MDIT POLYTECHNIC COLLEGE, ULLIYERI

A SEMINAR REPORT

on

IOT BASED LOW COST SMART INDOOR FARMING MANAGEMENT

SYSTEM USING AN ASSISTANT ROBOT AND MOBILE APP

Submitted in partial fulfillment of the requirement for the award

of

Diploma in Computer Engineering.

Submitted by

ASHWIN V

REG NO:19132357

DEPARTMENT OF COMPUTER ENGINEERING

MDIT POLYTECHNIC COLLEGE, ULLIYERI P.O, KOZHIKODE – 673620

MARCH 2022
 DEPARTMENT OF COMPUTER ENGINEERING MDIT
POLYTECHNIC COLLEGE, ULLIYERI

CERTIFICATE

This is to certify that the report of the seminar entitled “IOT BASED LOW-COST SMART

INDOOR FARMING MANAGEMENT SYSTEM USING AN ASSISTANT ROBOT AND

MOBILE APP” is a bonafide record of the work done by ASHWIN V (REG NO:19132357) in

partial fulfillment of the requirement for the award of Diploma in Computer Engineering from the

Board of Technical Education, Kerala.

GUIDED BY SEMINAR CO ORDINATOR HOD

INTERNAL EXTERNAL
 ACKNOWLEDGEMENT

I would like to thank GOD ALMIGHTY for giving me an opportunity to excel in my

efforts to complete this seminar on time.

I express my gratitude to Principal Mr. Basheer K, for making resources available at

right time without this seminar would not have been a success.

I am grateful to Ms. Ummu Haniya P, Head of the Department of Computer

Engineering for the cordial support and guidance in various stages of this task.

I express my gratitude to my Seminar Guide Ms. Pranisha T.O, Lecturer in Department

of Computer Engineering for her guidance and constant encouragement.

I express my thanks to all staff members and friends for all the help and coordination
extended in bringing out this seminar successfully in time.

I will be failing in duty if I do not acknowledge with grateful thanks to the authors of
the references and other literature referred to in this seminar report.

Last but not the least; I am very much thankful to my parents who guided me in every
step which I took.

Place: ULLIYERI
Date: ASHWIN V

i
 ABSTRACT

As the population grows and the quality of life of the people improves, leading to heightened

demand for salubrious food. As a result, indoor farming has become a very popular day by day

and the use of technology in agriculture is noteworthy. The most beneficial feature of indoor

farming is that it can produce healthy food using less farmland and workers. In this paper, a system

will be acquainted through which it is possible to manage an indoor farm automatically at a very

low cost.

Whereby it is possible to water the farm plants when required, provide specific light to

each plant for photosynthesis, constrain the concentration of CO2 on the farm, which is suitable

for the plants, etc. In addition, the whole arrangement can be managed from any location through

a mobile app. This system comprises an ancillary robot that provides fertilizer to the farm's plants

and real-time monitoring of the entire farm. The robot permits the user to define its task in advance

or users can give instructions from any place at any time by using the app.

ii
 CONTENTS

Contents Page No:

ACKNOWLEDGEMENT i

ABSTRACT ii

LIST OF FIGURES v

ABBREVIATIONS vi

Chapter 1 INTRODUCTION 1
Chapter 2 PROBLEM DEFINITION 3
Chapter 3 PROPOSED SYSTEM 4

Chapter 4 METHODOLOGY 6
4.1 Background Study 6
4.2 IoT in Agriculture 7
4.2.1 Overview 7
4.2.2 Context 8
4.2.3 Edge Computing 10
4.3 AI in Agriculture 10
4.3.1 Overview 10
4.3.2 Anomaly Detection 11
4.3.3 Type of Anomalies 12
4.3.4 Yield Prediction 13

iii
 4.4 System Description 14
4.5 System Overview 18
4.5.1 3D Simulation and System Prototype 18

Chapter 5 RESULT AND DISCUSSION 25

5.1 App Description 25

Chapter 6 CONCLUSION 30

REFERENCES

iv
 LIST OF FIGURES

No. Title Page No.

