2M) is the automatic exchange of data and

collaboration between devices, sensors and machines that does not require human
interaction. This technology has far-reaching implications for the IoT ecosystem, including
m

improved resource management, supply chain optimisation, predictive maintenance,

smart cities and industrial automation.
M2M communication allows for real-time tracking and monitoring of items, which
)A

improves logistics, reduces delays and improves inventory management. Machines that
are linked can transmit their operational status, allowing for improved maintenance and
increasing the life of equipment. Smart cities can benefit from better traffic management,
garbage disposal and infrastructure maintenance, while industrial automation can help
decrease bottlenecks and optimise productivity.
(c

Overall, M2M communication is an important component of the IoT ecosystem,

altering a wide range of businesses and applications.

Amity Directorate of Distance & Online Education

 196 Internet of Things

Autonomous Decision-Making
Notes

e
In the Internet of Things, autonomous decision-making involves devices and systems
making judgements based on data and predetermined criteria without human interaction.

in
AI, machine learning and real-time analytics advancements have made autonomous
decision-making more sophisticated. Real-time reaction, emergency response, precision
agriculture, environmental monitoring and healthcare and well-being are examples of
these.

nl
Autonomous vehicles are capable of analysing incoming data, detecting trends and
making rapid judgements, such as reacting to changing road conditions or crashes. IoT
devices can also activate emergency reactions like fire alarms and medical notifications.

O
Precision agriculture uses soil parameters and weather forecasts to alter sowing rates,
fertiliser application and irrigation. By changing water flow based on water quality
assessments, environmental monitoring can help to conserve ecosystems. Healthcare
and well-being devices can notify medical personnel or carers if patients’ health readings

ity
are aberrant.
As M2M communication and self-driving cars become more common, ethical
concerns about data privacy, security and unexpected repercussions must be addressed.
Nonetheless, the combined strength of these technologies has the potential to

rs
revolutionise industries, increase efficiency and improve quality of life by creating a more
interconnected and intelligent society.

Predictive Maintenance and Real-time Analytics

ve
The Internet of Things (IoT) has enormous potential to alter industries through
predictive maintenance and real-time analytics in the future of automation. This
convergence uses IoT devices, data analytics and automation to improve operational
efficiency, reduce downtime and improve decision-making processes.
ni

Predictive Maintenance
It entails predicting when equipment or machinery is likely to break using data from
U

IoT-connected sensors and devices. This proactive maintenance strategy replaces typical
reactive maintenance, resulting in less unplanned downtime and associated expenses.
●● Sensor Data Collection: Sensors with Internet of Things capabilities continuously
ity

monitor equipment conditions, gathering data on elements such as temperature,

vibration, pressure and performance indicators.
●● Machine Learning and Analytics: Advanced analytics and machine learning
techniques are used to examine the obtained data for trends and abnormalities. This
aids in the prediction of potential breakdowns and the estimation of equipment’s
m

remaining usable life.

●● Early Alerts and Optimisation: Maintenance teams receive early warnings when
the analytics detect indicators of approaching failure, allowing them to arrange
)A

maintenance actions during planned downtime. This avoids costly unplanned

downtime and lowers maintenance expenses.
●● Operational Efficiency: Predictive maintenance improves overall operating efficiency
by optimising equipment utilisation, reducing the requirement for spare parts
inventories and extending the lifespan of machines.
(c

Real-time Analytics
Real-time analytics entails processing and analysing data in real time, allowing for
Amity Directorate of Distance & Online Education
 Internet of Things 197

immediate insights and informed decision-making. Real-time analytics powered by IoT

Notes

e
have several applications across industries:
●● Manufacturing: In manufacturing, real-time analytics may monitor production lines,

in
detect errors, optimise quality control and alter procedures in real-time to ensure
optimal output.
●● Logistics and Supply Chain: The ability to watch shipments, inventory levels and

nl
demand patterns in real time allows for more efficient inventory management, route
optimisation and responsive supply chain operations.
●● Energy Management: Real-time analytics of energy consumption trends in buildings

O
and factories can detect inefficient usage and provide energy-saving methods.
●● Healthcare: In healthcare, IoT devices can continuously monitor patient vital signs
and send real-time notifications to medical staff in the event of an emergency or a
change in health conditions.

ity
●● Smart Cities: To optimise city operations and improve public services, real-time
analytics can monitor and manage urban infrastructure such as traffic lights, trash
management and emergency services.

Benefits
1.

rs
Cost Reduction: Predictive maintenance eliminates unplanned downtime, reduces
equipment failures and optimises maintenance plans, all of which result in cost
savings.
ve
2. Increased Productivity: Real-time analytics allow for quick responses to changing
situations, resulting in increased efficiency and production across industries.
3. Enhanced Safety: Predictive maintenance promotes worker safety by recognising
ni

potential dangers before they increase, which is especially important in industries

such as manufacturing and mining.
4. Improved Decision-Making: Real-time insights from IoT data enable businesses to
U

make educated decisions faster, resulting in improved outcomes and a competitive

advantage.
5. Sustainability: Using real-time analytics to optimise resource usage and reduce energy
consumption contributes to sustainability and environmental conservation.
ity

The capacity of predictive maintenance and real-time data to drive efficiency,

decrease costs, improve safety and alter industries across the spectrum is the future
of IoT automation. As IoT technology advances, its integration with data analytics and
automation will likely influence how businesses function and create value.
m

Intelligent Supply Chain Management and Logistics

Automation powered by IoT has enormous potential to revolutionise supply chain
)A

management and logistics. The incorporation of IoT technology into these sectors has
the potential to increase efficiency, improve visibility and optimise operations across the
whole supply chain. Let’s take a look at how IoT-powered automation is influencing the
future of supply chain management and logistics:

Intelligent Supply Chain Management

(c

1. Real-time Tracking and Visibility: IoT sensors and RFID tags can track items, shipments
and inventory in real time, allowing complete visibility into the movement and status of

Amity Directorate of Distance & Online Education

 198 Internet of Things

goods at every level of the supply chain. This real-time data allows for precise demand
Notes

e
forecasts, inventory optimisation and proactive problem resolution.
2. Predictive Analytics: IoT devices can collect data from a variety of sources, including

in
weather, production rates and transit routes. Supply chain managers may foresee
disruptions, optimise routes and change inventory levels using advanced analytics
and machine learning to reduce delays and costs.

nl
3. Smart Warehousing: Warehouse factors such as temperature, humidity and storage
density can be monitored using IoT devices. Automated systems can optimise storage
allocation, track product shelf life and make picking and packing more efficient.

O
4. Supplier Collaboration: Suppliers, manufacturers and distributors benefit from
seamless collaboration in IoT-enabled supply chains. Real-time data sharing and
communication improve coordination and allow for quick responses to changes in
demand or supply.

ity
5. Automated Reordering: When inventory levels hit predetermined criteria, IoT sensors
can automatically initiate reorders. This reduces stockouts and overstocks, resulting
in appropriate inventory levels.

