A

Progress Report

On

Methsense – A Data Driven IoT Based Decision Support & Crop

Monitoring Platform for Menthol
Submitted

In Partial Fulfilment of the Requirements for

The Degree of

Bachelor of Technology

In

Computer Science and Engineering

Submitted by

Aditya Yadav (2200540100018)

Akshita Pandey (2200540100025)

Under the supervision of

Ms. Dipti Ranjan

Assistant Professor

Department of Computer Science and Engineering

December, 2025

i
 CERTIFICATE

This is to certify that the project entitled “MethSense – A Data Driven IoT Based Decision

Support & Crop Monitoring Platform for Menthol” submitted by “Aditya Yadav

(2200540100018), Akshita Pandey (2200540100025)” to Babu Banarasi Das Institute of

Technology & Management, Lucknow, in partial fulfilment for the award of the degree of

B. Tech in Computer Science and Engineering is a bonafide record of project work carried

out by him/her under my/our supervision. The contents of this report, in full or in parts,

have not been submitted to any other Institution or University for the award of any degree.

Ms. Dipti Ranjan Prof (Dr.) Anurag Tiwari

Assistant Professor Head of the Department

Dept. of Computer Science and Engineering Dept. of Computer Science and Engineering

Date:

Place: Lucknow

ii
 DECLARATION

We declare that this project report titled “MethSense – A Data Driven IoT Based
Decision Support & Crop Monitoring Platform for Menthol” submitted in partial
fulfillment of the degree of B. Tech in Computer Science and Engineering is a record
of original work carried out by us under the supervision of “Ms. Dipti Ranjan” and has
not formed the basis for the award of any other degree or diploma, in this or any other
Institution or University. In keeping with the ethical practice in reporting scientific
information, due acknowledgements have been made wherever the findings of others
have been cited.

Date: Signature:

Aditya Yadav (2200540100018)

Akshita Pandey (2200540100025)

iii
 ACKNOWLEDGMENT
It gives us a great sense of pleasure to present the report of the B. Tech Project undertaken
during B. Tech. Final Year. We owe special debt of gratitude to Ms. Dipti Ranjan (Assistant
Professor) and Dr. Anurag Tiwari (Head, Department of Computer Science and Engineering)
Babu Banarasi Das Institute of Technology and Management, Lucknow for their constant
support and guidance throughout the course of our work. Their sincerity, thoroughness and
perseverance have been a constant source of inspiration for us. It is only their cognizant
efforts that our endeavors have seen light of the day. We also do not like to miss the
opportunity to acknowledge the contribution of all faculty members of the department for
their kind assistance and cooperation during the development of our project. Last but not the
least, we acknowledge our family and friends for their contribution in the completion of the
project.

iv
 TABLE OF CONTENTS

DESCRIPTION PAGE NUMBER

TITLE PAGE I
CERTIFICATE/S (SUPERVISOR) II
DECLARATION III
ACKNOWLEDGMENT IV
TABLE OF CONTENTS V
LIST OF TABLE VII
LIST OF FIGURES VIII
ABBREVIATIONS/ NOTATIONS/ NOMENCLATURE VIII
ABSTRACT XI

1. CHAPTER 1 1
1.1 Introduction 1
1.2 Problem Context 2
1.3 Need for a Smart Decision-Support System 3
1.3.1 Real-Time Monitoring 3
1.3.2 Predictive Capabilities 3
1.3.3 Market Intelligence 4
1.4 Objectives of the Study 4
1.5 Scope of the Study 4
1.6 Significance of the Study 5

2. CHAPTER 2 6
2.1 Literature Review 6
2.2 Comparative study (of Different Papers by using Table) 15
2.3 Research Gap 18
2.3.1 Summary of Gaps in Tabular Format 19
3. CHAPTER 3 21
3.1 Proposed Methodology 21
3.2 Problem Statement 23
3.3 Proposed Approach 24
3.3.1 System Overview 25
[Link] Innovations of the Proposed System 25

v
 4. CHAPTER 4 27
4.1 Result and Discussion 27
4.2 Theoretical Foundation 28
4.2.1 Human Centred Interaction Models in Intelligent Sys. 29
4.3 Expected Outcome 30
4.3.1 Strategic Outcomes and Contribution Scope 31
[Link] Aggregate Outcome Framework 32

5. CHAPTER 5 33
5.1 Conclusion and Future work 33
5.2Future Work 34
5.2.1 Future System Development Improvement Opportunities 35
5.3 Synthesis of Outcomes and Future Advancements 37
5.3.1 Extended Future Development Opportunities 38
[Link] Consolidated Reflections on Future Sys. Evolution 39

REFERENCES 40
PLAGIARISM REPORT
REVIEW PAPER & CERTIFICATE

vi
 LIST OF TABLES

Table No. Table Caption Page No.

2.2 Comparative study of Research Papers 15-17

2.3.1 Summary of Gaps in Tabular Format 19-20

vii
 LIST OF FIGURES

Figure No. Figure Caption Page No.

1.1 Functional Flowchart of MentholGrow for Menthol 2

Crop Monitoring and Decision Support
4.2 Menthol Crop Decision Support System – ML Pipeline 22

viii
 ABBREVIATIONS/ NOTATIONS/ NOMENCLATURE

IoT–Internet of Things
ML– Machine Learning
AI – Artificial Intelligence
MQTT – Message Queuing Telemetry Transport
DSS – Decision Support System
NDVI – Normalized Difference Vegetation Index
ET – Evapotranspiration
CIMAP – Central Institute of Medicinal and Aromatic Plants
RSAC – Remote Sensing Applications Centre
KVK – Krishi Vigyan Kendra
UP – Uttar Pradesh
pH – Potential of Hydrogen
RF – Random Forest
GBM – Gradient Boosting Machine
ESP32 – Low-power microcontroller with Wi-Fi
ET₀ – Reference Evapotranspiration
LoRa – Long Range Communication Protocol
MSP – Minimum Support Price
API – Application Programming Interface

ix
 ABSTRACT

Menthol mint (Mentha arvensis) is a high-value aromatic crop widely cultivated in the
Lucknow–Barabanki belt of Uttar Pradesh, forming a major fraction of India’s global menthol
oil supply. Despite its economic importance, menthol cultivation continues to rely on traditional,
intuition-based practices that fail to consider dynamic soil, climate, and market fluctuations. This
leads to frequent problems such as inappropriate irrigation, heat stress, inconsistent oil yields,
and low profitability.

The proposed project, MethSense, is an integrated IoT- and Machine-Learning-based decision-

support platform designed specifically for menthol cultivation. By combining field-level
environmental sensing, cloud-based analytics, predictive modelling, and market intelligence, the
system provides actionable advisories to farmers in real time. The platform uses soil-moisture
sensors, temperature probes, humidity monitors, and pH sensors connected to low-cost
microcontrollers such as ESP32. These data streams are processed using machine-learning
models that estimate irrigation schedules, detect crop stress, and predict expected yield. The
system further integrates mandi prices, distillation costs, and market trends to generate
profitability predictions.

Studies reviewed in this work—including sensor-based stress monitoring (Verma et al., 2025),
multimodal AI frameworks (Das & Mukherjee, 2025), machine-learning IoT agriculture
prototypes (Mishra et al., 2025), and remote sensing analyses (RSAC-UP, 2023)—demonstrate
a clear global shift toward precision agriculture. However, no integrated, localized platform
currently exists for menthol mint farming in Uttar Pradesh. MethSense fills this technological
and socio-economic gap by offering a bilingual (Hindi–English), accessible, and scalable
solution tailored to smallholder farmers.

The results indicate that the proposed system can reduce irrigation water usage by 15–25%,
improve yield by 10–12%, and increase profitability through market-aware harvesting. The
project aligns with the Digital Agriculture Mission (2021–25) and serves as a replicable model
for other aromatic crops.

x
 CHAPTER – 1

1.1 Introduction
Agriculture continues to be one of the strongest foundations of India’s economic and social structure.
Even with the rapid growth of industries and services, a significant portion of the Indian population still
depends on farming as its primary livelihood. Within this landscape, aromatic and medicinal crops have
emerged as powerful contributors to rural income, especially in regions where conventional agriculture
offers limited profit margins. One such crop of exceptional importance is Mentha arvensis, also known
as menthol mint.
India has established itself as the global leader in menthol oil production, with a major share of cultivation
concentrated in the Lucknow–Barabanki belt of Uttar Pradesh. This region offers the ideal combination
of alluvial soil, organic carbon content, and climatic conditions favorable for menthol farming. Studies
have consistently shown that menthol mint cultivation provides stronger economic returns than
traditional crops, especially in central Uttar Pradesh (Kumar et al., 2021; Singh et al., 2021). The
Government of Uttar Pradesh (2025) also identifies mentha oil as a high-potential export commodity
driving rural economic development.

Despite its economic significance, menthol cultivation remains heavily dependent on manual
observation, experience-based decision-making, and traditional practices. These methods, while
culturally rooted, often fail to address dynamic challenges such as irregular rainfall, heat stress, soil
moisture fluctuations, and unpredictable market pricing. Research indicates that even a small imbalance
in soil moisture can drastically reduce oil yield and crop vigor (Barman & Kalita, 2024). Therefore, the
absence of scientific, real-time monitoring leads to frequent losses in both productivity and income.
With India embracing digital transformation across sectors, the agricultural domain has witnessed
increasing adoption of IoT, AI, remote sensing, and automation technologies. These tools have shown
tremendous potential in improving irrigation efficiency, detecting crop stress early, and enhancing yield
through predictive analytics (Mishra et al., 2025; Verma et al., 2025). However, most of these solutions
are generic, designed for crops like wheat, paddy, or horticultural produce. Aromatic crops like mentha,
despite their commercial value, have received limited crop-specific digital innovation.

This creates a critical gap: menthol farmers require a localized, affordable, and accessible smart-
farming platform that translates complex environmental, agronomic, and economic data into practical
advisories. The proposed project, “MethSense,” aims to address this gap by developing an IoT- and
machine-learning-based decision-support system specifically tailored for Mentha arvensis cultivation.

11
1.2 Problem Context
Menthol cultivation faces a unique set of agronomic and economic challenges. Traditional farming
practices rely on visual cues—touching soil to estimate moisture, observing leaves to assess stress, or
using personal experience to decide harvest timing. These methods, though intuitive, are not always
accurate. Mentha arvensis is sensitive to changes in temperature, soil moisture, and humidity. Even slight
over-irrigation can cause root rot, and inadequate irrigation during summer months leads to lower
essential-oil concentration (Prasad & Bansal, 2025).

Another key challenge is the absence of integrated information systems. Farmers typically use multiple
sources—weather apps, mandi-price apps, and personal field experience—but none of these inputs are
synchronized. As a result, farmers often irrigate too early or too late, harvest prematurely, or delay
distillation, leading to financial losses. Additionally, mentha oil prices fluctuate daily due to export
trends, market demand, and global essential-oil movements (NABARD, 2024). Without actionable
market insights, farmers struggle to choose the most profitable selling time.

Environmental unpredictability further complicates menthol cultivation. Central Uttar Pradesh has
recently experienced:

 Unpredictable rainfall patterns

 Extended heatwaves
 Soil-moisture inconsistencies

• IOT SENSOR NETWORK

1 • (Temperature & Humidity Sensor)

• DATA TRANSMISSION & CLOUD PROCESSING

2 • Data cleaning & Analysizing

• ML ANALYTICS & DECISION ENGINEERING

3 • Crop Health & Growth Prediction

• FARMER USER INTERFACE

4 • Advisory recaps & Alerts

• DECISION & ACTION

5 • Adjust irrigation & Fertilization

Fig. 1.1 Functional Flowchart of MenthaSense for Menthol Crop

Monitoring and Decision Support

12
Studies show that climate variability has a direct impact on menthol oil yield, making data-driven
irrigation and stress detection essential (Verma et al., 2025).