4.1 Flow diagram of the system 15

4.2 Process diagram of controling the robot through app 17

4.3 3D Simulation of the system 18

4.4 Prototype of the farm 19

4.5 Light intencity (a) low (b) high 20

4.6 (a) Pump ON and (b) serial monitor data 21

4.7 (a) Circuit and (b) sensor data 23

4.8 Prototype of the assistant robot 24

5.1 (a) App menu and (b) System control from app 26

5.2 Robot control interface 27

v
 ABBREVIATIONS

LED Light-emitting diode

NodeMCU Node MicrocController Unit

GSM Global System for Mobile communication

SMS Short Message service

IFM Indoor Farm Management

IOT Internet of Things

AI Artificial Intelligence

vi
 CHAPTER 1

INTRODUCTION

Indoor agriculture is a great implement to meet the burgeoning demands of the world's food. It not

only fulfills the demand for fresh foods but also plays a significant role in producing nutritious

foods. Enormous amounts of pesticides are used to protect the crop from pests. As a result, the

crops that are being produced every year are not nutritious and toxic-free, so we suffer in

malnutrition and chronic disease. We also have to depend on our environment when it comes to

food production. Global warming is causing climate change rapidly, which is pernicious for our

agriculture. Indoor agriculture is very important for bringing to an end such problems.

Advanced technology is being used to make indoor farming more efficient and easier. By

using technology, it is very easy to change the rate of food production. These include automatic

monitoring of plants, managing adequate light and ventilation, taking various necessary

arrangements with a robot, etc. It is estimated that the population of the world will be about nine

billion by 2050.

Conventional cultivation is not enough to satisfy the requirements. Therefore, indoor

agriculture has chosen to meet the demand for healthy food in developed countries as well as in

developing countries. Several people are doing this as a hobby and many are adopting this method

for commercial reasons. They are using their land, roof, or porch for indoor agriculture.

1
We are able to handle indoor farming automatically through our project. It automatically supplies

water or essential materials to the farm's crops. It is also able to control temperature when the room

temperature rises and provides the light needed for the farm's plants.

As well as, there is an especially proficient line follower’s robot use as an assistant for the

overall management of the entire farm. By which, the owner of the farm can monitor it from

anywhere and anytime.

Moreover, if you tell the robot when to give fertilizer in crops, the robot will be able to do

it automatically. In addition, it can take pictures of the whole farm and upload it to the server at

the specified time. The entire system can be controlled through an app.

2
 CHAPTER 2

PROBLEM DEFINITION

Indoor farms are most often located close to urban areas and have access to electricity and a

reliable internet connectivity. Availability, affordability (both by the cost of sensors and resource

optimization) and reliability of necessary technologies are the barriers to widespread adoption of

IoT devices, especially in a small-to-medium scale experimental indoor farm. Developing a

solution with off-the-shelf sensors, implementing valid error handling, alerts, remote monitoring,

and running indoor farm operations through deployment of smart logarithms that integrate

monitored data for decision making, has a potential to motivate growers in their wide-scale

adoption.

The rationale of the study and detailed description of the design and implementation has

potential to serve as an end-to-end reference for an interested grower or an IoT system developer.

The ability to obtain detailed information on plant growth and yield without employing destructive

techniques could be particularly interesting as it provides an example for using automation to

reduce reliance on labor during critical operations.Though the testbed was used for studying the

effect of lighting, nutrition solution concentration, seed density, and day of harvest on microgreens

production but it has potential to be used for conducting experiments with other plant species or

treatment combinations.

3
 CHAPTER 3

PROPOSED SYSTEM

As the system starts , the moisture sensor, gas sensor MQ135, humidity sensor, and LED lights

will be turned ON immediately. The moisture sensor will monitor the amount of water in the

container of the plant. If the amount of water in the tub is less than required, the moisture sensor

will send a signal to Arduino to turn the pump ON. Moreover, if the amount of water is sufficient,

the pump will stop. MQ135 has been used on this system for two reasons, firstly to know the

amount of carbon in the air and secondly to know the amount of smoke in the room. If the amount

of carbon is not between 400 and 600, then the ventilation process will start. On the other hand, if

MQ135 detects smoke, two things will happen simultaneously.

First, the main electrical supply of the room will be disconnected with the help of the relay

module and secondly, through the NRF24L01 transmitter, the whole system's data will be sent to

the assistant robot. Meanwhile, the system has LED lights to provide the light needed for

photosynthesis, which can be reduced or increased with the intensity requirement. Humidity sensor

senses the humidity of the farm, if the humidity is less than required then the system gives the

signal to spray the sprinkle of water to meet the demand.

The system will be able to exchange information with an assistant robot via NRF24L01.

This allows the system to be controlled by an assistant robot .A line follower robot has been used

for continuous monitoring of the entire system. This robot has a few special abilities. This robot

has an esp32 camera module, NodeMCU module, GSM, and NRF20L01 module.

4
The ESP32 camera allows users to watch real-time video of his farm. NodeMCU is used to

manage the robot remotely through the app. This will allow the user to know all updates of his

farm, control the robot, and give fertilizer to the plants. The robot has a fertilizer container, which

is controlled by a servo motor and can provide fertilizer to a specific plant. One unique feature of

the robot is that it can work automatically.