Intelligent Logistics
1.

rs
Fleet Management: Vehicles and assets equipped with IoT can broadcast real-time
data on position, speed, fuel consumption and maintenance requirements. Routes
may be optimised, fuel usage reduced and total fleet efficiency improved.
ve
2. Condition Monitoring: During travel, sensors installed on vehicles or containers can
detect factors such as temperature, humidity and vibration. This guarantees that
perishable or sensitive goods are transported under the best possible circumstances,
reducing spoiling and damage.
ni

3. Last-Mile Delivery Optimisation: IoT data can enable dynamic delivery route
rerouting and scheduling based on real-time traffic conditions, weather and customer
preferences. This improves delivery precision and client happiness.
U

4. Cargo Security: Tracking and monitoring solutions based on IoT improve cargo
security by sending alarms in the event of tampering or unauthorised access during
transit.
ity

5. Environmental Impact: By optimising routes, minimising idle time and supporting

sustainable practises, IoT-driven logistics can help decrease the environmental effect
of transportation.
Integration of IoT in supply chain management and logistics has various advantages,
including lower operational costs, increased efficiency and improved customer
m

experiences. IoT technology advancements include advanced automation, autonomous

cars, drones and predictive maintenance models. These developments propel the
transition of traditional supply chains into agile, responsive and efficient networks capable
)A

of meeting the demands of a connected society.

5.2 IoT Analytics

IoT analytics is a powerful data analysis tool that processes and extracts valuable
(c

insights from the vast amounts of data collected by IoT devices. This technology plays a
significant role in the context of Industrial IoT (IIoT), where data is gathered from sensors
embedded in factory infrastructure, weather stations, smart meters, delivery vehicles and

Amity Directorate of Distance & Online Education

 Internet of Things 199

various machinery. Notably, IoT analytics finds applications in diverse sectors including
Notes

e
data centre management, retail and healthcare.

in
nl
O
ity
rs
[Link]
ve
IoT data shares similarities with big data, but the key distinction lies not only in the
volume of data but also in its diverse sources. The challenge is to effectively consolidate
this data from various origins into a cohesive and comprehensible format. This complexity
in data integration is where IoT analytics proves its value, although its implementation
can present challenges.
ni

5.2.1 Industrial and Commercial Networks

Industrial and commercial networks are critical components of the Internet of Things
U

(IoT) ecosystem, allowing the networking and communication of devices, sensors and
systems across industries. These networks serve as the foundation for data interchange,
automation and intelligent decision-making across industries, resulting in enhanced
ity

efficiency, production and creativity. Let’s have a look at industrial and commercial
networks in the context of IoT:

Industrial Networks
Interconnected systems in the manufacturing, energy and other industrial sectors
m

are referred to as industrial networks. IoT technologies have had a significant impact
on industrial processes, giving rise to the concepts of “Industry 4.0” and the “Industrial
Internet.”
)A

Key Features of Industrial Networks in IoT

1. Machine-to-Machine (M2M) Communication: Industrial Internet of Things (IoT)
networks provide seamless connectivity between machines, equipment and sensors.
This permits real-time data interchange for industrial process monitoring, control and
(c

optimisation.
2. Sensor Integration: Industrial networks include a variety of sensors (for example,
temperature, pressure and vibration) that collect data from equipment and

Amity Directorate of Distance & Online Education

 200 Internet of Things

environments. This information is essential for predictive maintenance, process

Notes optimisation and quality assurance.

e
3. Remote Monitoring and Control: Industrial networks powered by IoT enable remote

in
monitoring and control of machinery and operations. From a centralised place,
operators may access real-time data and make educated decisions, decreasing
downtime and enhancing operational efficiency.
4. Predictive Analytics: Data analytics and machine learning are used in industrial

nl
networks to predict equipment breakdowns, optimise maintenance schedules and
improve overall asset management.
5. Safety and Security: In industrial settings, ensuring the safety of personnel and assets

O
is critical. IoT networks can monitor dangerous situations, track staff movements and
discover anomalies in order to respond quickly.

Commercial Networks

ity
Retail, healthcare, hotels and smart buildings are all examples of commercial
networks. Integration of IoT into commercial networks improves consumer experiences,
resource management and sustainability.

Key aspects of commercial networks in IoT include

1.
rs
Smart Retail: Beacons, RFID and mobile apps are used by IoT-enabled retail
networks to manage inventory, analyse customer behaviour and provide
personalised shopping experiences. Smart shelving and checkout systems
ve
increase efficiency while decreasing stockouts.
2. Healthcare IoT: IoT networks connect medical devices, patient monitoring and
electronic health records in healthcare. This makes remote patient monitoring,
telemedicine and effective healthcare delivery possible.
ni

3. Smart Buildings: Commercial IoT networks help to optimise energy usage,

lighting, HVAC and security in smart building systems. Building managers can
remotely monitor and operate equipment to reduce energy use and improve
U

tenant comfort.
4. Hospitality: IoT networks improve hotel visitor experiences by allowing mobile
check-in, room customisation and smart amenities. Energy-saving measures
ity

and occupancy detection help to ensure sustainability.

5. Fleet Management: IoT networks in business logistics track truck locations, fuel
usage and driver behaviour. Real-time tracking and route optimisation boost
delivery efficiency while lowering operating expenses.
m

Benefits and Challenges

Industrial and commercial networks under the IoT paradigm provide several
advantages, such as better operational efficiency, cost savings, improved customer
)A

experiences and data-driven insights. These networks, however, present issues in terms
of interoperability, security, scalability and data privacy. For tapping the full potential
of IoT in industrial and commercial contexts, seamless integration, effective security
measures and efficient data management are required.
Industrial and commercial networks play critical roles in the Internet of Things
(c

ecosystem, altering traditional industries and company operations. In these networks,

the combination of connection, data analytics and automation drives innovation and
competitiveness, ushering in a new era of smart and interconnected systems.
Amity Directorate of Distance & Online Education
 Internet of Things 201

5.2.2 Common Machine Learning Models for IoT

Notes

e
The Internet of Things generates immense volumes of data through millions of
interconnected devices. This data serves as the fuel for IoT machine learning, a process

in
that extracts insights from the data. Machine learning within the IoT context involves
recognising patterns in historical data, constructing models to forecast future events and
behaviours. The synergy between IoT and machine learning is transformative, unveiling
insights that were previously concealed within the data. This synergy leads to quicker,

nl
more automated responses and enhanced decision-making capabilities. Leveraging
image, video and audio data, machine learning applied to IoT enables the anticipation of
upcoming trends, identification of anomalies and the enhancement of overall intelligence.

O
Why use machine learning for IoT?
The utilisation of machine learning in IoT is motivated by its ability to decipher
intricate patterns within extensive datasets through advanced algorithms. By employing

ity
machine learning, the analysis of vast amounts of IoT data becomes possible, revealing
concealed insights. This approach facilitates the automation of significant processes by
integrating statistically derived actions, thereby improving or even substituting manual
operations.

rs
Sample use Cases
In the realm of IoT, businesses are harnessing machine learning to enable predictive
capabilities across diverse use cases. This strategic application empowers companies to
ve
gain novel insights and elevate their automation capabilities.
You may use machine learning for IoT to:
™™ Ingest and transform data into a consistent format
ni

™™ Create a machine learning model.