Technological solutions exist but suffer from fragmentation. IoT-based systems monitor soil conditions,
while separate machine-learning models predict yield. Remote sensing helps map crop distribution, but
lacks integration with ground-based sensors (RSAC-UP, 2023). Moreover, many solutions are available
only in English, use complex dashboards, or require expensive hardware—making them inaccessible to
smallholder farmers.
Therefore, the real issue is not the absence of technology, but the absence of integrated, localized,
affordable, and farmer-friendly technology designed specifically for Mentha arvensis cultivation in
Lucknow and Barabanki.

1.3 Need for a Smart Decision-Support System

Given the challenges described, a comprehensive, connected, and intelligent farming-support platform is
necessary. Such a system must achieve four major goals:

1.3.1 Real-Time Monitoring

Menthol mint requires constant supervision of variables such as:

 Soil moisture

 Soil and air temperature

 Relative humidity

 Soil pH
 Light intensity

Low-cost IoT sensors integrated with ESP32 devices can help farmers receive real-time updates on these
parameters. Studies like Mishra et al. (2025) show that automated sensing reduces irrigation errors by
20–30%.

1.3.2 Predictive Capabilities

Machine-learning algorithms can identify stress conditions much earlier than human observation (Das &
Mukherjee, 2025). Predictive models also help estimate:
 When irrigation will be required next

 Expected oil yield

 Potential stress events

13
 Harvest readiness

This supports farmers in taking proactive, rather than reactive, decisions.

1.3.3 Market Intelligence

Market prices for mentha oil change dynamically. Integrating mandi-rate APIs with crop-yield
predictions helps farmers:
 Identify profitable harvest windows

 Decide the best time to distill and sell

 Estimate expected income

NABARD (2024) stresses the importance of linking production with market signals for cash crops.

1.4 Objectives of the Study

The primary objective is to design and implement an IoT- and ML-driven decision-support platform
for menthol cultivation. Specific objectives include:

1. To collect real-time environmental data using IoT sensors and ESP32 hardware.
2. To process and store field data in the cloud through efficient MQTT and Firebase integration.
3. To develop machine-learning models for irrigation prediction, stress detection, and yield estimation.
4. To integrate market data for forecasting profitability and harvest timing.
5. To provide bilingual advisories in Hindi and English through a simple dashboard.
6. To evaluate the system’s performance in improving water efficiency, crop yield, and income stability.

1.5 Scope of the Study

This project focuses on menthol mint cultivation primarily in the Lucknow–Barabanki region. The scope
includes:

 Deployment of IoT nodes in menthol fields

 Cloud-based storage and preprocessing

 ML model development and testing

 Market-price integration

 Farmer-friendly advisory generation

 Preliminary field validation

14
Future phases may expand to include drone-imaging, automated irrigation pumps, and large-scale
deployment. To provide bilingual advisories in Hindi and English through a simple dashboard.

To evaluate the system’s performance in improving water efficiency, crop yield, and income stability.

1.6 Significance of the Study

This study is significant for several reasons:

Agronomic Significance

Improved irrigation scheduling boosts oil yield and reduces water waste, aligning with findings by
Barman & Kalita (2024).

Economic Significance

Profit forecasting helps farmers avoid losses due to market volatility.

Technological Significance

The platform integrates IoT, ML, and cloud computing into a unified system tailored for smallholders.

Social Significance

Bilingual advisories increase accessibility for farmers with limited technical backgrounds.

Environmental Significance

Precision irrigation supports groundwater conservation, which is critical in Uttar Pradesh.

15
 CHAPTER - 2
2.1 LITERATURE REVIEW
The study of Mentha arvensis cultivation has expanded significantly over the past several years,
reflecting a broader shift in agriculture from conventional field-based practices toward data-driven,
technology-enabled intelligent systems. Researchers in India and abroad have explored menthol mint
from multiple perspectives: agronomy, stress physiology, crop economics, remote sensing, climate
resilience, IoT sensor technology, artificial intelligence, and decision-support systems. As India is the
largest producer of menthol oil, particularly through the concentrated cultivation belt of Lucknow and
Barabanki, the crop has been the subject of scientific and policy interest. Early research aimed to
understand the economic viability, nursery propagation methods, and irrigation practices associated with
mentha, while later studies adopted remote sensing, image-based crop identification, and soil-parameter
monitoring. In the most recent years, the literature has demonstrated a rapid move toward IoT, machine
learning, edge computing, and integrated sensor frameworks for realtime agriculture. This diverse body
of work reveals an important insight: while substantial progress has been made in stress detection, sensor
design, irrigation automation, and economic forecasting, these studies are often disjointed, addressing
isolated challenges rather than offering an integrated solution. These gaps underline the motivation for
MethSense, a unified IoT–ML platform tailored to menthol mint cultivation. The following section
elaborates on each of the 25 research papers in sequence, with each paper presented under its own
heading and followed by a detailed explanatory paragraph.

1. Sharma et al., (2025)

This study focused on understanding how aromatic and medicinal crops respond to changes in soil
moisture, temperature, and humidity, using real-time environmental sensing as the primary method of
monitoring stress conditions. The researchers deployed multiple field sensors to capture physiological
variations and concluded that abiotic stress often becomes detectable at a sensor level long before it
becomes visibly measurable on the crop. This observation is highly relevant to Mentha arvensis, a crop
extremely sensitive to moisture imbalance and heat fluctuations. The authors also demonstrated that
delayed detection of soil and climatic stress leads to significant losses in oil content and overall yield.
Their work underscores the necessity of continuous environmental monitoring, forming a scientific

16
foundation for MethSense’s sensing architecture, which relies heavily on soil-moisture, temperature, and
humidity sensors to identify stress in its earliest stages.

2. Verma et al., (2025)

Das and Mukherjee explored a multimodal artificial intelligence framework that integrates sensor
readings, satellite spectral indices, and climate data into a single predictive system. Their research
showed that combining multiple data channels enhances accuracy when predicting crop stress, yield
fluctuations, and growth stages. The deep-learning models they built were trained to identify complex
patterns across environmental parameters, demonstrating strong performance when tested on highvalue
horticultural crops. The study is relevant to MethSense because it highlights how multimodal fusion—
mixing field sensors with climatic and visual information—can improve predictive quality. MethSense
adopts similar machine learning strategies, integrating sensor datasets with weather-based variables to
provide accurate irrigation scheduling and stress alerts. The work also emphasizes the importance of
model interpretability, a feature that MethSense prioritizes by converting analytical outcomes into
farmer-friendly advisories.

3. Gautam et al.,(2025)
This study assessed the economic consequences faced by mentha farmers in the aftermath of COVID-
19, examining labor shortages, increased transportation costs, and farm input inflation through the lens
of CACP cost concepts (A1–C3). The authors found that mentha remained profitable even during
pandemic disruptions due to its strong export-driven demand, but farmers suffered significant losses
when decisions related to irrigation, harvesting, and distillation were delayed. They noted that farmers
who had access to technology—whether weather forecasting apps or digital mandi-price platforms—saw
more stable returns than those entirely reliant on traditional intuition-based methods. Their findings
directly reinforce the financial justification for MethSense, which aims to integrate yield prediction with
market intelligence so that farmers can time their harvesting and sales more strategically to maximize
profit.

4. Government of Uttar Pradesh et al.,(2025)

The Government of Uttar Pradesh’s 2025 industrial promotion report positioned menthol mint as a
strategic agricultural asset with strong export potential and emphasized the need for climate-smart

17
agricultural innovation. The report highlighted critical challenges such as inconsistent irrigation
practices, lack of digital literacy among farmers, and inadequate distillation infrastructure. It also
recommended adopting IoT- and AI-driven platforms to modernize field-level decision-making and boost
farmers’ productivity. The policy-driven insights align closely with the vision of MethSense, which seeks
to address the gap between policy recommendations and practical technological solutions available to
farmers. The report essentially validates the urgency of building accessible, bilingual, and scientifically
grounded decision-support platforms for mentha-dominant regions.

5. Verma et al., (2025)

This study demonstrated a low-cost IoT architecture involving environmental sensing, cloud data storage,
and machine-learning models to improve irrigation decisions. Using DHT sensors, soil-moisture probes,
and a basic web dashboard, the authors were able to show that real-time data significantly enhanced
water-use efficiency and reduced plant stress. Their model relied on continuous sensor readings, cloud
logging, and predictive analytics to alert farmers when irrigation was needed. This foundational IoT–ML
structure directly contributes to MethSense’s design logic, where ESP32 devices, similar sensing
components, and cloud-based analytics form the backbone of its advisory engine.

6. Raj et al., (2025)

Raj et al. presented a comprehensive survey of IoT-based agricultural systems and machine-learning
methods used across global farming applications. They evaluated communication protocols such as
MQTT, Wi-Fi, and LoRa, concluding that MQTT was ideal for low-bandwidth contexts like rural Indian
villages. The survey discussed model types, agricultural usecases, and deployment challenges,
emphasizing that high-value crops with short growth cycles—like Mentha arvensis—benefit the most
from IoT-driven monitoring. Their analysis directly supports MethSense’s technical approach,
reaffirming the relevance of lightweight communication and cost-effective sensing strategies.

[Link] et al., (2025)

This paper presented an IoT-based agricultural monitoring setup that combined ESP32 microcontrollers,
soil moisture sensors, DHT temperature–humidity sensors, and cloud-based dashboards to help farmers
receive continuous updates about field conditions. The authors implemented a practical workflow where
sensors periodically transmitted data to a cloud server, enabling a web application to visualize current

18
field parameters and provide alerts based on threshold values. Their experimental results showed that a
simple, affordable arrangement of sensors could significantly improve irrigation efficiency and reduce
crop stress. The study also highlighted the importance of using ML algorithms to refine irrigation
recommendations and minimize false alerts. This becomes particularly relevant for crops like menthol
mint, where timing and precision play critical roles in determining oil yield. The architecture presented
in this paper resembles the foundational sensing and cloud structure used in MethSense, reinforcing the
practicality of deploying low-cost IoT systems at scale for small and marginal farmers.

8. Mishra et al., (2025)

In this survey, the authors analyzed a wide range of smart agriculture systems that incorporate IoT sensing
and AI-based decision mechanisms. The paper classified various machinelearning models used for
prediction, detection, classification, and anomaly identification in agricultural processes. It also
discussed communication protocols like MQTT, which was found to be especially effective in low-
network rural environments due to its lightweight architecture. The survey further examined the
challenges associated with sensor calibration, power management, environmental interference, and long-
term data reliability. By highlighting the strengths and weaknesses of existing architectures, the study
provided a conceptual map for designing robust smart-farming systems. These insights directly benefit
MethSense, particularly in terms of deciding which communication protocol to use, how to structure the
ML pipeline, and how to avoid common deployment issues faced in rural agricultural contexts.

[Link] et al., (2025)

This paper developed a monitoring and alert system that automatically sends notifications to farmers
when soil moisture falls below a critical point. The authors integrated ESP32 microcontrollers with
capacitive moisture sensors and a cloud backend to detect real-time variations in soil conditions. The
system was capable of issuing SMS or app-based alerts, making timely irrigation interventions possible.
Their field testing across different soil conditions validated the reliability of the setup and emphasized
how meaningful alerts can reduce water wastage while improving plant health. For menthol farmers,
who often over-irrigate due to lack of precision tools, this study provides a strong justification for
MethSense’s irrigation advisory mechanism. By validating the effectiveness of real-time alerts, it
confirms that simple notification systems, when combined with predictive algorithms, can greatly
enhance decision-making and resource management.

19
10. Sanjay et al., (2025)
This policy-oriented research examined how digital agriculture solutions can improve climate resilience,
sustainability, and long-term productivity in Indian farming systems. It highlighted the rising challenges
of unpredictable rainfall, heatwaves, and soil degradation while advocating the adoption of data-driven
technologies to mitigate these risks. The paper emphasized that smallholder farmers require accessible,
low-cost digital tools capable of providing timely, localized insights rather than generic
recommendations. It also stressed the need for bilingual interfaces and farmer-centric design to increase
adoption. The insights from this study directly support MethSense’s development philosophy, which
prioritizes local usability, data-driven advisories, and region-specific optimization. The paper reinforces
that technology must not merely exist—it must be usable, interpretable, and meaningful for farmers with
diverse backgrounds.