The robot will automatically inform the user about the status of the farm twice a day. To

do this, the robot will first collect the entire firm's data such as moisture sensor data, MQ135 data,

etc. Data from each sensor on the system will be transmitted via an NRF20L01 module.

Transmitted data will be received by the NRF20L01 module located on the robot and will be sent

to the user via GSM.

Apart from the NodeMCU module, also, GSM is used here because if there is a fire on the

farm and the user does not have an internet connection on his phone at the same time, he will get

the information via SMS on his phone. An important and unique feature of the system is that when

the robot finishes its work, it will go to its stoppage where its battery will be able to charge

automatically so that the user will not have to worry about taking charge of the robot's battery.

5
 CHAPTER 4

METHODOLOGY

4.1 BACKGROUND STUDY

Numerous work of indoor farming has been done by a lot of researchers using different methods.

The authors of paper have worked on crop management that focuses on indoor farming. Raspberry

Pi 3 module, FC-28 sensor, and DHT22 sensor are used on this system. The sensors check soil

moisture, humidity, and temperature, respectively. The system can operate the pump according to

the moisture sensor value. In addition, it can store humidity and temperature sensor’s real-time

data in Excel sheets.

Another paper's authors have introduced wireless technology for indoor farming. They used

some sensors, routers, and RF modules on the system, so they could monitor it wirelessly. In this

farming system, if the water is required to crops, then it will give water through the solenoid valve.

Another important factor is that the buzzer will be ringing if the firm's temperature and humidity

are not in the optimum condition.

The humidity of the stored rice can be monitored by the Blink application through the capsule

designed in the paper. The mobile app will graphically show the data of each capsule. The user

will be notified when the device goes offline.

6
A farming box for a small house was designed by the authors of the paper. This system is

controllable by mobile apps and computer software. Here the author's emphasis on three things

that are the box's temperature, humidity, and light intensity for the crop.

"Agrobot" is a farmer's assistant robot. By which, it is possible to automatically plow, sow,

and water the land. This work is done with the help of Servo Motors, Ultrasonic Sensors, Arduino

Uno, Water Pump, etc.

Taking ideas from the above studies, we have designed a system to increase food production

keeping in mind the nutritional value of food and the challenges of indoor farming. Which will

also make indoor farming easier.

4.2 IOT IN AGRICULTURE

4.2.1 Overview

Organizations and researchers have defined IoT systems in a multitude of forms based on

applications and assets a specific proponent wants to emphasize. One such definition provided

by OASIS (Open Protocols, 2014) describes IoT as: “System where the Internet is connected to

the physical world via ubiquitous sensors.” A high-level functional model provided by the

Alliance for Internet of Things Innovation (AIOTI) consists of 3 layers – Application, IoT, and

Network. Many researchers and technology companies have built on it. Multiple architectures

have been proposed from various standpoints like domain, service, function, etc. (Ray, 2018;

Verdouw et al., 2019). In the field of agro-industrial and environmental applications, a review of

72 studies suggested an architecture by incorporating commonly used and relevant

methodologies.

7
This IoT architecture contains physical, communication, service and application layer and was

adapted in the design of the testbed in this project. IoT devices (e.g., wireless sensor networks,

network-connected weather stations, cameras, and smartphones) can be used to collate a vast

amount of environmental and crop performance data, ranging from time-series data from sensors

to spatial data from cameras to human observations captured via mobile smartphone applications

or other network-connected user interfaces. Analysis of this data in real-time from a remote

location, off the farm, is possible because of internet connectivity. By extending the sensing system

to include actuators, necessary control tasks like irrigation, ventilation, misting etc. can be

performed based on input(s) from a rule-based system, an AI algorithm, or a user action.

4.2.2 Context

Numerous benefits of IoT devices have led to their increased adoption and the domain of

IoT has been enhanced by research activities persistently over the past few years. However, there

are several unsolved issues about proper management and performance of IoT devices. Some of

them are related to availability, reliability, scalability, mobility, data confidentiality, security,

optimization, and compatibility of networks, etc. (Elijah et al., 2018; Khanna and Kaur, 2019; M

et al., 2020). Several peer-reviewed studies also note that within an IoT environment, cost,

constraint-free communication, proper deployment, and heterogeneity are additional

challengesthat are to be addressed by further research (Elijah et al., 2018; Khanna and Kaur,

2019; Ummesalma et al., 2020).