™™ Install this machine learning model on the cloud, the edge and the device.
For example, a corporation can use machine learning to automate quality inspection
U

and defect tracking on its assembly line, track asset activities in the field and estimate
consumption and demand patterns.
ity
m
)A
(c

[Link]
Figure: Machine Learning architectural overview

Amity Directorate of Distance & Online Education

 202 Internet of Things

Benefits of machine learning inference for IoT

Notes

e
Machine learning is a pivotal element within Software AG’s Cumulocity IoT platform,
a low-code, self-service solution. This comprehensive platform encompasses essential

in
components for rapid IoT implementation, including device connectivity, management,
application enablement, integration, streaming analytics and machine learning
model deployment. With deployment options spanning cloud, on-premises and edge
environments, Cumulocity IoT stands out as the sole platform supporting independent

nl
edge-only applications.

Key features of Cumulocity IoT Machine Learning include:

O
1. Effortless Model Training: Cumulocity IoT Machine Learning streamlines the creation
of new machine learning models. Leveraging AutoML, users can select optimal models
for their data—be it device data within the Cumulocity IoT platform or historical data
stored in big data repositories.

ity
2. Freedom of Data Science Libraries: Embracing flexibility, the platform accommodates
various data science libraries for model development (e.g., Tensorflow®, Keras,
Scikit-learn). Models can be crafted in the chosen framework, converted into industry-
standard formats and seamlessly integrated into Cumulocity IoT.
3.

rs
Swift Operationalisation: Models, whether built within Cumulocity IoT Machine Learning
or imported from other frameworks, can be deployed into production environments
with a single click. This deployment occurs wherever required, be it cloud-based or at
ve
the edge. Adaptable models allow swift modifications to account for changing patterns.
4. Datastore Connectivity: The platform facilitates data retrieval from operational and
historical datastores for model training. Cumulocity IoT Machine Learning ensures
data retrieval, conversion and model training are efficiently automated, integrating
ni

with Cumulocity IoT DataHub connectors for various data storage options.
5. Integration with Streaming Analytics: Cumulocity IoT Machine Learning collaborates
with Cumulocity IoT Streaming Analytics for real-time data scoring. The visual analytics
U

builder incorporates a “Machine Learning” component that employs preset models to

score real-time IoT data, seamlessly fusing machine learning with streaming analytics
workflows.
6. Notebook Integration: The integration with Jupyter Notebook—an industry-standard
ity

data science environment—enables users to interactively handle data, train models

and validate results. Cumulocity IoT Machine Learning harmonises with this open-
source web application for comprehensive functionality.
Cumulocity IoT’s Machine Learning capabilities empower users to swiftly create,
m

deploy and operationalise machine learning models across various data sources,
streamlining processes and fostering data-driven insights within IoT applications.
So, you have individualised models that have been trained across an army of
)A

IoT devices. This lends credence to the notion that a large number of simple models
outperform a single, more sophisticated model. Even if this is usually true, certain
individual models may perform below average due to inadequate or missing data for that
device. Here are some suggestions for expanding on this:
●● You may readily distinguish between particular models that perform well and
(c

models that perform poorly using the number of records provided in the training and
evaluation metrics. This data can be used to switch between a per-device model
and a model trained on the entire fleet.
Amity Directorate of Distance & Online Education
 Internet of Things 203

●● Training an ensemble model using per-device model predictions, fleet-wide model

Notes

e
predictions and metadata such as evaluation metrics and number of records per
device. This would result in a final prediction that would enhance the individual
models that were underperforming.

in
5.2.3 Exploratory Data Analysis
Exploratory Data Analysis (EDA) is a data analysis approach that involves visually

nl
and statistically exploring a dataset to understand its structure, characteristics and
potential patterns. EDA aims to uncover insights, relationships, anomalies and trends in
the data, which can inform further analysis and decision-making. In the context of IoT

O
analytics, EDA involves applying this approach to the vast and often complex datasets
generated by Internet of Things (IoT) devices and sensors.

Purpose of Exploratory Data Analysis in IoT Analytics

ity
The goal of exploratory data analysis in IoT analytics is multifaceted and fulfils
numerous critical goals:
1. Data Understanding: EDA assists analysts and data scientists in gaining a thorough
grasp of the IoT dataset. It enables them to appreciate the data’s nature, distribution

2.
and variables involved.

rs
Data Quality Assessment: Data quality concerns such as missing values, outliers and
inconsistencies can be identified via EDA. This is crucial for assuring the precision
ve
and dependability of subsequent analysis.
3. Pattern Discovery: EDA allows for the detection of hidden patterns, trends and
correlations in IoT data. It aids in the discovery of correlations and dependencies that
can be used to anticipate outcomes or make educated decisions.
ni

4. Anomaly Detection: EDA can help detect abnormalities or variations from expected
behaviour in IoT data. Outliers or unusual trends may indicate equipment breakdowns,
security breaches, or other noteworthy events.
U

5. Feature Selection: EDA assists the selection of significant features or variables for future
modelling and analysis in IoT analytics, where data dimensions can be considerable.
It aids in the reduction of dimensionality while keeping critical information.
ity

6. Hypothesis Generation: Hypotheses about prospective insights or research directions

are generated by EDA. These hypotheses can be used to lead more specific and
focused studies or experiments.
7. Visualisation: EDA relies heavily on visualisation to convey complicated IoT data
m

in intuitive and interpretable visual representations. Visualisations help to effectively

communicate findings and insights.
8. Decision Support: The outputs of EDA help decision-makers grasp the consequences
)A

of IoT data. It serves as a foundation for making educated decisions about resource
allocation, process optimisation and risk management.
9. Iterative Process: EDA is frequently an iterative process in which initial ideas lead
to additional exploration and refinement. As new patterns or abnormalities emerge,
(c

analysts can go deeper to uncover fresh insights.

Exploratory Data Analysis is critical in IoT analytics for uncovering hidden insights,
guaranteeing data quality and guiding subsequent analytical efforts. It enables analysts
Amity Directorate of Distance & Online Education
 204 Internet of Things

and organisations to make educated decisions, optimise processes and derive value from
Notes

e
massive and complicated datasets created by IoT devices and sensors.

Techniques for Exploratory Data Analysis

in
1. Descriptive Statistics: Basic statistical measurements including mean, median, mode,
standard deviation and range provide an introduction to data distribution and variability.
2. Data Visualisation: Data visualisations such as histograms, box plots, scatter plots

nl
and time series charts aid in the discovery of data patterns, trends and anomalies.
Time series visualisations are especially useful for understanding temporal patterns in
IoT data.