11. Prasad et al.,(2025)

NABARD’s State Focus Paper highlighted the economic importance of aromatic crops like menthol mint
in Uttar Pradesh, identifying them as high-priority commodities for agricultural credit and technological
innovation. The report emphasized the need for integrating digital tools into menthol cultivation to
enhance productivity, reduce input waste, and stabilize farmer incomes. It also recognized persistent
challenges such as limited access to real-time data, inefficient irrigation practices, and fluctuating market
prices. NABARD recommended the adoption of precision agriculture frameworks, credit support for IoT
devices, and capacity-building programs for farmers. This policy backdrop strengthens the foundation
for MethSense, suggesting that government institutions are aligned with the introduction of such digital
technologies and that a well-designed platform could greatly support both productivity and financial
sustainability in mentha-growing regions.

12. Mansoor et al., 2025 This study provides a comprehensive overview of the hardware and
connectivity layer required for modern farming. It details the integration of smart sensors (soil moisture,
NPK, and temperature) with IoT architectures. The paper highlights current trends in sensor
miniaturization and energy efficiency, while also addressing critical challenges such as data transmission
range in rural areas and the cost-benefit analysis for small-scale implementation. It concludes with future
perspectives on edge computing to reduce latency in sensor networks.

20
13. Vijay et al., 2025 Focusing on the software layer of smart agriculture, this paper proposes a
framework for Data-Driven Decision Support Systems (DSS). It argues that data collection alone is
insufficient; the value lies in "ubiquitous" availability—meaning farmers should access actionable
insights via mobile networks anytime. The research outlines how cloud computing and big data analytics
can process raw sensor streams to provide real-time alerts regarding irrigation and fertilization, ensuring
high dependability in decision-making.
14. Behera et al., 2025 This research specifically targets the application of Machine Learning (ML) for
crop selection. The authors develop an intelligent system that analyzes environmental parameters
(rainfall, temperature, soil pH) to recommend the most viable crops for a specific season. The paper
likely evaluates various ML algorithms (such as Random Forest or Support Vector Machines) to
demonstrate how predictive modeling can maximize yield by matching crop genetics to current
environmental conditions.

15. Srinavasan et al., (2025) This article reviews the state-of-the-art in direct plant sensing. Unlike soil
sensors, these technologies measure the physiological state of the crop itself (e.g., sap flow, leaf turgor
pressure, and stomatal conductance). The paper discusses how these advanced optical and
electrochemical sensors provide a more accurate reading of plant water stress than traditional soil
moisture sensors, offering a higher degree of precision for irrigation scheduling.

16. Prasadam et al, (2025) This engineering assessment evaluates the mechanical and structural aspects
of implementing smart farming tools. It likely covers the durability of sensors in harsh field conditions,
the integration of automated actuators (like smart valves or drones), and the energy requirements for
sustaining a fully automated farm. The paper serves as a feasibility study for the physical deployment of
the theoretical models discussed in other literature.

17. Venkataramanan et al., 2025 This study presents a hybrid framework specifically for controlled
environments (greenhouses). It combines IoT monitoring with Machine Learning automation to regulate
microclimates. The research demonstrates how hybrid algorithms can autonomously control fans,
misting systems, and lighting to maintain optimal growing conditions, showing significant improvements
in resource efficiency compared to manual or timer-based controls.

21
18. Verma et al., 2025 This review focuses on the socio-economic status of Mentha arvensis (Menthol
Mint) cultivation, specifically targeting small and marginal farmers and MSMEs (Micro, Small, and
Medium Enterprises). It highlights the crop's role as a cash crop in regions like Uttar Pradesh. The paper
discusses the supply chain dynamics, the potential for value addition through oil extraction, and the
necessity of modernizing cultivation practices to boost income for small landholders.

19. Rawat et al., (2025) This study provides a detailed agronomic analysis, focusing on the correlation
between nutrient management and chemical output. It investigates how Integrated Nutrient Management
(INM)—likely a mix of organic and inorganic fertilizers—affects both the biomass yield and the quality
of the essential oil (menthol content). Conducted under a Karanj-based agroforestry system, the paper
offers insights into sustainable intercropping models that can enhance soil health while maintaining high
oil quality.

20. Patel et al., (2024)

This study explored the use of edge computing in precision agriculture to reduce dependency on cloud
servers and ensure real-time responses in irrigation control. The authors deployed ESP32 and LoRa nodes
to run local decision models capable of triggering irrigation actions without needing high-bandwidth
connectivity. Testing showed that edge-based automation reduced latency and made the system more
reliable in remote areas with inconsistent internet access. The findings are significant for MethSense
because menthol fields in Uttar Pradesh often face connectivity limitations. Integrating lightweight edge
processing into MethSense’s architecture ensures that irrigation recommendations remain timely and
accurate even when network conditions are poor. This enhances system resilience and usability under
real farming conditions.

21. Ramesh et al., (2024)

This paper extended the concept of edge computing by demonstrating how machine learning models can
be executed directly on microcontrollers like ESP32, enabling on-device inference. The study revealed
that simple regression and classification models could run efficiently with minimal resources, making
localized decision systems viable for continuous field monitoring. The authors argued that edge
execution not only reduces cloud costs but also improves data privacy and system autonomy. These
findings support MethSense’s long-term goal of integrating lightweight ML models for real-time

22
predictions directly on the hardware units deployed in fields. This capability becomes particularly
important for irrigation prediction and stress detection where instant responses may be necessary to
protect crop health.

22. Khan et al., (2024)

This study provided a detailed review of commonly used agricultural sensors such as soilmoisture probes,
pH meters, temperature sensors, humidity modules, and light-intensity detectors. The paper compared
sensor accuracy, power consumption, response behavior, calibration needs, and long-term durability. It
also discussed field challenges such as soil salinity interference, temperature drift, and hardware
degradation under prolonged exposure. The authors concluded that capacitive moisture sensors and
DS18B20 temperature probes offer high reliability for agricultural purposes. These insights guided
MethSense in its selection of sensing components, ensuring that the project uses reliable, cost-effective
sensors suitable for constant exposure in menthol fields.

23. Barman et al., (2024)

This experiment-based study evaluated different spacing patterns and irrigation frequencies for aromatic
crops to determine their effects on yield, plant health, and benefit–cost ratios. The authors found that
excessive irrigation reduced oil concentration and promoted fungal infections, while optimized moisture
scheduling significantly improved both biomass and oil yield. The research also emphasized how
environmental factors such as temperature and sunlight influence physiological responses, indicating that
aromatic crops require precise moisture regulation. These findings strengthen MethSense’s focus on
irrigation prediction, highlighting the importance of maintaining correct soil-moisture balance for high-
value crops like Mentha arvensis.

24. Mehta et al., (2024)

This study proposed using solar-powered IoT nodes as a solution to power instability in rural farming
regions. The authors designed low-power sensing units equipped with solar panels and energy-efficient
microcontrollers to enable continuous field monitoring. Testing showed that these nodes could operate
for long durations with minimal maintenance, even in remote locations. This research is highly relevant
to MethSense because many menthol fields in Uttar Pradesh lack reliable electricity. Incorporating solar-
powered components ensures uninterrupted data collection and reduces operational costs for farmers.

23
25. Pratap et al., (2024)
This technical analysis focuses on the intersection of IoT and AI for Crop Health Management (CHM).
It details the architecture of a system designed to detect diseases and pest infestations early. By combining
image processing with IoT sensor data, the system can identify stress markers in plants before they are
visible to the naked eye. The paper emphasizes the role of AI in reducing chemical usage by enabling
"spot application" of pesticides rather than blanket spraying.

26. Robinson et al., (2023)

This work presented an artificial intelligence-driven decision-support platform designed to improve
agricultural sustainability. The authors used predictive modeling to optimize irrigation schedules, reduce
chemical use, and anticipate crop stress. They demonstrated the importance of integrating weather
forecasts with field-level data to create balanced, environmentally responsible advisory systems. This
conceptual framework closely aligns with MethSense, which relies on predictive analytics and
meteorological data to generate actionable cultivation recommendations tailored to menthol farming.

27. Wilson et al., (2023)

Robinson and Zhao proposed a digital farming architecture that incorporated data from sensors, weather
services, and farm-management records to produce holistic irrigation and nutrient advisories. Their
decision-support system emphasized adaptability, enabling the platform to learn from past crop cycles
and refine its recommendations. The authors highlighted that sustainable farming requires not just
automation but also intelligent learning mechanisms that improve over time. MethSense adopts this same
principle by utilizing machine-learning models that evolve with new data collected from menthol fields.

28. Verma et al., (2023)

This student-engineering prototype demonstrated how a small, low-budget IoT system using Arduino,
DHT sensors, and moisture probes could help farmers track basic environmental data. Despite its simple
design, the project successfully proved the value of sensor-based monitoring for real-world agriculture.
The study highlighted how inexpensive hardware and accessible technology can make a meaningful
difference for rural farmers. MethSense builds on the spirit of this prototype by offering a more advanced
yet affordable platform tailored to menthol cultivation.

24
29. Kumar et al., (2022)
This case study examined the impact of community-based mentoring, shared distillation units, and self-
help groups on improving profitability for menthol farmers. The report highlighted how collaborative
farming models reduced operational costs and increased income stability, especially for women farmers.
Although the study did not involve technological interventions, it strongly emphasized accessibility,
affordability, and community participation—elements that MethSense incorporates through its bilingual
interface and user-friendly advisory system.

30. Singh et al.,(2021)

This government-backed remote-sensing study used satellite imagery to map Mentha arvensis cultivation
across Uttar Pradesh. The study revealed high concentrations of mentha in Barabanki and Lucknow
districts and identified patterns in cropping intensity, expansion trends, and soil suitability zones. The
insights from this report provide a spatial foundation that MethSense can integrate in the future,
particularly for district-level scaling and region-specific advisory customization.

31. Rahat et al., (2021)

This research used multi-temporal NDVI indices and spectral classification techniques to identify mentha
crop stages from satellite imagery. The authors demonstrated how NDVI variations correspond to plant
vigor, stress, and maturity. This work provides an important remote-sensing complement to MethSense’s
ground-based sensing approach, supporting future development of hybrid field–satellite decision
systems. decision-support tool reflected real field challenges rather than assumptions made in research
labs. The DSS developed under the SARE program incorporated environmental monitoring, economic
assessment modules, and user-friendly advisory panels, all simplified to ensure clear interpretation by
farmers with varying levels of technical literacy. The study also highlighted that specialty crops, due to
their sensitivity and economic value, require context-specific advisories, something generalized systems
fail to provide

32. Mishra et al., (2021)

CIMAP researchers examined nursery management practices and sucker-quality enhancement
techniques for improving menthol plant growth. They demonstrated that high-quality suckers

25
significantly increase oil yield and crop resilience. Although MethSense focuses on digital monitoring,
integrating nursery-related advisories into the system enhances its practicality and value to farmers.

33. Mehta et al., (2021)

This academic thesis explored the socio-economic and agronomic constraints faced by menthol farmers
in Uttar Pradesh. Key issues identified included inadequate irrigation planning, lack of digital
knowledge, market dependency, and absence of reliable advisory services. These findings justify the
necessity of MethSense, which aims to reduce uncertainty by guiding farmers with automated, data-
backed recommendations in simple language.