8
A lot of literature exists about application protocols (Glaroudis et al., 2020) and connectivity

methods (like Wi-Fi, Satellite, Bluetooth, cellular) (Ahmad et al., 2019; Vannieuwenborg et al.,

2018). Data security and privacy issues are emerging as a domain that are primarily been of interest

to computer scientists and financial institutions across the world. To deal with availability and cost

of commonly available sensors/components, numerous solutions have been proposed (Danita et

al., 2019; Ruengittinun et al., 2017). However, there is a noticeable void in the literature on how

to improve and adapt IoT solutions for real-world indoor farms beyond simple prototypes. Direct

scaling in most cases implies that the number of hardware components grows very quickly with

the size of the system. Careful requirement analysis and system design are indispensable for a

viable IoT solution in a real-world scenario. Along with it, a series of checks in the system over

hardware and software are required to steer through the reliability challenges that are especially

associated with the use of low-cost sensors.

Data management is yet another aspect of IoT devices that calls for attention. Sensors and

devices can connect indirectly through the cloud, where data is centrally managed or utilize the

concept of edge computing wherein data is sent directly to other local devices to collect, store, and

analyze, and then share selected findings or information through the cloud. An Edge/Cloud

combination offers a particularly attractive approach to deliver an efficient workflow. However,

validation of edge-driven services on operational farms is limited (O’Grady et al., 2019). Along

with reaping the benefits of the technology, exploration of the paradigm of edge-computing in an

indoor farm is required to identify systemic challenges.

9
4.2.3 Edge Computing

In the context of IoT, with increasing number of devices and data in a network, traditional

cloud computing has limitations in terms of latency, bandwidth, round the clock accessibility and

update frequency of devices, and data security (Shi et al., 2019). Increased energy consumption

and scalability challenges at cloud servers have pushed the rapid growth of edge computing over

the past 4 years (Shi et al., 2019). Despite a rapid growth, the concept of edge computing is still in

its infancy (Varghese et al., 2016). A summary of motivation, challenges, and opportunities in

edge-computing (Varghese et al., 2016). A survey of 46 state-of-the-art research papers on the use

of edge computing for precision dairy, irrigation, fire detection, wildlife surveillance etc.

highlights utilization of techniques for reducing latency and computational offloading (O’Grady

et al., 2019). The survey points out that the systems reviewed were mostly prototypes and the

application of edge computing was still in initial stages. In the current study, a requirement to

perform real-time control led to performing computation at the edge.

4.3 AI IN AGRICULTURE

4.3.1 Overview

Both indoor and row-crop farms produce a multitude of data points every day on

temperature, humidity, soil moisture, water usage, etc. that can be leveraged and analyzed in real-

time with the help of AI for obtaining useful insights (Liakos et al., 2018). The importance of such

AI enabled applications increases with scale as it is infeasible for a human analyst to handle all

the generated data manually. These applications can plug-in at multiple phases in the farming data

pipeline and improve a wide range of tasks.

10
By automatically processing the field data, AI-enabled systems can help improve the overall crop

quality, yield, and resource utilization (Gertphol et al., 2018; Liakos et al., 2018). Weather

forecasting can help crop planning, crop health monitoring can help take preventative measures,

and farm robots can make tasks more efficient.

In the context of increasing reliability of IoT systems and requiring clean data for

predictive analytics, the current study is also focused on implementing anomaly detection

algorithms for monitoring sensor data and for automated controls. In the context of moving

towards closed loop and adaptive systems, AI for yield monitoring is also explored.

4.3.2 Anomaly Detection

As much as IoT paves a way for convenient data collection, the data needs to be accurate to make

inferences and decisions about the farm. Abnormalities in the data captured by a sensor can occur

because of multitude of reasons and their sources can be organized into two major categories (Ou

et al., 2020)-

Sensor failures: Sensor misses or fails to capture data. This can directly cause an automated control

system to make wrong decisions or indirectly be a source of bias in the data resulting in wrong

inferences.

System failures: Sensor correctly captures the data of an anomalous event occurring in the system.

This means, there is a fault occurrence in the system and corrective measures might be needed

immediately and on a long-term basis to prevent crop/ equipment losses on farm.

11
Anomaly detection is the process of identification of such events in the system that deviate from

the expected behavior. The relevance of anomaly detection is increasing by the day with

deployment of increasing number of sensors on both outdoor and indoor farms and a general shift

towards automation, thereby, making it extremely troublesome to manually check a huge amount

of data acquired from the automated farms. Anomaly detection is based on the hypothesis that

anomalies stand out trivially/ explicitly from rest of the data and can be quantitatively

differentiated from the clean data. Different detection and handling techniques have been

developed across diverse domains based on their nature and application. The types of anomalies,

popular techniques used to handle them in different IoT applications and related applications in

agriculture are discussed below. This section considers anomaly detection only in the context of

an IoT-based system and exclude other management issues affecting crops in the farm such as

presence of obstacles, weeds, diseases, etc.