O
3. Correlation Analysis: Correlation analysis between variables can provide insight into
how they are related. Correlation matrices and heatmaps aid in the identification of
probable relationships.

ity
4. Outlier Detection: Understanding anomalies and potential data quality issues requires
identifying outliers in IoT data. Techniques such as the z-score, Tukey’s fences and
clustering-based procedures are available.
5. Dimensionality Reduction: IoT data frequently has several aspects. Principal
Component Analysis (PCA) and t-SNE are two techniques that can help decrease

6.
rs
dimensions for visualisation and analysis.
Time Series Analysis: Data from IoT devices usually reveal temporal trends. Time
series analysis techniques such as autocorrelation, trend decomposition and
ve
seasonality analysis aid in the discovery of long-term patterns.

Tools for Exploratory Data Analysis

1. Python: Python, coupled with modules like as Pandas, Matplotlib, Seaborn and Plotly,
ni

creates a rich ecosystem for data processing and visualisation.

2. R: R is yet another sophisticated statistical analysis and visualisation language. EDA
typically employs packages such as ggplot2 and dplyr.
U

3. Jupyter Notebooks: Jupyter notebooks enable for interactive and iterative data
exploration by combining code, visualisations and explanations in a single document.
4. Tableau: Tableau is a well-known data visualisation platform that allows for the
ity

interactive and intuitive examination of IoT data via drag-and-drop interfaces.

5. Power BI: For IoT analytics, Microsoft Power BI provides interactive dashboards and
reports for data visualisation and investigation.
6. Apache Zeppelin: Zeppelin is an open-source data analytics notebook that supports
m

various languages and can be used with Spark to handle large-scale IoT data.
7. Excel: Excel, while not as complex as other programmes, can nevertheless be useful
for basic EDA tasks, particularly for smaller datasets.
)A

Case Studies Showcasing Benefits in Industrial and Commercial Networks

Exploratory Data Analysis (EDA) is a crucial phase in data analysis that entails
visualising, summarising and comprehending the patterns, correlations and anomalies
contained within a dataset. EDA is critical in finding insights, optimising processes and
(c

making informed decisions in the context of industrial and commercial networks inside
the Internet of Things (IoT). Let’s look at several case studies that demonstrate the
advantages of EDA in industrial and commercial networks:

Amity Directorate of Distance & Online Education

 Internet of Things 205

Case Study 1: Predictive Maintenance in Manufacturing

Notes

e
Scenario: IoT sensors are used in a manufacturing plant to monitor the condition of
machines on the production line. The goal is to deploy predictive maintenance solutions

in
in order to reduce downtime and maintenance expenses.

Benefits of EDA
●● Anomaly Detection: The factory identifies odd patterns in machine behaviour,

nl
suggesting probable defects, by applying EDA on past sensor data.
●● Downtime Reduction: EDA aids in the prioritisation of maintenance activities
depending on the severity of abnormalities, hence decreasing unplanned downtime

O
and optimising maintenance schedules.
●● Cost Savings: Predictive maintenance reduces the frequency of unnecessary
maintenance procedures, resulting in lower parts and labour costs.

ity
Case Study 2: Customer Behaviour Analysis in Retail
Scenario: To better understand client behaviour in their stores, a retail chain
incorporates IoT devices and customer tracking technologies.

rs
Benefits of EDA
●● Foot Traffic Analysis: Peak shopping hours, popular store sections and consumer
flow patterns are revealed by EDA, allowing for optimised store design and
personnel.
ve
●● Conversion Rate Insights: EDA discovers factors impacting conversion rates by
linking foot traffic with sales data, enabling tailored marketing and promotion efforts.
●● Product Placement Optimisation: EDA assists in determining appropriate product
ni

positioning to maximise product visibility and sales.

Case Study 3: Healthcare Data Analytics

Scenario: A healthcare facility uses IoT-enabled medical devices to remotely monitor
U

patient health data.

Benefits of EDA
●● Patient Monitoring: EDA assists healthcare providers in analysing patient data
ity

trends, allowing for prompt treatments for at-risk patients.

●● Treatment Effectiveness: EDA assesses the effectiveness of various interventions
by comparing patient data with treatment plans, resulting in improved patient
outcomes.
m

●● Resource Allocation: EDA aids in resource allocation optimisation by identifying

high-demand periods and assigning employees and resources accordingly.
Exploratory Data Analysis is used in each of these case studies to deliver useful
)A

insights and actionable information for optimising operations, improving decision-making

and generating efficiency improvements in industrial and commercial networks within
the IoT ecosystem. EDA serves as a basic step in guiding organisations to data-driven
strategies and informed decisions.
(c

5.2.4 Predictive Analysis

Predictive analytics is a type of data analytics that uses historical and real-time

Amity Directorate of Distance & Online Education

 206 Internet of Things

data to estimate future outcomes, trends, or events. Predictive analysis in the context
Notes

e
of Internet of Things (IoT) analytics is leveraging data acquired from IoT devices,
sensors and systems to generate predictions about future behaviours, circumstances,
or occurrences. Predictive analysis in IoT is designed to deliver actionable insights that

in
enable proactive decision-making, optimise operations and improve outcomes. Let’s look
at the definition and application of predictive analytics in IoT analytics:
Predictive analysis in IoT analytics entails using advanced statistical and machine

nl
learning approaches to find patterns, connections and trends in historical and present
data from IoT devices. These insights are then utilised to build models that can accurately
anticipate the outcomes of future occurrences.

O
Purpose in IoT Analytics
The goal of predictive analysis in IoT analytics is multidimensional, with several
primary goals:

ity
1. Proactive Maintenance: By analysing data from IoT sensors, predictive analysis
can predict equipment breakdowns or maintenance requirements. This enables
organisations to schedule maintenance tasks ahead of time, lowering downtime and
expenses.
2.

rs
Anomaly Detection: Predictive analysis can detect abnormalities in IoT data by
detecting deviations from usual behaviour. These anomalies could suggest security
breaches, malfunctions, or other serious issues.
ve
3. Optimised Resource Management: Based on predicted demand and usage patterns,
predictive models can assist organisations in optimising resource allocation, such as
energy consumption, inventory levels and workforce scheduling.
4. Demand Forecasting: Predictive analysis in retail and supply chain management
ni

leverages IoT data to forecast product demand, allowing for more effective inventory
management and production planning.
5. Health Monitoring and Personalisation: Predictive analysis in healthcare can forecast
U

patient health trends and customise treatment regimens based on IoT-collected health
data.
6. Risk Assessment and Fraud Detection: Predictive analysis in banking and insurance
helps assess risk and identify potential fraud by analysing IoT data for unexpected
ity

trends or actions.
7. Quality Control: Predictive analysis in manufacturing enhances product quality by
spotting probable flaws or variations from conventional operations using IoT data.
8. Customer Behaviour Prediction: Based on IoT data, predictive analysis may estimate
m

user preferences and behaviours, assisting in targeted marketing and improving

customer experiences.
)A

Benefits of Predictive Analysis in IoT Analytics

●● Proactive Decision-Making: Predictive analysis enables organisations to take
proactive measures in reaction to predicted future occurrences, minimising reactive
responses and increasing overall efficiency.
●● Cost Savings: Predictive analysis saves money on operations and maintenance by
(c

forecasting repair needs, optimising resource allocation and minimising downtime.