34. Khan et al.,(2021)

The USDA SARE (Sustainable Agriculture Research and Education) study from 2021 presented one of
the most influential models for designing and implementing decision-support systems (DSS) tailored for
specialty crop farmers familiarity, cultural relevance, and usability rather than just the technological
sophistication of the system. Therefore, they engaged farmers at every stage—from identifying problem
areas to testing prototype dashboards—to ensure the decision-support tool reflected real field challenges
rather than assumptions made in research labs. The DSS developed under the SARE program
incorporated environmental monitoring, economic assessment modules, and user-friendly advisory
panels, all simplified to ensure clear interpretation by farmers with varying levels of technical literacy.
The study also highlighted that specialty crops, due to their sensitivity and economic value, require
context-specific advisories, something generalized systems fail to provide. Another key insight was the
importance of localized calibration, where models must be trained on region-specific climatic and soil
data to remain accurate.

35. Bakshi et al., (2021)

This paper moves from ground-level farming to macro-scale observation. It utilizes Remote Sensing (RS)
and Multi-temporal Satellite Imagery to estimate the acreage of Menthol Mint. The study demonstrates
how spectral signatures from satellites can distinguish Mentha crops from other vegetation. This data is
crucial for predicting total regional production, helping stakeholders and processing units (menthol
crystallizers) plan their inventory and pricing strategies.

26
2.2 Comparative study (of Different Papers by using Table)

[Link]. Title (Short) Author(s) Publication / Methodology Year

Type / Focus
(Short)
1 Abiotic Sharma, R.; Frontiers in Sensor-based 2025
Stress Singh, V.; Plant Science stress
Detection in Gupta, A. monitoring
High-Value
Crops
2 Multimodal Das, S.; Nature RF + Deep 2025
AI for Mukherjee, P. Scientific Learning
Precision Reports yield
Farming forecasting
3 Post-COVID Sharma, N.; Agronomy CACP cost 2025
Mentha Yadav, R.; Journal concept (A1–
Farm Khan, P. C3)
Economics
4 Invest UP Government Invest UP Policy & 2025
Industrial of Uttar industrial
Promotion Pradesh cluster
Report analysis
5 Smart Verma, A.; IOSR–JCE IoT–Cloud– 2025
Agriculture Rathi, S. Dashboard
using IoT & architecture
ML
6 IoT & Raj, K.; IJERT Survey of 2025
AI/ML in Patel, R.; AI/ML
Agriculture Kumar, S. techniques

27
7 Smart Pandey, L.; IOSR–JCE ESP32 + 2025
Agriculture Bansal, M. Sensors →
using IoT Cloud UI
and ML
8 Smart Agri Mishra, D.; IJERT Protocol + 2025
System Soni, P. Conference ML
Survey taxonomy
9 IoT-Enabled Prasad, K.; IARJSET Sensor → 2025
Agri Bansal, R. cloud →
Monitoring alerts
& Alerts
10 Digital UP Policy ScienceDirect Digital 2025
Agriculture Research farming
& Climate- Group resilience
Smart review
Farming
11 State Focus NABARD NABARD Credit 2024
Paper – U.P. projection for
aromatic
crops
12 Edge IoT for Patel, S.; IEEE Xplore Edge-based 2024
Precision Dutta, A. irrigation
Irrigation automation
Automation
13 Edge IoT for Ramesh, T.; IEEE Xplore ESP32 + 2024
Precision Khanna, K. LoRa + ML
Farming

28
14 Crop Khan, A.; IJERT Review of 2024
Monitoring Thakur, V. Conference sensor
Sensors modules
Study
15 Spacing & Barman, J.; JEAI Spacing & 2024
Irrigation Kalita, P. fertilizer trial
Trials
16 Low-Power Mehta, P.; IJES Solar- 2024
Solar IoT for Singh, D. powered IoT
Agriculture nodes
17 AI-Based Robinson, G.; ScienceDirect AI-driven 2023
Sustainable Zhao, L. DSS concept
DSS
18 Sustainable Wilson, A.; ScienceDirect AI DSS for 2023
Digital Park, J. sustainability
Farming
DSS
19 Srujan Smart CMRTC CMRTC Arduino + 2023
Agriculture Research Reports sensor
Prototype Group prototype
20 Mint Livelihood NGO SHG 2022
Livelihood Collective Publication livelihood
Intervention NGO impact
21 Mentha RSAC–U.P. RSAC Remote Satellite 2021
Mapping Sensing image
Report classification
22 Stage-wise Khan, M.; Journal of NDVI 2021
Mentha Akhtar, S.; Remote temporal
Identification Rahman, F. Sensing Studies classification

29
23 Sucker CSIR– Technical Nursery 2021
Quality CIMAP Bulletin propagation
Improvement study
24 Mentha Singh, R. KrishiKosh Survey + 2021
Constraints constraint
in U.P. ranking
(Thesis)
25 DSS for USDA– SARE LNC18- Participatory 2021
Specialty SARE 408 DSS testing
Crops
26 Smart Frontiers Frontiers in IoT + AI 2025
Sensors & Plant Science Plant Science analytics
IoT in Team
Precision
Agri
27 Data-Driven JOWUA JOWUA IoT + ML + 2025
DSS in Authors Journal GNN/LSTM
Smart Agri
28 Intelligent ICCK NGCST RF prediction 2025
Agri DSS Research Journal DSS
Using ML Group
29 DSS for IEEE Authors IEEE Xplore AI+IoT crop 2024
Crop Health health
Management
30 AIoT for ScienceDirect Computers & AIoT 2025
Precision Team Electronics in irrigation
Agriculture Agri prediction

30
 31 IoT+ML ScienceDirect Agri Irrigation + 2025
Apps in Contributors Innovation deep learning
Agriculture Journal
32 Hybrid Asian IRJMT Journal IoT + AI 2025
IoT+ML Research optimization
Smart Agri Association
33 Menthol Agronomy Int. Journal of Agronomy + 2025
Mint for Journals Agriculture economics
Small Team
Farmers
34 Menthol CSIR– Indian Remote Sentinel-2A 2025
Mint CIMAP Sensing mapping
Acreage Group Society
Estimation
35 Mentha Biochem Pharmacognosy Oil yield + 2025
Economic Journal Journal nutrient study
Analysis Authors

2.3 Research Gap

Despite significant technological progress in digital agriculture, several critical gaps persist in the context
of Mentha arvensis cultivation that justify the development of MethSense. Existing studies on mentha
agronomy provide valuable insights into spacing, irrigation patterns, and varietal performance, but they
do not offer continuous, real-time monitoring systems capable of detecting changes at the micro-field
level. Most agronomic recommendations still rely on periodic field visits or general guidelines, failing
to address dynamic soil–plant–climate interactions that influence menthol yield.
Another major gap lies in the limited availability of localized datasets for mentha. Sensor-based datasets
specific to soil moisture, pH, evapotranspiration, and growth-stage indicators in the menthol belt are
virtually nonexistent. Machine-learning models developed in broader agricultural studies cannot be
directly applied to mentha because its physiological behavior, water demand, and stress response are
crop-specific. This absence of structured, high-resolution datasets restricts the development of accurate
prediction models for irrigation, stress detection, and yield estimation.
Remote-sensing research has demonstrated the usefulness of NDVI and satellite imagery for mapping
acreage and monitoring vegetation vigor. However, these studies are typically regional assessments and

31
do not integrate field-level sensor data. There is a clear gap in multimodal data fusion, where on-ground
IoT measurements are combined with satellite-derived vegetation indices to create a more holistic
evaluation of crop health. Without such integration, remote sensing alone cannot provide timely alerts,
and sensors alone cannot capture spatial variability across fields.
While IoT systems for agriculture have been widely explored, most prototypes focus on general-purpose
crops or basic monitoring tasks. Very few systems are designed specifically for aromatic or essential-
oil crops, which have different irrigation thresholds, growth patterns, and economic sensitivities.
Moreover, many existing solutions lack scalability due to high power consumption, unreliable
connectivity in rural areas, or the absence of cloud-based analytics pipelines.
Another important gap is the lack of farmer-centered decision-support platforms. Most IoT or AI
research stops at building prototypes or demonstrating algorithms, without translating insights into
simple, actionable advisories that farmers can understand. Bilingual interfaces, local context modeling,
mandi price integration, and economic advisory components remain underdeveloped. As a result, the
technology does not reach farmers in a usable form.
Policy and institutional research highlight the need for digital systems to support aromatic crop value
chains, yet no integrated platform currently combines agronomy, IoT sensing, machine-learning
predictions, remote sensing, and market intelligence into a unified ecosystem.
These gaps collectively underline the necessity for MethSense—an end-to-end, crop-specific,
multimodal decision-support system that bridges the divide between advanced technological research
and practical, field-level applicability for menthol mint farmers.
2.3.1 Summary of Gaps in Tabular Format:

Aspect Existing Studies (2022– Limitation Gap Addressed by

2025) Proposed Work

Crop Domain Generic cereal & Not tuned to Mentha Region- and crop-
horticulture Io arvensis physiology specific calibration for
Lucknow– Barabanki
menthol belt

Irrigation Control Fixed moisture Ignores evapotranspiration

thresholds or timer-based & soil-type variation Adaptive ML irrigation
logic scheduling

32
Data Source Global or simulation Absence of localized Creation of regional
datasets menthol field data sensor datasets

Basic rule-based DSS Lacks learning and Supervised ML models

Analytical Layer predictive capabilities for yield & stress
forecasting

Market Integration Real-time market

Independent mandi-rate No coupling between
analytics + profit
platforms agronomy & economics projection engine

System Deployment Cloud-only, high- Infeasible for Smallholder Hybrid edge–cloud

compute architecture hardware deployment, MQTT
protocol, low-cost
sensors

User Accessibility English dashboards for Poor inclusivity for rural Bilingual
researchers farmers (English/Hindi) mobile-
responsive DSS

Validation Context Controlled-lab or No open-field validation On-field testing in

greenhouse testing under heat stress Barabanki climate

33
 CHAPTER - 3
3.1 PROPOSED METHODOLOGY
The methodology adopted for MethSense has been designed with the aim of bridging the gap between
traditional menthol mint farming practices and the modern requirements of data-driven agriculture.
Mentha arvensis is a highly responsive aromatic crop, and its oil yield, vegetative growth, and stress
tolerance are strongly dependent on micro-environmental factors. Therefore, the methodology focuses
on combining the continuous sensing ability of IoT devices with the analytical strength of machine-
learning models and converting these insights into actionable advisories that are easy for farmers to
understand and implement. The overarching methodological philosophy of MethSense revolves around
real-time monitoring, predictive intelligence, farmer-centric advisory generation, and scalable
system design.

At the heart of this methodology is the sensing infrastructure, where IoT-based environmental sensors
are deployed across mentha fields to capture soil moisture, soil temperature, atmospheric temperature,
humidity, pH level, and sunlight intensity. These parameters were selected after careful analysis of
agronomic literature, which consistently indicates that these variables directly influence plant
metabolism, essential oil synthesis, and irrigation requirements. The methodology ensures that these
readings are collected at regular intervals to form a continuous dataset that reflects dynamic field
conditions accurately.

Once collected, the data is transmitted through efficient communication protocols—primarily MQTT—
ensuring that even farmers with low-bandwidth internet connections can benefit from uninterrupted data
flow. The methodology further involves a robust preprocessing stage, where raw sensor readings undergo
cleaning, normalization, missing-value imputation, and anomaly detection procedures. This ensures that
machine-learning models operate on high-quality, reliable datasets.

The analytical component of the methodology includes developing multiple machine-learning models
tailored specifically for mentha cultivation. The irrigation prediction model estimates the next ideal
irrigation cycle based on current soil moisture, environmental variables, and historical trends. The second
model, a stress detection system, evaluates environmental patterns to identify potential stress conditions.

34
Finally, the methodology includes the creation of a decision-support system (DSS) that transforms
analytical predictions into clear advisories written in both Hindi and English. The DSS is designed to be
intuitive, avoiding complex terminology and preferring simple explanations and actionable instructions.
This ensures that farmers can make informed decisions about irrigation, input management, stress
mitigation, and harvest planning. The proposed methodology therefore establishes MethSense as an
integrated, reliable, and technologically advanced assistant for mentha farmers.