4.3.3 Types of Anomalies

Numerous classifications exist for differentiating anomalies in a dataset based on nature of input

data and domain of application (such as agriculture, security, banking, etc.) (Foorthuis, 2020). In

the context of time-series and spatial data, a broad classification approach that

differentiatesbetween three diverse categories is described below along with a graphical

representation of anomalies for a sample temperature sensor data (Chandola et al., 2009)

Point anomalies (Global anomalies) – An individual data instance that is significantly different

from rest of the dataset wherever it occurs.

12
Contextual anomalies (Conditional anomalies) – An individual data instance that is anomalous in

a specific context but an identical instance in another context need not be considered so.

Collective anomalies – A collection of data instances together deviate from the rest of the dataset;

the individual values need not be anomalies in themselves either globally or in the context.

4.3.4 Yield Prediction

Yield prediction is one of the most essential components of precision agriculture and spans the

topics of yield mapping, yield estimation, matching of crop supply with demand, and crop

management to increase productivity (Liakos et al., 2018). Traditionally, mathematical, or

empirical crop models that use certain crop growth variables as inputs have been used to predict

plant growth for research applications. However, those models required extensive agronomy and

physiology experiments that limited their use for various crops.

With the rise of ML, more efforts have been directed towards the use of data-centric models

that are based on correlation but do not necessarily deal with causation (Basso et al., 2013).

Supervised learning models have been developed that use crop yield data from previous years to

predict future yield (Liakos et al., 2018). Deep Learning models utilizing either satellite or UAV

imagery (hyper-spectral, multi-spectral, RGB) for yield mapping are increasingly becoming

popular. Imaging systems for various phenotyping applications are also gaining momentum where

yield is one such variable proxied by stem diameter, leaf count, leaf area etc. (Li et al., 2020;

Mochida et al., 2018). Nevertheless, vision-based techniques for indoor farms still have a lot of

room for exploration to help understand the effects of lighting, nutrition, and other environment

factors on plant growth and yield

13
4.4 SYSTEM DESCRIPTION

When system starts in figure.4.1, the moisture sensor, gas sensor MQ135, humidity sensor, and

LED lights will be turned ON immediately. The moisture sensor will monitor the amount of water

in the container of the plant. If the amount of water in the tub is less than required, the moisture

sensor will send a signal to Arduino to turn the pump ON. Moreover, if the amount of water is

sufficient, the pump will stop. MQ135 has been used on this system for two reasons, firstly to

know the amount of carbon in the air and secondly to know the amount of smoke in the room. If

the amount of carbon is not between 400 and 600, then the ventilation process will start. On the

other hand, if MQ135 detects smoke, two things will happen simultaneously.

First, the main electrical supply of the room will be disconnected with the help of the relay

module and secondly, through the NRF24L01 transmitter, the whole system's data will be sent to

the assistant robot. Meanwhile, the system has LED lights to provide the light needed for

photosynthesis, which can be reduced or increased with the intensity requirement. Humidity sensor

senses the humidity of the farm, if the humidity is less than required then the system gives the

signal to spray the sprinkle of water to meet the demand. The system will be able to exchange

information with an assistant robot via NRF24L01.

14
Figure.4.1. Flow diagram of the system.

15
This allows the system to be controlled by an assistant robot that is shown in figure.4.2. A line

follower robot has been used for continuous monitoring of the entire system. This robot has a few

special abilities. This robot has an esp32 camera module, NodeMCU module, GSM, and

NRF20L01 module. The ESP32 camera allows users to watch real-time video of his farm.

NodeMCU is used to manage the robot remotely through the app. This will allow the user to know

all updates of his farm, control the robot, and give fertilizer to the plants. The robot has a fertilizer

container, which is controlled by a servo motor and can provide fertilizer to a specific plant. One

unique feature of the robot is that it can work automatically. The robot will automatically inform

the user about the status of the farm twice a day.

To do this, the robot will first collect the entire firm's data such as moisture sensor data,

MQ135 data, etc. Data from each sensor on the system will be transmitted via an NRF20L01

module. Transmitted data will be received by the NRF20L01 module located on the robot and will

be sent to the user via GSM. Apart from the NodeMCU module, also, GSM is used here because

if there is a fire on the farm and the user does not have an internet connection on his phone at the

same time, he will get the information via SMS on his phone. An important and unique feature of

the system is that when the robot finishes its work, it will go to its stoppage where its battery will

be able to charge automatically so that the user will not have to worry about taking charge of the

robot's battery.