●● Enhanced Customer Experiences: Predictive analytics powered by IoT enables

Amity Directorate of Distance & Online Education

 Internet of Things 207

businesses to personalise services and goods, increasing customer pleasure and

Notes

e
loyalty.
●● Improved Risk Management: Predictive analysis assists organisations in mitigating

in
risks by spotting possible difficulties before they become serious ones.
●● Informed Strategy Formulation: Strategic planning is guided by predictive insights,
allowing organisations to match their operations with future trends and expectations.

nl
Predictive analysis in IoT analytics uses data to forecast future scenarios and
results. Organisations may optimise operations, make informed decisions and improve
overall performance and competitiveness by exploiting IoT-generated data. The capacity
to predict future trends and events is a critical driver of IoT innovation and success.

O
Techniques and Algorithms for IoT Analytics
Predictive analysis is the process of combining historical and real-time data to make
educated guesses about future occurrences or outcomes. Predictive analysis is critical

ity
in Internet of Things (IoT) analytics for projecting trends, optimising processes and
improving decision-making. Let’s look at some of the methodologies and algorithms that
are often utilised in predictive analytics for IoT analytics:

Techniques for Predictive Analysis in IoT

1.

rs
Time Series Analysis: Predictive analysis frequently employs time series data from IoT
sensors. Moving averages, exponential smoothing and the autoregressive integrated
moving average (ARIMA) model, for example, can estimate future values based on
ve
historical trends.
2. Regression Analysis: Regression models examine variable relationships in order to
predict a target variable. To forecast numerical outcomes such as energy consumption
or equipment performance, linear regression, multiple regression and polynomial
ni

regression are utilised.

3. Classification: Categorical outcomes are predicted using classification algorithms.
This can include forecasting occurrences such as equipment breakdown or classifying
U

products based on sensor data in IoT. Decision trees, random forests and support
vector machines (SVM) are examples of common algorithms.
4. Clustering: Clustering techniques aggregate together similar data elements. This can
be used for consumer segmentation, anomaly detection and discovering trends in
ity

IoT data. K-means and hierarchical clustering are two approaches that are commonly
employed.

Real-World Applications in Industrial and Commercial Networks

m

Predictive analysis is a strong data analysis technique that makes predictions about
future occurrences or outcomes using historical data and statistical algorithms. Predictive
analysis is critical in the context of industrial and commercial networks within the Internet
of Things (IoT) for optimising operations, minimising downtime and boosting overall
)A

operational efficiency. Let us look at some real-world examples of predictive analysis in

industrial and commercial networks:

Real-World Applications in Industrial Networks

1. Predictive Maintenance in Manufacturing
(c

™™ Scenario: IoT sensors are used in manufacturing plants to monitor equipment

status and performance.

Amity Directorate of Distance & Online Education

 208 Internet of Things

™™ Application: Predictive analysis identifies trends in prior sensor data that

Notes

e
indicate impending equipment breakdowns. Maintenance teams are sent alerts
and recommendations, allowing them to do maintenance jobs before they fail.

in
™™ Benefits: Reduced downtime, optimised maintenance schedules, extended
equipment lifespan and cost savings.
2. Energy Consumption Forecasting in Utilities

nl
™™ Scenario: Energy usage patterns in large-scale industrial facilities are
monitored by utility companies.
™™ Application: Predictive analysis models anticipate future use by analysing

O
historical energy consumption as well as external factors (such as weather and
manufacturing schedules). This assists utility companies in efficiently planning
energy generation and distribution.
™™ Benefits: Improved energy allocation, reduced waste, optimised energy

ity
purchasing and enhanced sustainability.
3. Quality Control in Manufacturing
™™ Scenario: IoT sensors are used by manufacturers to check product quality and
manufacturing operations.
™™ Application:

rs
Predictive analysis identifies links between manufacturing
parameters and product faults. Manufacturers can alter processes in real time
to ensure consistent quality.
ve
™™ Benefits: Product quality was improved, waste was reduced, rework was
reduced and customer satisfaction was increased.

Real-World Applications in Commercial Networks

ni

1. Inventory Management in Retail

™™ Scenario: Retailers manage enormous product inventories across several sites.
™™ Application: Predictive analysis forecasts product demand by using previous
U

sales data, seasonal trends and external influences (e.g., holidays, promotions).
Retailers optimise inventory levels and replenishment procedures accordingly.
™™ Benefits: Stockouts were reduced, overstock problems were minimised,
inventory turnover was efficient and revenue was increased.
ity

2. Customer Churn Prediction in Telecom

™™ Scenario: Telecommunications firms aim to keep their consumers and reduce
customer churn.
™™ Application: Predictive analysis examines consumer usage patterns, service
m

history and interactions to detect churn factors. Following that, businesses can
develop targeted retention efforts.
™™ Benefits: Customer retention rates are higher, customer satisfaction is higher
)A

and revenue is higher.

3. Fraud Detection in Financial Services
™™ Scenario: Financial institutions keep an eye out for fraudulent transactions.
™™ Application: Models of predictive analysis Learn from past fraud cases
(c

and apply what you’ve learned to real-world transactions. Anomalies and

questionable patterns generate alarms for more research.

Amity Directorate of Distance & Online Education

 Internet of Things 209

™™ Benefits: Financial losses due to fraud were reduced, security was improved
Notes

e
and customer trust was increased.
4. Occupancy Prediction in Smart Buildings

in
™™ Scenario: Building managers strive to maximise energy efficiency and occupant
comfort.
™™ Application: Predictive analysis anticipates building occupancy using previous

nl
occupancy data, weather forecasts and event schedules. HVAC and lighting
systems change as needed to reduce energy waste.
™™ Benefits: Energy savings, increased passenger comfort and less environmental

O
impact.
Predictive analysis in industrial and commercial networks enables organisations to
make educated decisions, optimise resources and achieve higher levels of efficiency in
these real-world applications. Predictive analysis transforms data into actionable insights

ity
by integrating IoT-generated data and advanced analytics, resulting in superior outcomes
across multiple industries.

5.2.5 Big Data Platforms - Hadoop, Storm and Spark

rs
ve
ni

[Link]

Big Data platforms such as Hadoop, Storm and Spark are powerful tools for
U

processing, analysing and extracting insights from massive amounts of data generated
by Internet of Things (IoT) devices. These platforms handle the issues of dealing with
large datasets and real-time streaming data, making them critical components of IoT
analytics. Let’s look at Hadoop, Storm and Spark in the context of IoT analytics:
ity

Hadoop
Hadoop is an open-source system for storing and processing big datasets on
commodity technology in distributed clusters. The Hadoop Distributed File System
(HDFS) is used for data storage and the MapReduce programming model is used for
m

distributed data processing.