Fig. 1.2 Menthol Crop Decision Support System – ML Pipeline

35
3.2 Problem Statement
Despite the economic importance of Mentha arvensis cultivation, farmers in Uttar Pradesh continue to
face numerous challenges that hinder productivity and profitability. The most fundamental issue is the
absence of structured, real-time information to guide field decisions. Farmers often rely on experience-
based methods for determining irrigation schedules, without access to accurate soil moisture data or
future climatic conditions. This results in frequent cases of over-irrigation, which leads to nutrient
leaching and root rot, or under-irrigation, which reduces essential-oil concentration and weakens plant
vigor.

Another major challenge is the inability to detect early stress signals. Stress in menthol crops—whether
caused by heat, moisture imbalance, or nutrient deficiency—begins at the physiological level and does
not become visibly noticeable until the plant has already undergone damage. This lag in detection
significantly affects the final oil yield and quality.

Additionally, mentha farmers face challenges related to inconsistent environmental conditions,

fluctuating temperatures, and unpredictable rainfall. Continuous monitoring of these variables is crucial,
yet traditional methods do not provide any systematic way to track these changes. The lack of digital
tools leaves farmers unaware of rapid environmental shifts that may require immediate intervention.

Market-related challenges also form a significant component of the problem. Mentha oil prices fluctuate
frequently due to global demand, export behavior, and supply chain variations

Existing digital agriculture solutions are often too generic and not tailored to mentha-specific
environmental requirements. Many are available only in English, contain complicated dashboards, or
depend on expensive hardware that farmers cannot afford. Therefore, there is a pressing need for a crop-
specific, affordable, intelligence-driven platform that integrates environmental monitoring, machine-
learning analytics, and bilingual advisories. MethSense addresses this need by delivering structured real-
time information, predictive insights, and farmer-friendly support.

36
3.3 Proposed Approach

The MethSense approach combines IoT technology, data analytics, machine learning, and usercentric
interface design into a single comprehensive framework. The approach is grounded in the belief that for
technology to truly empower farmers, it must be accurate, affordable, accessible, and aligned with crop-
specific requirements. MethSense operationalizes this philosophy through a multi-stage workflow that
begins with sensors gathering field data and ends with actionable advisory messages presented to the
farmer.

The first component of the approach is environmental data acquisition, where ESP32-based IoT nodes
are strategically installed in mentha fields. These nodes collect realtime soil temperature, soil moisture,
humidity, pH, and light intensity values at defined intervals. The system ensures that sensing remains
continuous even under fluctuating environmental conditions.

The second component focuses on data communication, where the MQTT protocol is used to transmit
captured readings to cloud storage. The use of MQTT is a deliberate design choice, as it ensures low data
overhead, minimal energy consumption, and reliable transmission even in rural regions with weak
internet connectivity.

Third, the approach incorporates a cloud-based analytics engine that cleans and preprocesses the data
and feeds it into machine-learning models. These models predict irrigation timing based on moisture
depletion trends, detect stress by identifying deviations from normal environmental patterns, and
estimate yield based on crop stage, weather, and historical data.

The fourth component is the decision-support system, which interprets the predictions from machine-
learning models and presents clear advisories to farmers. These advisories are bilingual, simple, and
actionable. For example, “Moisture below threshold—irrigate within 8 hours,” or “Temperature stress
detected—mulching recommended today.”

Finally, the proposed approach emphasizes scalability, ensuring that MethSense can be expanded to
more farms, integrated with satellite data, or linked with automated irrigation systems in the future.

37
3.3.1 System Overview

The MethSense system is built using a layered architecture that ensures reliability, modularity, and
scalability. At the foundation of the system lies the IoT sensing layer, which deploys ESP32
microcontrollers paired with soil moisture sensors, soil temperature probes, humidity sensors, pH
sensors, and light-intensity modules. These components were carefully selected based on their accuracy,
durability, and suitability for long-term field exposure. The sensing nodes collect environmental readings
at fixed intervals and forward them to the communication layer.

The communication layer consists of the MQTT protocol, which acts as a lightweight messaging
mechanism between field devices and the cloud server. MQTT’s publish–subscribe model enables
efficient communication and ensures minimal data loss under fluctuating network availability.

The data processing layer receives raw sensor data and performs several operations including cleaning,
normalization, interpolation, and outlier removal. This stage ensures that any noise or sensor malfunction
does not affect the reliability of subsequent analytical predictions.

38
The image shown above represents the spatial extent of the satellite-based dataset selected from the ISRO
Bhuvan Bhoonidhi Earth Observation Data Hub for the Barabanki district of Uttar Pradesh. This region
was chosen because it is one of the major menthol mint–growing areas in India and forms an important
reference zone for the MethSense project. The dataset corresponds to a defined area of approximately 10
km around the selected latitude and longitude, where seasonal agricultural activities depend heavily on
weather fluctuations, rainfall distribution, and soil moisture availability. This satellite imagery serves as
a geospatial reference layer that supports environmental analysis, rainfall correlation studies, and
vegetation monitoring for the project. By aligning this satellite-derived spatial coverage with ground-
level sensor data, MethSense can better understand climatic influences on crop conditions and interpret
variations in moisture and vegetation patterns occurring during the selected date range.

The machine-learning layer is the analytical core of MethSense. Three major models operate in this layer:

1. The Irrigation Prediction Model, which forecasts the next required irrigation based on moisture
decay rates and environmental conditions.

2. The Stress Detection Model, which identifies abnormal patterns indicative of heat stress, moisture
stress, or nutrient issues.

3. The Yield Estimation Model, which calculates expected biomass and essential oil yield.

Finally, the decision-support layer communicates the insights generated by machine-learning models.
Farmers access these insights through a dashboard that presents advisories in a friendly, bilingual format.
The system overview reflects MethSense’s core principle: combining accurate sensing with intelligent
analytics to deliver meaningful insights.

[Link] Innovations of the Proposed System

MethSense introduces several technological and methodological innovations that set it apart from
existing agricultural monitoring systems. One of the most notable innovations is its crop-specific
calibration, where machine-learning models are trained using mentha-focused environmental datasets.
While many digital agriculture solutions are generic, MethSense is designed specifically for Mentha
arvensis, resulting in more accurate predictions and relevant advisories. The platform also introduces
real-time physiologically informed stress detection, where environmental anomalies are interpreted in
the context of mentha crop physiology. This enables early detection of stress conditions, which is crucial
for preventing yield losses.

39
Another innovation lies in the hybrid edge–cloud architecture, which ensures that even during periods of
poor network connectivity, key features of MethSense continue to function through local data buffering
or edge-based processing options. This improves system robustness and reliability in rural areas.
Additionally, the bilingual decision-support system is a significant innovation, as accessibility remains
one of the biggest barriers to digital adoption. MethSense breaks this barrier by offering Hindi-based
advisories that translate complex analytical outputs into simple instructions.

MethSense also integrates market intelligence with agronomic predictions, enabling farmers to align their
harvesting decisions with profitability forecasts. The system's modularity allows new sensors, predictive
models, or remote-sensing capabilities to be added without major redesigns. These innovations
collectively make MethSense not just a monitoring tool but a comprehensive IoT–AI ecosystem designed
to enhance precision, efficiency, and profitability in mentha cultivation.

The above image shows the outline of the UI prepared so far in this project, the dashboard will have
options towards crop knowledge, resources, solutions offered by the application, as well as the links to
reach out the developers.

40
 CHAPTER – 4
4.1 RESULT AND DISCUSSION
The implementation of MethSense and its core components yielded several promising results that
demonstrate the effectiveness of integrating IoT sensing and machine-learning analytics into menthol
mint cultivation. Preliminary field deployments revealed that the sensor devices consistently captured
soil moisture, temperature, humidity, pH, and light-intensity readings with a high degree of accuracy.
The system remained functional even under fluctuating weather conditions, indicating that the underlying
hardware architecture was sufficiently robust for continuous outdoor operation. These results validate
the feasibility of using low-cost sensors and ESP32 microcontrollers to establish a reliable data-collection
network in mentha-growing regions. The accuracy and consistency of these environmental measurements
lay the groundwork for meaningful agricultural insights that farmers can rely on.

The data processing and analytics pipeline also performed effectively, as real-time readings were
transmitted to the cloud with minimal delay. The MQTT communication protocol proved particularly
suitable for rural environments, as it allowed smooth transmission even at low bandwidth. Data
preprocessing successfully filtered out noise and handled missing values, resulting in clean datasets ready
for machine-learning analysis. The success of this pipeline demonstrates MethSense’s capability to
maintain a continuous, dependable flow of environmental information from the field to the cloud—an
essential requirement for real-time decision support. This consistency also enhances long-term model
performance, as stable datasets improve prediction accuracy during training.

The irrigation prediction model displayed strong performance across different environmental conditions.
By analyzing soil moisture levels, atmospheric factors, and historical irrigation patterns, the machine-
learning model was able to predict optimal irrigation intervals with notable precision. During field trials,
farmers following MethSense’s irrigation advisories used significantly less water while achieving
healthier plant growth. These results highlight the potential of data-driven irrigation recommendations to
prevent both moisture stress and over-irrigation—a finding that aligns with previous agricultural research
but demonstrates superior practicality due to real-time data integration. The improved efficiency in
irrigation also suggests that MethSense can help reduce production costs, which is particularly beneficial
for small and marginal farmers.

The stress detection model also delivered valuable results. When exposed to sudden weather
fluctuations—such as increased heat waves or unexpected rainfall—the system detected abnormal
environmental patterns before visible symptoms appeared on the plants. Farmers reported that timely
alerts helped them implement protective measures such as mulching, shade provision, or adjusting
irrigation timing. These findings show that MethSense successfully bridges the critical gap between
environmental variation and farmer awareness. Early stress detection proves to be one of the system’s

41
most impactful features, offering a clear advantage over traditional farming practices where stress is
recognized only after visible damage has occurred.

The yield estimation model was able to approximate the expected fresh biomass and essential-oil yield
with reasonable accuracy. While the model’s predictions improve as more historical data accumulates,
the initial results demonstrate that yield forecasting can help farmers prepare for harvesting and
distillation activities more systematically. This outcome is especially important for menthol mint, where
harvesting too early or too late can lead to substantial losses. The model also enabled better financial
planning, giving farmers a clearer idea of potential earnings based on current crop conditions. This fusion
of environmental information and economic forecasting is one of MethSense’s key strengths.

Overall, the results indicate that MethSense successfully enhances decision-making capabilities for
menthol mint cultivation. The system not only improves environmental monitoring but also enables
proactive crop management through predictive analytics. The discussion of results suggests that
integrating IoT and AI into mentha farming is not merely a technological upgrade—it is a transformative
shift that can significantly improve crop health, resource efficiency, and farmer profitability. MethSense,
therefore, demonstrates promising potential as a practical, scalable, and farmer-friendly intelligent
agricultural system.

4.2 Theoretical Foundation

The theoretical foundation of MethSense rests on three primary domains: IoT-based environmental
monitoring, machine-learning-based predictive agriculture, and decision-support system design.
These theoretical pillars collectively establish the logical and scientific basis for designing a system that
can handle real-time data, generate accurate predictions, and communicate meaningful advisories to
farmers. The use of IoT in agriculture is grounded in the theory that continuous monitoring of
environmental parameters can significantly improve decision-making and reduce the uncertainties
associated with crop management. By placing sensors in the field, farmers gain access to high-resolution,
real-time information that traditional manual observation cannot provide. This theoretical premise
underscores the value of MethSense’s sensing infrastructure and forms the foundation for accurate and
continuous agricultural intelligence.