16
 Figure.4.2. Process diagram of controling the robot through app.

Interest in precision indoor farming is growing due to advances in artificial sources of

illumination efficient heating and cooling technologies water-efficient farming methods and the

use of IoT devices for monitoring and controlling various functionalities of the farm interest is

reinforced by a shrinking skilled workforce available for farm work. However, indoor farming is

3-5 times more expensive than traditional farming owing to high initial capital cost, energy, and

labor expenses The present study was undertaken to explore development of a low-cost, indoor

farm, that uses IoT devices and relies on open-source-based monitoring and control systems

because of their promising benefits.

17
4.5 SYSTEM OVERVIEW

4.5.1 3D Simulation And System Prototype

The 3D simulation of the full system is shown in figure.4.3. The fan on the right in the

figure is used for ventilation. The line follower robot follows the path of the black color and

stops in the black circle in the middle. Through which live monitoring of the farm and

fertilization of crops is possible.

Figure.4.3. 3D Simulation of the system.

An agricultural robot is a robot deployed for agricultural purposes. The main area of

application of robots in agriculture today is at the harvesting stage. Emerging applications of

robots in agriculture include weed control planting seeds, harvesting, environmental

monitering and soil analysis

18
Figure.4.4 is the fully functional prototype of the system. We set up all the sensors and the whole

system in this 3-bed prototype area to inspect whether the system is working or not. Moreover,

figure .4.5,figure. 4. 6 show the LED light control and water pumping in the soil respectively. In

addition, figure.4.7 shows the humidity and temperature sensors sense the farm data and shown

in the serial monitor.

Figure.4.4. Prototype of the farm.

The GSM Shield is used as a method of communication in this proposed technology. The

GSM Shield is used to send text messages to the farmer’s cell phone about the conditions of the

temperature, humidity soil and field area. A separate coding has to be done for GSM shield in

Arduino.

19
 a b

Figure.4.5. Light intencity (a) low and (b) high.

Meanwhile, the system has LED lights to provide the light needed for photosynthesis, which can

be reduced or increased with the intensity requirement. In the above figure.5.5. light intencity of

the system is shown. Figure (a) low light intencity and (b) high light intencity.

The advances in LED technology have made it possible to create the perfect environment

to grow vegetables at a large scale with shorter growing cycles and higher yields. In fact, LED is

becoming the de facto source of lighting to create the most advantageous controlled environment

for indoor farming. The advances in LED technologies have made indoor cultivation of vegetables

very energy-efficient.

While most indoor farms use LEDs these days, many are yet to completely convert to LEDs

due to the relatively high price tag. But the gradually decreasing prices of LEDs should help those

farmers convert to LED as it offers a significant energy cost saving.

20
 Figure.4.6. (a) Pump ON (b) serial monitor data

A system using far flung sensors that screen exceptional situation of stages of environment like

water degree, humidity, temperature, movement of animals, moisture content in soil, and so on.

The Arduino model at the side of GSM defend is used.The field situation is dispatched to the

farmer through mobile text messages.With this machine,sensor node failure and electricity saving

are controlled.A machine is proposed for sensible agriculture monitoring based totally on IoT era.

The machine plays statistics shooting, processing, transmission and reception functions.

The aim of their experiments is to realize smart agriculture device, in which the machine

efficiency is to control the surroundings region and lower the cash and farming fee and also saves

electricity. In a nutshell, the layout realizes faraway clever tracking and control of environmental

conditions and additionally replaces the traditional stressed generation to Wi-Fi, also reduces

manpower price.A system is proposed for crop growth which can be monitored using thermal

imaging technique. Here, the irrigation temperature distribution size (ITDM)technique has been

put into movement. In actual time, the captured information comprising of captured values gives

better irrigation

21
The device can capture environmental parameters along with temperature of air and humidity of

air.In shorttime, with the use of AT command, this gadget can also realize SMS automatic sending

and receiving,environmental situations overrun alarm and inadequate balance alarm. Through the

device putting, the message can be dispatched to the user-special cellular cellphone automatically

no matter what the customer’splace is. This gadget as a normal application of IoT inside the

agriculture has got some efficient consequences in the real operation.

Arduino is a microcontroller to control the working of the sensors and manage the working

of the device. The version of Arduino is implemented in this project. It was developed by Arduino

CC. The Arduino board comes with various number of pins. The pins are categorized as output

and input pins. The input pins accept digital as well as analog data. It has 14 digital pins and 6

analog pins. It accepts 7 to 20 volts of power for working. It also has an USB port. The U was the

first version of Arduino to be introduced in the Arduino familyThe Arduino board processes the

information and displays it on Arduino IDE.The results are also sent to GSM module so that it

passes in the information or data or values to the farmer via GSM network.Here, the farmer does

not need to have a smart phone, the normal cell phones also can work to share information with

farmer from fields. This gives the biggest advantage to the proposed system.