Role in IoT Analytics: Hadoop is an excellent choice for batch processing of
historical IoT data. It is capable of storing and managing massive amounts of sensor data
)A

generated over time. Hadoop’s features help IoT analytics application cases involving
data aggregation, batch processing and complicated transformations.

Storm
Apache Storm is a real-time stream processing framework that allows high-velocity,
(c

real-time data streams to be processed. It provides complicated event processing and

has low latency data processing.

Amity Directorate of Distance & Online Education

 210 Internet of Things

Role in IoT Analytics: Storm is suited for real-time IoT analytics, where data must be
Notes

e
processed in real time. Storm’s features can be used in IoT applications that require quick
replies, event detection and real-time notifications. Real-time sensor data monitoring,
anomaly identification and predictive maintenance are some examples.

in
Spark
Apache Spark is a flexible distributed computing platform that may be used for batch

nl
processing, interactive queries, real-time streaming and machine learning. It supports in-
memory data processing to provide speedier analytics.
Role in IoT Analytics: Spark is ideal for a wide variety of IoT analytics jobs. It is

O
capable of batch and real-time processing, making it ideal for IoT use cases requiring
flexibility. Spark’s machine learning library (MLlib) is useful for IoT data predictive
analytics, anomaly detection and pattern recognition.

ity
Comparison and Use Cases
Hadoop: Hadoop is best suited for dealing with massive amounts of historical data
and batch processing. Historical data analysis, data warehousing and data archival for
compliance are examples of use cases.

rs
Storm: Storm excels at real-time streaming data processing with low latency. It is
perfect for real-time monitoring, fraud detection and dynamic pricing in IoT scenarios that
require quick answers.
ve
Spark: Spark offers a unified platform for IoT analytics workloads ranging from
real-time streaming to batch processing and machine learning. It is appropriate for
applications such as predictive maintenance, sentiment analysis and recommendation
systems.
ni

Features and Capabilities

Big Data platforms such as Hadoop, Storm and Spark offer robust tools for
processing and analysing massive amounts of data. Each platform offers unique
U

characteristics and capabilities that make it appropriate for various use cases in industrial
and commercial networks.

Hadoop Features and Capabilities

ity

1. Distributed Storage: HDFS separates data into blocks, which are then distributed
across a cluster of commodity hardware, ensuring fault tolerance and scalability.
2. Batch Processing: MapReduce facilitates parallel data processing by dividing jobs
into smaller subtasks that are distributed across cluster nodes.
m

3. Scalability: The distributed architecture of Hadoop enables for easy scaling by adding
more nodes to the cluster.
4. Data Locality: The data localisation principle of Hadoop ensures that calculations are
)A

done on nodes where the data is stored, decreasing network overhead.

5. Ecosystem: Hive (SQL-like querying), Pig (data processing) and HBase (NoSQL
database) are among the many tools and libraries available in the Hadoop ecosystem.

Storm Features and Capabilities

(c

1. Real-Time Processing: Storm handles real-time streaming data, making it suited for
applications such as real-time analytics, fraud detection and event processing.

Amity Directorate of Distance & Online Education

 Internet of Things 211

2. Scalability: Storm can handle increasing data volumes and processing needs by
Notes

e
scaling horizontally.
3. Reliability: Storm ensures data reliability by processing each item of data at least

in
once and implementing fail-safe procedures.
4. Complex Event Processing: Storm can find patterns and correlations in streaming
data thanks to its capabilities for advanced event processing.

nl
5. Integration: Storm may be connected with a wide range of data sources, including
messaging systems such as Apache Kafka and it supports a wide range of programming
languages.

O
Spark Features and Capabilities
1. In-Memory Processing: Spark keeps data in memory, allowing for faster processing
than older disk-based systems such as Hadoop MapReduce.

ity
2. Unified Platform: Spark SQL (structured data), Spark Streaming (real-time data) and
MLlib (machine learning) all have a common API that may be used for a variety of data
processing applications.
3. Interactive Analytics: Spark supports interactive data querying and processing,

rs
making it ideal for ad hoc exploration and visualisation.
4. Machine Learning: The MLlib package in Spark provides tools for developing and
deploying machine learning models at scale.
ve
5. Graph Processing: Spark’s GraphX library can process and analyse graphs, making it
helpful for applications like as social network analysis and recommendation systems.
6. Ease of Use: Spark supports a variety of programming languages, provides interactive
shells for experimentation and has simple APIs.
ni

Hadoop, Storm and Spark are powerful big data platforms that each have their
own set of features and capabilities. Hadoop excels at batch processing and distributed
storage, Storm is built for real-time stream processing and Spark provides a flexible
U

framework for a variety of data processing jobs such as in-memory analytics, interactive
querying and machine learning. The platform used is determined by the specific needs
of industrial and commercial networks, such as data volume, processing speed and
ity

analytical requirements.
Hadoop, Storm and Spark: Use Cases and Benefits in Industrial and Commercial
Networks
Big Data systems such as Hadoop, Storm and Spark play an important role in
m

handling and processing huge amounts of data created by industrial and commercial
networks as part of the Internet of Things (IoT). Each platform has its own set of
characteristics that cater to varied data processing requirements.
)A

5.2.6 Usage of Databases

Databases are critical in IoT analytics because they provide a structured and
organised mechanism to store, manage and retrieve vast amounts of data created by IoT
devices and networks. For numerous reasons, databases are critical in IoT analytics:
(c

1. Data Storage and Management: Databases serve as a centralised repository for

handling and organising massive amounts of data generated by IoT devices and

Amity Directorate of Distance & Online Education

 212 Internet of Things

networks. They provide data integration by allowing different types of data to be kept
Notes

e
together, allowing for cross-domain analysis and insights.
2. Scalability: IoT environments necessitate databases that can handle increasing data

in
volumes without sacrificing performance. Scalable databases can handle increasing
data volumes without sacrificing performance.
3. Real-Time Access: For real-time decision-making, IoT analytics require prompt access

nl
to data and databases provide mechanisms for quick data retrieval and processing.
4. Query and Analysis: Databases allow users to execute complex operations on IoT
data, such as filtering, aggregation and correlation, to generate valuable insights.

O
5. Historical Data: Databases keep historical data for lengthy periods of time, allowing
for trend analysis, pattern identification and predictive modelling.
6. Data Consistency and Reliability: Databases enforce constraints and provide
transaction management and rollback tools to protect data integrity.

ity
7. Security and Privacy: To safeguard data, databases provide security features such as
encryption, access limits and user authentication.
8. Machine Learning and Advanced Analytics: Machine learning models can be hosted
in databases, allowing predictive and prescriptive analytics to be performed directly on
stored data.

rs
10. Regulatory Compliance: Databases make data governance and audit trails easier in
order to meet regulatory standards.
ve
11. Visualisation and Reporting: Databases lay the groundwork for integrating visualisation
tools and presenting information to stakeholders.
12. Data Retention and Archiving: Databases enable effective data retention and archiving
procedures, ensuring critical data is maintained for historical analysis and compliance.
ni

Databases serve as the foundation of IoT analytics, providing the infrastructure

required to store, manage and retrieve data for analysis and decision-making. They
enable businesses to capitalise on the value of IoT-generated data, unearth actionable
U

insights and drive innovation across industrial and commercial networks.