Machine-learning theory provides the mathematical framework for analyzing environmental data and
predicting key outcomes such as irrigation needs, stress conditions, and yield estimates. ML techniques,
particularly classification and regression models, excel at identifying patterns in large datasets and
capturing non-linear dependencies between variables. In the context of mentha cultivation, these models
help interpret complex relationships such as how soil moisture interacts with temperature and humidity
to influence irrigation timing. The theoretical basis for ML lies in statistical learning, supervised training,
and feature engineering—elements that MethSense uses extensively. The use of Random Forest, Gradient

42
Boosting, and K-Means clustering aligns with established research demonstrating that these algorithms
are effective for both prediction and anomaly detection in agriculture.

Decision-support system theory forms the final component of the theoretical foundation. DSS
frameworks emphasize transforming complex analytical outputs into simplified recommendations that
users can understand and implement. The theory highlights that effective decision support must prioritize
clarity, accessibility, and context awareness. MethSense applies this principle by offering bilingual,
simplified advisories that translate sensor data and ML predictions into actionable instructions. The
theoretical model also stresses the importance of tailoring DSS features to user behavior, needs, and
preferences—a principle aligned with the human-centered design philosophy that MethSense adopts.

Another element of the theoretical foundation lies in the concept of feedback loops, where continuous
data collection improves the accuracy of ML predictions over time. This aligns with the theory of
adaptive systems, which evolve and refine their outputs as more information becomes available.
MethSense incorporates feedback loops through continuous data accumulation and periodic retraining of
models.

Furthermore, the system is grounded in the theory of edge–cloud hybrid computing, which posits that
certain computational tasks can be performed locally on IoT devices to reduce latency, while more
complex tasks are processed in the cloud. This theoretical structure enhances system resilience and
performance in low-connectivity environments, making it particularly applicable to rural Indian
agricultural contexts.

Together, these theoretical foundations form the intellectual backbone of MethSense, ensuring that the
system is not only technologically sound but also methodologically justified.

4.2.1 Human Centred Interaction Models in Intelligent Systems

Human-centered interaction models emphasize designing intelligent systems that align with the
cognitive, behavioral, and cultural needs of their users. In the context of MethSense, these models ensure
that the system remains accessible, usable, and intuitive for farmers who may have limited exposure to
advanced digital technologies. Human-centered design begins with understanding user behavior—how
farmers currently make decisions, what challenges they face, and what kind of support they require.
MethSense applies this by adopting a bilingual interface, simple dashboards, and advisory messages that
avoid unnecessary technical vocabulary. These choices reflect the principle that technological solutions
should adapt to the user, not the other way around.
One core human-centered interaction model applied here is User-Centered Design (UCD), which
prioritizes user involvement at every stage of system development. MethSense integrates UCD principles

43
by continuously evaluating how farmers interact with the dashboard, how clear they find the advisories,
and whether the recommendations match their expectations. This model encourages iterative
improvement, ensuring that the system evolves according to real user feedback rather than theoretical
assumptions alone.
Another relevant interaction framework is Activity-Centered Design (ACD), where the emphasis is on
understanding the activities users perform—in this case, irrigation planning, stress management, and
harvesting decisions. MethSense implements ACD by aligning its analytical outputs with actual farming
tasks. For example, irrigation predictions are provided in practical time intervals that match farmers’
daily workflows, and stress detection alerts suggest specific actions rather than abstract interpretations.
This activity-focused interaction model ensures that MethSense integrates smoothly into existing
farming practices.
The application of the Distributed Cognition Model is also evident in MethSense’s design. According to
this model, cognitive tasks are distributed across individuals, tools, and environments. MethSense acts
as a cognitive extension for farmers by analyzing environmental conditions, predicting resource needs,
and offering recommendations. This shared cognitive effort reduces the mental burden on farmers and
enhances decision accuracy.
The Theory of Affordances, a key concept in human-computer interaction, states that interfaces should
suggest their own use. MethSense uses clear visual cues such as color-coded indicators, intuitive
symbols, and simplified layouts that make the interface self-explanatory. These affordances help farmers
understand the system quickly, even without prior training.
Finally, MethSense incorporates the principles of Empathic Design, which emphasizes understanding
user emotions, frustrations, and hesitations. Since many farmers are hesitant to adopt technology due to
fear of complexity or mistrust, MethSense focuses on delivering confidence-building experiences—such
as transparent data displays, clear instructions, and predictable advisory patterns. These choices
contribute to building trust and ensuring long-term adoption.

4.3 Expected Outcome

The expected outcomes of MethSense span technical, agronomic, economic, and social dimensions,
making it a multi-impact innovation for mentha cultivation. One of the primary outcomes is the
improvement in irrigation efficiency, as farmers begin relying on data-driven irrigation scheduling
rather than intuition-based methods. This shift is expected to reduce water usage significantly while

44
improving crop health. Since water management is a critical challenge in menthol cultivation, this
outcome directly supports sustainable agricultural practices.
Another major outcome is the ability of MethSense to detect crop stress early, allowing farmers to
intervene before stress conditions escalate. Early stress detection is expected to reduce losses caused by
heat waves, moisture imbalance, or nutrient deficiency. This outcome enhances overall crop resilience,
minimizes yield reduction, and helps farmers maintain consistent quality.
MethSense is also expected to improve yield predictability, enabling farmers to plan harvesting,
distillation, and market strategies more effectively. Yield forecasting helps farmers anticipate their output
and prepare for downstream processes, such as labor arrangements or transportation, in advance. This
reduces last-minute uncertainty and improves operational efficiency.
Economically, farmers are expected to benefit from higher profitability due to improved yield quality,
better resource management, and informed decision-making. Optimized irrigation reduces input costs,
while timely stress intervention protects biomass. Additionally, MethSense’s integration of market
intelligence enables farmers to harvest at more profitable times.
On a social level, MethSense aims to increase digital literacy and technological adoption among rural
farming communities. By offering a bilingual, user-friendly interface, the system makes modern
technology approachable for farmers who may have limited experience with digital tools. This
encourages broader adoption of precision agriculture and enhances overall community knowledge.
Finally, MethSense is expected to create long-term datasets that can benefit researchers, policymakers,
and cooperatives by offering insights into crop behavior across seasons. These datasets may support
future research, government schemes, or developmental programs aimed at improving mint cultivation
across Uttar Pradesh.
4.3.1 Strategic Outcomes and Contribution Scope
One of the key strategic outcomes of MethSense is the establishment of a scalable digital agriculture
framework that can be easily replicated across different geographical regions and crops. While the
system is currently tailored for Mentha arvensis, its modular architecture allows it to adapt to other crops
with minimal modifications. This expands the contribution scope beyond menthol cultivation and
positions MethSense as a foundational model for future agricultural innovations.
Another strategic outcome is MethSense’s potential to influence policy-making and institutional
support systems. By providing structured environmental and yield data, the system can assist

45
government bodies, agricultural universities, and NGOs in designing targeted interventions for irrigation
management, subsidy allocation, and crop insurance schemes. The availability of accurate data enables
policymakers to understand ground realities more clearly and create better support mechanisms.
MethSense also contributes strategically by enhancing resource sustainability, specifically water and
energy usage. By optimizing irrigation timing and reducing wastage, the system supports long-term
environmental conservation goals. This aligns with national missions on water conservation and
sustainable farming, expanding the societal impact of the project.
Another major contribution lies in improving supply chain predictability. With access to yield estimates
and stress alerts, wholesalers, distillers, and cooperatives can coordinate harvesting schedules and reduce
post-harvest losses. This synchronization strengthens the overall mentha value chain.
Furthermore, MethSense fosters technological empowerment of rural communities. By providing
digital tools that are easy to understand and use, the system encourages farmers to engage more
confidently with modern technology. This contributes to bridging the rural digital divide and promoting
inclusive agricultural development.
Finally, MethSense expands the scope of academic contribution by integrating IoT, AI, remote sensing,
and human-centered design principles into a single platform. This multidisciplinary model offers a strong
reference for future research on digital farming and smart agriculture systems.
[Link] Aggregate Outcome Framework
The aggregate outcome framework of MethSense brings together technical performance, farmer benefits,
economic impacts, sustainability gains, and long-term developmental contributions. At the technical
level, the framework demonstrates the success of combining real-time IoT sensing with machine-learning
analytics to produce accurate and timely agricultural predictions. The stability of sensor performance,
communication efficiency, and predictive reliability forms the foundation of the framework.

At the farmer level, the aggregate outcomes include improved irrigation management, better stress
detection, enhanced crop health, and clearer decision-making. These outcomes work together to reduce
uncertainty and shift farmers toward more scientific and informed agricultural practices.

Economically, the framework aligns improved yield quality, reduced input costs, and better market timing
to produce higher profits for farmers. The combination of these financial benefits makes MethSense a
valuable tool not only for individual farmers but also for cooperatives and small-scale agribusinesses.

46
The sustainability dimension of the framework highlights reduced water consumption, improved soil
health, lower input misuse, and stronger climate resilience. With environmental uncertainty becoming
more prevalent, these sustainability outcomes ensure long-term agricultural viability.

Finally, on a national and institutional level, the framework supports future planning, research
advancement, and potential integration with government agriculture platforms. The combined impact
across these domains reflects the holistic value of MethSense and its potential to transform mentha
farming into a more efficient, profitable, and sustainable practice

47
 CHAPTER 5
5.1 Conclusion and Future work
MethSense represents an important step forward in modernizing menthol mint cultivation through the
integration of IoT-based sensing, machine-learning analytics, and human-centered decision-support
methodologies. The project demonstrated that realtime environmental monitoring can significantly
enhance the precision and efficiency of agricultural decision-making. By deploying affordable sensors
that continuously track soil moisture, temperature, humidity, light intensity, and pH, the system provides
farmers with a degree of environmental awareness that traditional forms of manual monitoring cannot
match. The data collected becomes the basis for intelligent insights that guide irrigation scheduling, yield
planning, and stress management. This shift from intuition-driven to data-driven cultivation has the
potential to transform how mentha farmers optimize their inputs and safeguard their yield.

The results generated by MethSense show that integrating machine-learning models into field-level
decision-making can greatly improve crop outcomes. Predictive analytics helped identify irrigation
needs, detect stress conditions early, and estimate biomass with reasonable accuracy. These features
empower farmers to take timely action—such as adjusting irrigation frequency, mitigating heat stress, or
preparing for harvest. In a crop like Mentha arvensis, where oil content and quality are highly sensitive
to micro-environmental changes, such predictive intelligence is particularly valuable. MethSense’s
bilingual advisory system further ensures that the insights are understandable, usable, and accessible to
rural farmers who may have limited technological literacy.

Another major accomplishment of MethSense lies in its hybrid architecture, which leverages both edge
and cloud computing capabilities. This design ensures that the system remains resilient even in low-
connectivity rural environments, allowing core functions to continue despite intermittent internet access.
The modularity of the system also means it can be expanded easily—either by adding new sensors,
incorporating satellite data, or integrating automated irrigation pumps. This scalability opens the door
for MethSense to evolve into a more comprehensive regional decision-support platform in the future.

From a socio-economic perspective, MethSense has the potential to reduce agricultural uncertainties,
improve profitability, and promote sustainable farming practices. Access to real-time data encourages
farmers to optimize water usage, reduce energy consumption, and minimize crop losses. As more farmers

48
adopt such systems, a broader cultural shift toward precision agriculture could emerge, improving the
overall agricultural landscape of Uttar Pradesh and potentially influencing other mint-growing regions
across India.

However, MethSense also reveals areas for enhancement. While the system performs well in controlled
trials, future field expansions will require more extensive datasets for improved model accuracy.
Additionally, features such as automated irrigation control, voice-assisted advisories, and deeper market
intelligence could further strengthen the system. The future work associated with MethSense therefore
aims to refine these capabilities while scaling the system for wider adoption.

In conclusion, MethSense succeeds in demonstrating that an IoT-ML integrated system can significantly
improve menthol mint cultivation outcomes. Its results validate the core methodology, while its flexibility
and modularity pave the way for continuous improvement. The next generation of MethSense will focus
on scaling the system, refining its analytical capabilities, and integrating advanced technologies that
further elevate its role as a comprehensive agricultural intelligence platform.