The Soil Moisture sensor is used to sense moisture content in the soil. It checks the volume

of water content or moisture present in the soil. The calculations are done in the soil moisture

sensor through coefficients. It estimates the volume of water content in the soil. It detects the water

content in the soil and gets and sends the analog signals which is shown digitally. It transmits the

signals containing information or data or values of the condition of soil to Arduino to further

process it and display.

22
 a b

Figure.4.7. (a) Circuit .(b) Sensor data.

To prevent loss of data, it was stored at multiple locations. It was captured and stored temporarily

at the edge to account for network discontinuities and retrieved at the end of the cycle. It was also

stored in a local DB which was periodically synced with the cloud and both the copies were

retained.

Temperature and Humidity sensor is used to sense temperature and humidity present in the

atmosphere. It has 3 pins generally. One pin is used for transmitting signals, the next pin is used

to receive signals and the last pin is data transfer. It can be used for prolonged time period. It gives

approximate results. It regularly sends information to Arduino UNO. The information consists the

signals which contain the values gathered about temperature and humidity. It is reliable on nature.

It gives a very fast response.

23
In figure.4.8, the assistance robot of the system is demonstrated. This robot can take photos, stream

video, and give manure on the farm. An Arduino based servomotor is used to control the camera

and gate of the fertilizer container.

Figure.4.8. Prototype of the assistant robot

A line follower robot has been used for continuous monitoring of the entire system. This

robot has a few special abilities. This robot has an esp32 camera module, NodeMCU module,

GSM, and NRF20L01 module. The ESP32 camera allows users to watch real-time video of his

farm. NodeMCU is used to manage the robot remotely through the app. This will allow the user

to know all updates of his farm, control the robot, and give fertilizer to the plants. The robot has a

fertilizer container, which is controlled by a servo motor and can provide fertilizer to a specific

plant. One unique feature of the robot is that it can work automatically. The robot will

automatically inform the user about the status of the farm twice a day.

24
 CHAPTER 5

RESULT AND DISCUSSION

5.1 APP DESCRIPTION

Real-time information of the whole system can be found through the Indoor Farm Management

(IFM) app, as well as it can control the full system from any place with the help of the internet.

The menu bar of the IFM app shows in figure.5.1. (a), it consists of Notification, Robot Control,

Photo Gallery, and Help.

The “Notification” gives all the current status of the censors of the farm. As seen in

figure.5.1. (b), the app can give control of the pump and light intensity of the farm to the user. It

also shows real-time data of temperature and humidity of the soil. We can control the line follower

robot and camera from the “Robot Control” section. In addition, we can see a real-time video of

the farm and manually fertilize the crop by “Fertilizer Switch” which is shown in figure.5.2.

Another feature of the app “Photo Gallery”, where saves pictures of the farm taken by

assistance robots. The line follower robot automatically inspects the entire farm every 6 hours and

picks up the photos. Lastly, the user can go to the “Help” option for the user manual or if they

have some queries.

25
 a

Figure. 5.1. (a) App menu and (b) System control from app.

26
 Figure.5.2 . Robot control interface.

The proposed assignment may be further greater with the aid of including pump to the machine to

facilitate computerized irrigation. The automated irrigation device may be triggered when soil

moisture content is going under the brinkstage. The threshold degree can be decided in the written

Arduino code. Hence, whenever the fee for moisture goes under the brinkdegree, the pump gets

mechanically on and the irrigation is performed.

To improve the efficiency andeffectiveness of the machine, the noted recommendations

can be placed into attention. Alternative of water level controlling, the pump may be given to the

farmer by way of which they are able to turn on or off the pump to start or prevent the irrigation

manner without being there on farm at that gift time. The farmer can knowearlier about the

negative climate situations. In such instances, the farmer might also want toforestall the machine

remotely or routinely.

27
The concept of the usage of IOT for irrigation can be prolonged in addition to other tasks in

farming together with farmanimal management, fireplace detection and climate manage.This

could limit human intervention in farming sports. Designed an automated indoor farm production

system that facilitated automated lighting, irrigation, nutrient application and monitoring

environmental parameters (temperature, relative humidity (RH), Carbon Dioxide (CO2)) in the

greenhouse. This system was the foundation for growing multiple batches of microgreens and

source for acquiring data from the sensors.