Types of Databases Commonly Used in IoT Analytics

ity

Databases play an important role in IoT analytics because they store, handle and
retrieve massive amounts of data created by networked devices and sensors. To manage
the varying data requirements of IoT analytics, various types of databases are routinely
employed. Let’s look at some of the databases that are often utilised in IoT analytics:
1. Relational Databases: Relational databases store data in tables with rows and
m

columns in an organised way. They work well with structured and well-defined data,
making them ideal for managing IoT data that adheres to a consistent format.
)A

Use Cases
™™ Sensor Data Management: Sensor readings, timestamps and associated
metadata can be stored in an organised way in relational databases.
™™ Historical Data Analysis: Relational databases are useful for storing historical
data that needs to be queried and analysed, such as temperature trends or
(c

energy usage patterns.

2. Time-Series Databases: Time-series databases are designed to store and retrieve

Amity Directorate of Distance & Online Education

 Internet of Things 213

time-stamped data points. They are optimised for rapid inserts and time-based queries,
Notes

e
making them excellent for managing IoT data with constant time-related metrics.

Use Cases

in
™™ IoT Sensor Data: Time-series databases are capable of storing and retrieving
sensor readings, device statuses and events over time in an efficient manner.
™™ Real-Time Analytics: Time-series databases are ideal for analysing flowing data

nl
in real time, such as monitoring equipment performance or tracking changes in
climatic conditions.
3. NoSQL Databases: NoSQL databases provide flexible and schema-less data storage,

O
making them ideal for unstructured or semi-structured data seen in IoT applications.

Use Cases
™™ Unstructured Data: NoSQL databases may hold a wide range of data formats,

ity
such as photos, videos, text and JSON documents, allowing IoT systems to
handle a wide range of data sources.
™™ IoT Data Streams: Because NoSQL databases allow for real-time data
ingestion and analysis, they are useful for applications such as social media
sentiment analysis and IoT-generated event streams.
4.

rs
Graph Databases: Graph databases store data in the form of nodes and edges, making
them suitable for expressing and querying complex interactions and dependencies
that are common in IoT networks.
ve
Use Cases
™™ Network Topology: Graph databases can simulate the links between
IoT devices, sensors and gateways, making network optimisation and
troubleshooting easier.
ni

™™ Supply Chain Tracking: Graph databases can be used to describe the flow of
commodities, suppliers and distribution networks, hence improving supply chain
visibility and management.
U

5. In-Memory Databases: In-memory databases store data in RAM for faster access,
making them ideal for real-time processing and analytical applications.

Use Cases
ity

™™ Real-Time Decision-Making: Low-latency data retrieval is supported by in-

memory databases, providing real-time analytics for time-critical applications
such as fraud detection or dynamic pricing.
™™ IoT Data Fusion: In-memory databases may aggregate data from various
m

sources in real time, allowing businesses to acquire complete insights from a

variety of IoT data streams.
The type of database used for IoT analytics is determined by the nature of the
)A

data, the needed query performance and the IoT application’s specific use cases.
Organisations may easily manage and gain insights from the massive and varied data
generated by IoT devices and sensors by selecting the proper database type.

Summary
(c

●● In the future, a vast multitude of physical devices, embedded with intelligent

sensors, will seamlessly interact with each other on a Machine-to-Machine basis,
operating without the need for human intervention.

Amity Directorate of Distance & Online Education

 214 Internet of Things

●● IoT technology plays a pivotal role in ensuring road safety through systems like
Notes

e
collision detection, lane change alerts, smart traffic signal control and intelligent
traffic scheduling protocols.

in
●● The National Science Foundation developed a collaborative industry/university
research centre on intelligent maintenance systems (IMS).The goal is to achieve
near-zero breakdown through IoT-based manufacturing.
●● Through the utilisation of discrete wearable sensors and advanced machine

nl
learning algorithms, an automated system was developed to identify and
eliminate outlier points caused by movement artefacts in blood pressure (BP) and
photoplethysmogram (PPG) signals.

O
●● The rapid evolution of wearable technology has been propelled by advancements in
energy-efficient mobile networks, miniaturisation of electronic devices and sensors
and the advantageous capabilities offered by smart wearables.

ity
●● Gamers receive performance insights and guidance to enhance their skills.
Messages are conveyed through wearable screens, displays, audio, or haptic
indicators, fostering better understanding of performance quality and improvement
strategies.
●● OTDOA calculates the time difference of arrival of Positioning Reference Signals

rs
(PRSs) from multiple nodes relative to a reference node’s transmission. This data is
converted into a geographical hyperbola, with the device’s position approximated at
the intersection point of hyperbolas.
ve
●● Lux sensors play a pivotal role in gauging luminance and hold the potential to
initiate an array of actions, spanning from corroborating motions to activating lights
in response to diminishing light levels.
●● When it comes to selecting a protocol for smart home development, it’s crucial to
ni

begin by identifying the specific needs of the project and then refining the options.
The commonly utilised protocols include Bluetooth Low Energy, Z-Wave, Zigbee and
Thread.
U

●● IoT technology advancements include advanced automation, autonomous cars,

drones and predictive maintenance models. These developments propel the
transition of traditional supply chains into agile, responsive and efficient networks
capable of meeting the demands of a connected society.
ity

●● Predictive analysis is the process of combining historical and real-time data to

make educated guesses about future occurrences or outcomes. Predictive analysis
is critical in Internet of Things (IoT) analytics for projecting trends, optimising
processes and improving decision-making.
m

●● Predictive analysis is critical in the context of industrial and commercial networks

within the Internet of Things (IoT) for optimising operations, minimising downtime
and boosting overall operational efficiency.
)A

●● Big Data platforms such as Hadoop, Storm and Spark are powerful tools for
processing, analysing and extracting insights from massive amounts of data
generated by Internet of Things (IoT) devices.
●● Databases play an important role in IoT analytics because they store, handle and
retrieve massive amounts of data created by networked devices and sensors. To
(c

manage the varying data requirements of IoT analytics, various types of databases
are routinely employed.

Amity Directorate of Distance & Online Education

 Internet of Things 215

Glossary
Notes

e
●● Collaborative: produced or conducted by two or more parties working together
●● E-maintenance: e- maintenance is an emerging concept generally defined as “a

in
maintenance management concept whereby assets are monitored and managed
over the Internet.”
●● Multimodal: characterised by several different modes of activity or occurrence.

nl
●● Physiological: relating to the branch of biology that deals with the normal functions
of living organisms and their parts.
●● Accelerometer: an instrument for measuring acceleration, typically that of an

O
automobile, ship, aircraft, or spacecraft, or that involved in the vibration of a
machine, building, or other structure.
●● Leveraging: using borrowed capital for (an investment), expecting the profits made
to be greater than the interest payable.

ity
●● Inventory: a complete list of items such as property, goods in stock, or the contents
of a building.