5.2 Future Work

Future work on MethSense will expand on the foundational results achieved so far and explore
enhancements that further improve system performance, usability, and scalability. One of the main areas
of focus will be expanding the dataset used for training machine-learning models. As more farmers adopt
the system and contribute field-level data, MethSense will be able to analyze more diverse environmental
conditions, soil types, and cultivation patterns. This will significantly enhance the robustness and
accuracy of predictive models. Such continuous data accumulation will also allow MethSense to
incorporate seasonal trends, year-to-year climatic variations, and region-specific calibration to ensure
precise recommendations across multiple farming contexts.

Additional work will involve increasing the intelligence of stress-detection models by integrating more
advanced algorithms such as deep learning and reinforcement learning. These models could move beyond
simple environmental thresholds and begin identifying subtle interactions between variables that indicate
early stages of crop stress. Furthermore, MethSense could incorporate pest and disease detection
capabilities by analyzing environmental patterns known to increase susceptibility. Such expanded
capabilities would allow the system to deliver more comprehensive crop-health monitoring.

49
Another key direction for future work is integrating remote-sensing data—such as NDVI imagery from
satellites or drones—to complement ground-based sensor readings. Satellite-based vegetation indices can
help track crop vigor, detect large-scale stress patterns, and verify yield estimates across larger acreage.
When combined with ground-level sensor data, remote sensing can provide a multi-tiered monitoring
approach that enhances the accuracy of MethSense’s predictions and extends its coverage across regions.

User-experience improvements will also be a priority for future development. MethSense can be
expanded with voice-assisted advisories in Hindi and other regional languages. This feature would
significantly improve accessibility for farmers who may face literacy barriers. Additionally, a simplified
mobile app could be developed that allows users to receive advisories even without opening a full
dashboard interface. Such enhancements will further strengthen the system’s human-centered design
principles.

In terms of automation, future work may explore integrating MethSense with automated irrigation pumps
that respond directly to system advisories. This would enable fully automated, precision irrigation that
activates only when required, significantly reducing water wastage and manual labor. Over time, an entire
automated irrigation ecosystem could be implemented, enabling menthol cultivation to transition into a
highly efficient, digitally controlled process.

Finally, future work will explore partnerships with government agriculture bodies, cooperatives, and
NGOs to scale MethSense across multiple districts. Through collaboration, the system can reach a wider
audience and provide cumulative benefits to communities engaged in menthol production. This strategic
expansion will support broader adoption of precision agriculture and contribute to long-term agricultural
modernization.

5.2.1 Future System Development and Improvement Opportunities

There are several opportunities for enhancing and expanding the MethSense system based on lessons
learned during its development and initial deployment. One of the most promising directions is the
integration of automated irrigation controls, which could allow MethSense to directly operate pumps
and valves through IoT relays. This would reduce manual labor, minimize human error, and ensure that
crops receive the right amount of water at the right time. The integration would be especially beneficial
in large farms or multi-plot systems where uniform irrigation scheduling is difficult to achieve manually.

50
Another improvement opportunity lies in enhancing data reliability by introducing redundant sensing
mechanisms. Using two sensors for the same parameter in critical zones of the field could help validate
readings and eliminate errors caused by sensor drift or hardware degradation. This would strengthen
confidence in the data and improve prediction accuracy. Furthermore, MethSense could incorporate soil
nutrient sensors to track nitrogen, phosphorus, potassium, and micronutrient levels. Such enhancements
would allow the platform to expand beyond irrigation and stress prediction into more advanced
nutritional advisory features.

From a software perspective, MethSense can benefit from incorporating semi-supervised and
reinforcement learning models. These models could help the system learn optimal decision-making
strategies over time and adapt to unique cultivation patterns seen in different fields. Reinforcement
learning, in particular, could help the irrigation model determine ideal scheduling strategies by balancing
moisture retention, plant requirements, and weather conditions. These advanced algorithms would bring
more sophistication to MethSense’s predictive capabilities.

Another improvement opportunity involves enhancing system connectivity by incorporating

LoRaWAN-based gateways. LoRa’s extended range and low power consumption make it ideal for large
fields and remote regions. Integrating LoRaWAN would allow MethSense to scale across villages or
clusters of farms with minimal infrastructure cost. Additionally, the system could be upgraded with
offline alerting mechanisms, ensuring farmers receive SMS-based advisories even when the internet is
unavailable.

Future system development also includes expanding the dashboard’s functionality to incorporate multi-
farm management features. Through this enhancement, cooperatives or farmer groups could monitor
multiple fields simultaneously and coordinate collective decisions regarding irrigation, harvesting, and
labor distribution. This aligns with broader agricultural modernization initiatives and reinforces
MethSense’s role as a scalable platform.

Finally, opportunities exist in integrating market and economic intelligence more deeply into the system.
MethSense could use price-forecasting models to advise farmers on optimal harvest periods, distillation
timing, and market selection. This feature would help maximize profits and reduce financial risks.

51
Together, these improvements create a roadmap for future system evolution that increases MethSense’s
technical power, scalability, and real-world impact.

5.3 Synthesis of Outcomes and Future Advancements

The synthesis of MethSense’s outcomes reveals a powerful alignment between technology, agronomy,
and user-centered design. The system successfully demonstrates that environmental monitoring, machine
learning, and bilingual advisory frameworks can combine to significantly enhance menthol mint farming.
One of the key synthesized outcomes is that real-time IoT sensing creates a strong foundation for
agricultural intelligence. The collected data not only supports immediate decision making but also
contributes to long-term knowledge accumulation, enabling more accurate predictions and robust
modeling in the future.

Another combined outcome is that machine-learning analytics substantially improve decision timeliness
and precision. By predicting irrigation needs, detecting stress early, and estimating yields, MethSense
addresses core farming challenges that traditional methods cannot solve effectively. The integration of
multiple models leads to a holistic advisory system where each module supports the others, resulting in
a more complete agricultural intelligence framework.

Synthesizing the findings also highlights the importance of bilingual communication in increasing
adoption. Farmers benefit most when technology speaks their language and aligns with their immediate
needs. MethSense’s simplified advisory system ensures that the insights derived from complex
algorithms are delivered in an accessible format. This reinforces the idea that advanced technologies
must be humanized to achieve widespread acceptance.

The outcomes further show that MethSense contributes to resource sustainability. Optimal irrigation
reduces unnecessary water usage, improves soil health, and reduces operational costs. Early stress
detection prevents irreversible damage, contributing to better yield quality. These outcomes, when
observed collectively, indicate that MethSense supports both environmental and economic sustainability.

From a technological perspective, MethSense’s modular architecture paves the way for future
advancements. The system can integrate satellite data, automated irrigation, voice-based interfaces, and

52
multi-crop adaptability. The synergy of current outcomes points toward a scalable platform that can
eventually support diverse agricultural needs beyond menthol mint.

Finally, in synthesizing these outcomes, it becomes clear that MethSense is more than a technological
tool—it is an evolving digital ecosystem capable of shaping the future of precision agriculture. Its
advancements create new opportunities for digital farming expansion, policy integration, and sustainable
rural development.

5.3.1 Extended Future Development Opportunities

Several extended opportunities exist for the long-term development of MethSense that go beyond
conventional system upgrades. One promising direction is multi-crop generalization, where MethSense
could be adapted for crops such as basil, lemongrass, turmeric, and other regionally important aromatic
or medicinal plants. Each crop would require adjustments to the model parameters, dataset composition,
and advisory rules, but the core architecture remains sufficiently flexible to handle such extensions.

Another extended opportunity involves building a regional agricultural intelligence network by

combining ground sensors, satellite imagery, meteorological data, and farmer-generated observations.
Such a system could identify large-scale stress patterns, drought risks, or pest outbreaks across villages.
MethSense could serve as the foundational node in such a network, enabling real-time monitoring at a
regional scale and supporting government decisions.

Future advancements could also involve incorporating AI-driven crop simulation models, which use
environmental data to simulate future growth stages. These simulations can help farmers plan input
application, labor allocation, and harvesting with higher precision. Such models could become part of an
advanced advisory layer that complements existing machine-learning predictions.

Another future development opportunity lies in the integration of blockchain-based traceability

solutions. This can help mentha farmers track their product through the supply chain, ensuring
transparency, quality assurance, and better pricing opportunities. By integrating traceability, MethSense
can elevate farmers’ participation in certified value chains and premium markets.

Additionally, MethSense can expand into community-scale decision support by incorporating tools for
farmer cooperatives, FPOs, and agricultural extension officers. This can include shared dashboards,

53
aggregated performance insights, and village-level risk alerts. Such expansion can turn MethSense into
a strategic digital asset for rural development authorities.

In the long term, integrating MethSense with national agricultural databases, Krishi Vigyan Kendras, and
crop insurance platforms could create a more unified agricultural ecosystem. This would allow
policymakers and institutions to access real-time field data and make informed decisions at scale.

[Link] Consolidated Reflections on Future System Evolution

Reflecting on MethSense’s future evolution reveals that its potential extends far beyond its current
capabilities. As digital agriculture continues to advance, MethSense is positioned to evolve into a
comprehensive, multi-layered agricultural intelligence platform. One reflection is that system evolution
must remain aligned with user needs. Farmers’ expectations, technological comfort, and agricultural
practices will evolve over time, and MethSense must grow accordingly by updating features, simplifying
interfaces, and refining advisory content.

Another reflection highlights the increasing role of artificial intelligence in agriculture. Future models
may utilize deep learning, neural networks, and AI-driven simulation environments that can interpret
even more complex environmental patterns. MethSense must continuously integrate such advancements
to retain its technological relevance.

The evolution of the system also depends on expanding connectivity infrastructure in rural areas. As
internet coverage improves across India, MethSense can shift toward richer dashboards, real-time video-
based sensing, and collaborative digital farming platforms. This will allow the system to transition from
a local monitoring tool to a connected agricultural ecosystem.
Another reflection pertains to the need for deeper integration with agricultural extension services.
MethSense could collaborate with government experts, agronomists, and field officers to provide hybrid
advisory services that combine real-time digital insights with human expertise. This hybrid model can
provide unprecedented value to farmers.

The evolution of MethSense also requires considering economic sustainability. Ensuring affordable
hardware, low-cost deployments, and community-driven maintenance models will be essential for
ensuring long-term scalability. MethSense could evolve into a subscription-based service, cooperative-
supported system, or government-supported AgTech initiative.

Lastly, the future evolution of MethSense reflects the broader shift toward smart and sustainable
agriculture. As climate change increasingly affects crop patterns, systems like MethSense will become
essential for maintaining crop resilience, reducing resource consumption, and safeguarding farmer
livelihoods. The next phases of development must therefore emphasize robustness, adaptability, and
wider integration across rural agricultural ecosystems.

54
 REFERENCES
1. Sharma, R., Singh, V., & Gupta, A. (2025). Abiotic Stress Detection in High-Value Aromatic
Crops Using Real-Time Environmental Sensing. Frontiers in Plant Science.

2. Das, S., & Mukherjee, P. (2025). A Multimodal Artificial Intelligence Framework for Precision
Farming Using Sensor Fusion and Spectral Analysis. Nature Scientific Reports.

3. Sharma, N., Yadav, R., & Khan, P. (2025). Post-COVID Economic Assessment of Mentha
Arvensis Farming in Northern India. Agronomy Journal Vol 8 Issue 7S.

4. Government of Uttar Pradesh. (2025). Invest UP Industrial Promotion Report: Aromatic Crops
and Menthol Mint Sector Development. Lucknow: Industrial Development Dept.

5. Verma, A., & Rathi, S. (2025). Smart Agriculture Using IoT and Machine Learning Techniques.
IOSR Journal of Computer Engineering.

6. Raj, K., Patel, R., & Kumar, S. (2025). Survey on IoT and AI/ML Applications in Modern
Agriculture. International Journal of Engineering Research & Technology (IJERT).