Developed a dashboard from scratch and hosted on Purdue server. It facilitated two-way

communication between the user and the sensors/ actuators for implementing real-time control of

the indoor farm. Installed a set of gantries for moving red, blue, green (RGB) sensors and thermal

sensors over the growing trays to capture images of plant canopy. The gantry was remote

controlled and a cartesian co-ordinate system was used to guide the motion. Designed calibration

setups and procedures for light, flow, temperature, and RH sensors to determine the accuracy of

measured data and make a choice about a particular make and model of the sensors.

28
The global farming industry is trying various measures for mitigating impacts due to climate

change, dwindling natural resources due to rapid urbanization, and meeting demands to achieve

food and nutritional security for a rapidly increasing population. These challenges are being met

head-on by farmers through the adoption of advances in agro-technologies that enable practicing

of science-based, efficient, and sustainable farming to complement years of traditional farming

experience (Astill et al., 2020; Schimmelpfennig, 2016). Advances in the field of machinery and

Internet of Things (IoT) devices are helping generate various kinds of sensor data from agricultural

operational environments that are creating opportunities for development/application of Artificial

Intelligence (AI) methods to interpret and understand this huge amount of data.

Today, there is a growing trend towards the design and establishment of controlled

environments for agriculture (CEA), involving automated monitoring and control, to increase the

capacity, economic viability, and efficiency of indoor farms (Gómez et al., 2019). Many of the

modern indoor farms rely on sensors and controls for specific purposes that inspired other

researchers to create new concepts and experimental models of closed-loop systems (Jaiswal et

al., 2019; Zamora-Izquierdo et al., 2019). This study created one such closed-loop system

equipped with low-cost sensors, controls, and smart monitoring (approximate total cost $10k, for

details refer to Appendix A) for an experimental indoor farm, that was used for investigating effect

of different treatments on microgreens production.

29
 CHAPTER 6

CONCLUSION

The world's population is growing as well as the food needed for these people to survive, in

addition, people are now more concerned about the nutritional value of food, which in turn

increases the demand for nutritious food. Nevertheless, the amount of arable land in the world is

not increasing. In this critical moment, indoor farming has shown a new light of hope for food

security. For this, we have designed an intelligent farming management system to make indoor

farming easier and cheaper. Through which the whole farm can be monitored very easily through

a line follower robot. In addition, the primary tasks of the farm such as fertilizing the trees,

watering, providing adequate lighting, etc. can be done automatically or manually. Since the

system uses easily available sensors, the system will be relatively low cost than other existing

devices. Whereas so far our device is a prototype, so there are still some issues that we are working

on, e.g. precise fertilizer management of the crops and the ability to work at the upper level of the

farm by the robot. Finally, this system will facilitate indoor farming at a reasonable price.

30
 REFERENCES

[1] “The rise of indoor cropping,” greenwatchbd, 12-Oct-2013. [Online]. Available:

https://greenwatchbd.com/the-rise-of-indoor-cropping/. [Accessed: 01-May-2020].

[2] “Tech innovation in agriculture - Observer TeCH,” The Daily Observer, 27-Nov-2019.

[Online]. Available: https://www.observerbd.com/details.php?id=230356. [Accessed: 01-May-

2020].

[3] F. N. Sabri, N. H. H. M. Hanif and Z. Janin, "Precision Crop Management for Indoor Farming,"

2018 IEEE 5th International Conference on Smart Instrumentation, Measurement and Application

(ICSIMA), Songkla, Thailand, 2018, pp. 1-4.

[4] G. Panda and T. Saha, "Building of Low Cost Reliable Wireless Sensor Network for Smart

Indoor Agriculture Products," 2018 2nd International Conference on Electronics, Materials

Engineering & Nano-Technology (IEMENTech), Kolkata, 2018, pp. 1-5.

[5] P. Serikul, N. Nakpong and N. Nakjuatong, "Smart Farm Monitoring via the Blynk IoT

Platform : Case Study: Humidity Monitoring and Data Recording," 2018 16th International

Conference on ICT and Knowledge Engineering (ICT&KE), Bangkok, 2018, pp. 1-6.
[6] N. Duangban, K. Noinan and S. Wicha, "Farming box: The integrated of vegetable production

system for food safety in small household," 2019 Joint International Conference on Digital Arts,

Media and Technology with ECTI Northern Section Conference on Electrical,

Electronics,Computer and Telecommunications Engineering (ECTI DAMT-NCON), Nan,

Thailand, 2019, pp. 378-382.

[7] M. D. I. Sujon, R. Nasir, M. M. I. Habib, M. I. Nomaan, J. Baidya and M. R. Islam, "Agribot:

Arduino Controlled Autonomous Multi-Purpose Farm Machinery Robot for Small to Medium

Scale Cultivation," 2018 International Conference on Intelligent Autonomous Systems (ICoIAS),

Singapore, 2018, pp. 155-159.