Check Your Understanding

rs
1. IoT devices can also take charge of controlling crucial infrastructure elements, like
____________.
a) Airplanes
ve
b) Bridges that facilitate ship docking
c) Roads and highways
d) Parks and gardens
ni

2. Creating a stress-management strategy involving self-monitoring becomes imperative

in adapting to the advancements in the _____________.
a) Financial planning
U

b) Agricultural practices
c) Social media engagement
d) Internet of Medical Things (IoMT).
ity

3. A portable IoT device, embedded in a smart vest, constantly monitors the respiratory
rate of Chronic Obstructive Pulmonary Disease (COPD) patients during ____________.
a) Exercise equipment
b) Food delivery services
m

c) Traffic signals
d) Rest periods
4. By curbing power consumption and embracing energy harvesting, the endurance of
)A

____________ can be significantly prolonged.

a) Household appliances
b) Cars
c) Wearable IoT devices
(c

d) Musical instruments

Amity Directorate of Distance & Online Education

 216 Internet of Things

5. A smart wheelchair with smartphone interaction allows remote monitoring by carers.

Notes

e
Stroke patients benefit from a rehabilitation approach that merges ____________.
a) Cleaning robots

in
b) Self-driving cars
c) Smart refrigerators
d) Gaming and wearables

nl
6. Tracking daily physical activities and body movements, including actions like walking,
running and sitting, is a primary application of IoT ____________.

O
a) Cooking techniques
b) Musical compositions
c) In sports

ity
d) Literature analysis
7. Rapid detection of falls is essential, requiring the system to discern between falls and
routine activities while ____________.
a) Cooking recipes

rs
b) Stock market trends
c) Minimising false positives
d) Movie reviews
ve
8. A modified safety smart helmet captures data and transmits it to a local server, where
a random forest classification algorithm determines the severity of the ____________.
a) Hair styling techniques
ni

b) Gardening practices
c) Sleep situation
d) Physical therapy exercises
U

9. Incorporating smart devices like thermostats, light bulbs and security cameras into
IoT-enabled ____________ is commonplace.
a) Baking methods
ity

b) Construction techniques
c) Home automation
d) Astrophysics research
10. Platform developers are progressively expanding access to their ____________ for
m

developers.
a) Airport construction
b) Libraries
)A

c) Automotive design
d) Textile manufacturing
11. Float level sensors and other integrated circuits, such as the LM1830, provide IoT
developers with more precise ____________.
(c

a) Weather forecasts
b) Measuring capabilities
Amity Directorate of Distance & Online Education
 Internet of Things 217

c) Networking skills
Notes

e
d) Music composition techniques
12. The integration of IoT and AI has enormous promise for the future of ____________.

in
a) Cooking recipes
b) Environmental conservation

nl
c) Industrial manufacturing
d) Automation
13. In the Internet of Things, autonomous decision-making involves devices and systems

O
making judgments based on ____________ without human interaction.
a) Random factors
b) Instincts

ity
c) Data and predetermined criteria
d) Emotional responses
14. Real-time analytics entails processing and analysing data in ____________, allowing
for immediate insights and informed decision-making.
a)
b)
c)
Slow motion
Batches
Real time
rs
ve
d) The past
15. The capacity of predictive maintenance and real-time data to drive efficiency, decrease
costs, improve safety and alter industries across the spectrum is the ____________.
ni

a) Historical trend
b) Present scenario
c) Future of IoT automation
U

d) Challenge of modern technology

16. IoT analytics finds applications in diverse sectors including ___________, retail and
healthcare.
ity

a) Space exploration
b) Financial management
c) Data centre management
d) Artistic expression
m

17. Interconnected systems in the manufacturing, energy and other industrial sectors are
referred to as ____________.
)A

a) Social networks
b) Industrial networks
c) Data highways
d) Entertainment platforms
(c

18. The synergy between IoT and machine learning is transformative, unveiling insights
that were previously ____________.

Amity Directorate of Distance & Online Education

 218 Internet of Things

a) Ignored
Notes

e
b) Overlooked
c) Concealed within the data

in
d) Available to everyone
19. With deployment options spanning cloud, on-premises and edge environments,
Cumulocity IoT stands out as the sole platform supporting ____________ applications.

nl
a) High-performance computing
b) Independent edge-only

O
c) Data storage
d) Cloud-based
20. Exploratory Data Analysis is critical in IoT analytics for uncovering hidden insights,

ity
guaranteeing data quality and guiding subsequent ____________ efforts.
a) Marketing
b) Financial
c) Analytical
d)

Exercise
Administrative

rs
ve
1. What are the three key applications that wearable IoT safety sensors primarily serve?
Explain.
2. What are the main characteristics of industrial networks within the context of the
Internet of Things (IoT)?
ni

3. Give a short note on Intelligent Supply Chain Management.

4. How can the teamwork of Internet of Things (IoT) and Artificial Intelligence (AI) change
industries, make businesses work in new ways and transform how we interact with the
U

world around us?

5. Explain shortly the usage of Databases. Also mention the types of Databases that are
generally Used in IoT Analytics.
ity

6. Elaborate on the techniques and tools used to carry out Exploratory Data Analysis.

Learning Activities
1. Evaluate your knowledge of the uses and advantages of IoT-enabled health monitoring
systems. Write a quick analysis of how IoT technology could completely transform
m

healthcare and wellbeing. Include your opinions on the difficulties and opportunities
that these systems bring.
2. Talk about the moral issues surrounding data security and privacy in IoT analytics.
)A

What role do databases play in ensuring data security and legal compliance? Think of
ways that organisations can make better decisions and spur innovation by utilising the
power of IoT-generated data housed in databases.

Check Your Understanding- Answers

(c

1. b) 2. d) 3. d) 4. c)
5. d) 6. c) 7. c) 8. c)

Amity Directorate of Distance & Online Education

 Internet of Things 219

9. c) 10. b) 11. b) 12. d)

Notes

e
13. c) 14. c) 15. c) 16. c)
17. b) 18. c) 19. b) 20. c)

in
Further Readings and Bibliography
1. “IoT Analytics: Opportunities for Business and Research” by Ajith Abraham, et al.

nl
2. “Designing for the Internet of Things” by Adrian McEwen and Hakim Cassimally
3. “Internet of Things: Principles and Paradigms” by Rajkumar Buyya, et al.
4. Borgia, E. (2014). The Internet of Things vision: Key features, applications and

O
open issues. Computer Communications, 54, 1-31.
5. Atzori, L., Iera, A., & Morabito, G. (2010). The Internet of Things: A survey.
Computer Networks, 54(15), 2787-2805.

ity
rs
ve
ni
U
ity
m
)A
(c

Amity Directorate of Distance & Online Education

 220 Internet of Things

Notes

e
in
nl
O
ity
rs
ve
ni
U
ity
m
)A
(c

Amity Directorate of Distance & Online Education