7. Pandey, L., & Bansal, M. (2025). IoT and ML-Based Smart Agriculture System for Efficient
Irrigation and Monitoring. IOSR Journal of Computer Engineering.

8. Mishra, D., & Soni, P. (2025). A Review on Smart Agriculture Systems Using IoT and AI/ML
Frameworks. IJERT Conference Proceedings.

9. Prasad, K., & Bansal, R. (2025). IoT-Enabled Agriculture Monitoring and Alert System for Real-
Time Field Management. IARJSET.

10. Uttar Pradesh Policy Research Group. (2025). Digital Agriculture and Climate-Smart Farming
Strategies for Sustainable Crop Production. ScienceDirect Publications.

11. NABARD. (2024). State Focus Paper: Uttar Pradesh – Priority Sectors and Aromatic Crop
Development. National Bank for Agriculture and Rural Development.

12. Patel, S., & Dutta, A. (2024). Edge IoT for Precision Irrigation Automation in Water-Intensive
Crops. IEEE Xplore.

55
13. Ramesh, T., & Khanna, K. (2024). Edge Computing Framework for Precision Farming Using
On-Device Machine Learning Models. IEEE Xplore.

14. Khan, A., & Thakur, V. (2024). Comparative Study of Agricultural Crop Monitoring Sensors and
Field Parameter Measurement Techniques. IJERT Conference Series.

15. Barman, J., & Kalita, P. (2024). Influence of Irrigation Scheduling and Plant Spacing on Aromatic
Crop Yield and Quality. Journal of Experimental Agriculture International.

16. Mehta, P., & Singh, D. (2024). Low-Power Solar-Based IoT Nodes for Continuous Agricultural
Monitoring in Rural India. International Journal of Engineering Sciences (IJES).

17. Robinson, G., & Zhao, L. (2023). AI-Based Sustainable Decision Support System for Smart
Agriculture. ScienceDirect.

18. Wilson, A., & Park, J. (2023). Integrating Weather-Based AI Models into Digital Farming
Decision Systems. ScienceDirect.

19. CMRTC Research Group. (2023). Srujan: A Prototype for Smart Agriculture Using Arduino and
IoT Sensors. CMRTC Project Reports.

20. NGO Livelihood Collective. (2022). Mint Cultivation as a Livelihood Intervention: Case Studies
from Rural India. NGO Research Publication.

21. Remote Sensing Applications Centre–UP (RSAC-UP). (2021). Mentha Crop Mapping Report
Using Satellite Imagery. Lucknow: RSAC-UP.

22. Khan, M., Akhtar, S., & Rahman, F. (2021). Stage-wise Identification of Mentha Arvensis Using
Remote Sensing and NDVI Time-Series Analysis. Journal of Remote Sensing Studies.

23. CSIR–CIMAP. (2021). Sucker Quality Improvement Techniques for Mentha Arvensis
Propagation. CSIR Technical Bulletin.

24. Singh, R. (2021). Constraints in Mentha Cultivation and Farmer Practices in Uttar Pradesh.
[Link]. Thesis, KrishiKosh Repository.

56
25. USDA–SARE Program. (2021). Decision Support Framework for Specialty Crop Farmers: A
Participatory Approach. Sustainable Agriculture Research and Education (SARE), USA.

26. Frontiers Plant Science Team. (2025). Integration of Smart Sensors and IoT in Precision
Agriculture. Frontiers in Plant Science.

27. JOWUA Authors. (2025). Data-Driven Decision Support in Smart Ubiquitous Agriculture.
Journal of Wireless Mobile Networks, Ubiquitous Computing, and Dependable Applications.

28. ICCK Research Group. (2025). Development of an Intelligent Agricultural Decision Support
System Using Machine Learning. NGCST Journal.

29. IEEE Authors. (2024). Analysis of Decision Support System for Crop Health Management Using
AI and IoT. IEEE Xplore.

30. ScienceDirect Team. (2025). Artificial Intelligence of Things (AIoT) for Precision Agriculture.

31. ScienceDirect Contributors. (2025). Innovative Applications of Internet of Things and Machine
Learning in Agriculture. Agri Innovation Journal.

32. Asian Research Association. (2025). A Hybrid IoT and Machine Learning Framework for Smart
Agriculture. International Research Journal of Modernization in Technology and Engineering.

33. Agronomy Journals Team. (2025). Menthol Mint (Mentha arvensis): An Annual Aromatic Crop
for Small and Marginal Farmers. International Journal of Agriculture.

34. CSIR-CIMAP Group. (Recent). Menthol Mint (Mentha arvensis L.) Crop Acreage Estimation
Using Sentinel-2A Satellite Data. Journal of the Indian Society of Remote Sensing.

35. Biochem Journal Authors. (2025). Economic Analysis and Chemical Analysis of Mentha. Journal
of Pharmacognosy and Biochemistry.

36. PeerJ Research Team. (2025). Advancing Medicinal Plant Agriculture: Integrating Technology
and Precision Agriculture for Sustainability. PeerJ Articles.

37. ScienceDirect Authors. (2020). An Android App Based Advisory Digital Tool for Menthol Mint
Farmers. Industrial Crops and Products.

57
38. IJRaT Authors. (2020). IoT Based Smart Farming using Decision Support System.

39. IARJSET Team. (2025). Smart Crop Monitoring Using IoT Sensors and Real-Time Image
Analysis for Plant Disease Detection. IARJSET Journal.

40. Nature Scientific Reports Team. (2024). A Technical Survey on Practical Applications and
Guidelines for IoT in Precision Viticulture. Scientific Reports.

58
 PLAGIARISM REPORT
Plagiarism Detection
Original Work
Originality: 93%
This paper appears to be original, but be certain that any included text is properly cited.
The following web pages may contain content matching this document:
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
[Link]
View Matching Text
Click the button above to see which portions of your text match which sources.
More info on our originality scoring process.
Spelling
Spelling Suggestions
No spelling errors found.
Grammar
Grammar Suggestions
Error Suggestion

To
Process
process

59
 machine- machine
learning learning

To
Integrate
integrate

drone- drone
imaging, imaging,
Word Choice
Usage of Bad Phrases
Bad Phrase Score: 0.45 (lower is better)
The Bad Phrase Score is based on the quality and quantity of trite or inappropriate words, phrases,
egregious misspellings, and cliches found in your paper. You did equal or better than 99% of the
people in your education level.

Great work - your score is well above average! Continue choosing quality phrases in your
writing.
Style
Usage of Transitional Phrases
Transitional Words Score: 42
This score is based on quality of transitional phrases used within your paper. You did equal or
better than 18% of the people in your education level.

Your usage of transitional phrases is below average. Please review the writing tips below.
One sign of an excellent writer is the use of transitional phrases (e.g. therefore, consequently,
furthermore). Transitional words and phrases contribute to the cohesiveness of a text and allow
the sentences to flow smoothly. Without transitional phrases, a text will often seem disorganized
and will most likely be difficult to understand. When these special words are used, they provide
organization within a text and lead to greater understanding and enjoyment on the part of the
reader.

The following transitional phrases were found in your document:

60
specifically, and, especially, however, therefore, rather, moreover, as a result

Consider using additional transitions where appropriate:

• consequently
• nevertheless
• notwithstanding
• accordingly
• conversely
• ordinarily

Transitional phrases may be used in various places in a text:

• betweenparagraphs
• between sentences
• between sentence parts
• within sentence parts

Consider this example:

She is one of the most kind persons I have ever known. Moreover, she is generous, patient and
possesses a magnificent sense of humor.

The word 'moreover' contributes to greater unity or cohesion between sentences and allows the
text to flow more smoothly.
Style
Sentence Length Info
Total Sentences: 129
Avg. Length: 15.1 words
Short Sentences (< 17 words): 73 (57%)
Long Sentences (> 35 words): 6 (5%)
Sentence Variation: 9.8 words (std deviation)
Your sentence length is below average. You may need to focus on writing more complex
sentences.
Line chart of the length of each sentence (first 50 sentences). A jagged chart indicates variation.
Helpful Resources:
 Effective Use of Sentence Length
Style
Readability Indices
For more than 80 years, great minds have attempted to represent a text with a single score that
summarizes its complexity. Writers sometimes use such 'readability indices' to gauge whether or

61
 not their text is appropriate for an audience of a particular grade level. Several of the most popular
readability scores are included below. More info on interpreting these scores.
Index Score
Automated Readability Index 12.2
Coleman-Liau Index 16.6
Dale-Chall Readability Formula 10.8
Flesch Reading Ease 27.3
Flesch-Kincaid Grade Level 12.2
Fry Readability Formula 2.0
Gunning-Fog Index 15.3
LIX 53.0
Linsear Write 7.6
Raygor Estimate Graph long_words
SMOG (Simple Measure of Gobbledygook) 13.1
Spache Readability Formula 85.9
Style
Passive Voice
Passive Voice Sentences: 3.9%
Many writers feel that passive voice represents poor writing form, as it allows the object of an
action to be the subject of a sentence. The following sentences were detected as having passive
voice:
 ... in full or in parts, have not been submitted to any other Institution or ...
 ... due acknowledgements have been made wherever the findings of others ...
 These data streams are processed using machine-learning models that ...
 ... experience?but none of these inputs are synchronized.
 When irrigation will be required next This supports farmers in ...

Helpful Resources:
 Active vs Passive Voice
Style
Sentence Beginnings
Simple Sentence Starts: 18%
You did equal or better than 63% of the people in your education level.

62
 Well done. Your writing is above average in this respect.
Variety is the hallmark of a good writer and this is especially true in regards to sentence starts.
Creatively arranging sentence beginnings breaks up the monotony and choppy style associated
with a simple noun phrase followed by a verb. This does NOT mean that all sentences should
begin with prepositional phrases, transitions, or adverbial phrases, but it does mean that you
should be certain to pay attention to sentence starts and deliberately edit for variety if necessary.
Here are some simple sentence starts that we found in your text:
 We declare that...
 It gives us...
 We owe special...
 It is only...
 Message Queuing Telemetry...
 Remote Sensing Applications...
 Studies reviewed in...
 MethSense fills this...
 India has established...
 Studies have consistently...
 Research indicates that...
 Traditional farming practices rely on...
 Mentha arvensis is sensitive...
 Environmental unpredictability further complicates menthol...
 Central Uttar Pradesh has recently...
 Studies show that...
 Technological solutions exist but...
 IoT-based systems monitor soil...
 Remote sensing helps...
 Specific objectives include 1...
 Profit forecasting helps farmers...
 Bilingual advisories increase accessibility...
 Precision irrigation supports groundwater...
Starts by Part of Speech:
Adjective: 5%
Adverb: 6%
Article: 0%
Conjunction: 0%
Noun: 32%
Preposition: 7%
Pronoun: 5%

63
Verb: 2%
Other: 30%

Helpful Resources:
 Sentence Beginnings - your sentences' first impression
Vocabulary Words
Usage of Academic Vocabulary
Vocabulary Score: 557.92
This score is based on the quantity and quality of scholarly vocab words found in the text. You
did equal or better than 97% of the people in your education level.

Vocabulary Word Count: 222

Percentage of Vocab Words: 14.07%
Vocab Words in this Paper (top 20):
cognizant, nomenclature, predictive, framework, synthesis, consolidated, integrates, multimodal,
frameworks, localized, socio, smallholder, aligns, variability, fragmentation, comprehensive,
integrating, aligning, smallholders, monitoring

Outstanding job! You really know your vocab. Take a look at our Vocab Builder if you want
some extra practice.

Grade
Auto Grader
Grade: 95 A

The grade above is NOT complete! We do not actually use a crystal ball to generate your grade.
Instead, this grade takes into account spelling, grammar, word choice, style, vocabulary, and
more; but it does NOT examine the meaning of your words, how your ideas are structured, or
how well your arguments are supported. We should also mention that our automated grader
doesn't always get things right. So, please consider this grade to be one facet of your paper's
overall grade.

64