AnnaMrita: An IoT and AI-Based Solution for Reducing Food Waste
and Ensuring Food Safety in India
Dr.Vikas Kaul1, Tejas S. Yadav2, Vinayak S. Vishwakarma3,Mohammed Kaif Tungekar4,
Ankit L. Yadav5
HoD, Dept of Information Technology,Shree LR Tiwari College of Engineering (SLRTCE), Mumbai, India1
Student, Dept. of Information Technology, Shree LR Tiwari College of Engineering (SLRTCE), Mumbai, India2
Student, Dept. of Information Technology, Shree LR Tiwari College of Engineering (SLRTCE), Mumbai, India3
Student, Dept. of Information Technology, Shree LR Tiwari College of Engineering (SLRTCE), Mumbai, India4
Student, Dept. of Information Technology, Shree LR Tiwari College of Engineering (SLRTCE), Mumbai, India5
Abstract:
Food waste is a pressing global issue contributing to environmental degradation and food insecurity,
particularly in India. This paper presents AnnaMrita, an innovative IoT and AI-based system designed to
minimize food waste and ensure food safety through real-time spoilage detection and optimized food
redistribution. The system utilizes an MQ-135 sensor to monitor volatile organic compounds (VOCs) and
moisture levels, enabling precise classification of food as “fresh” or “spoiled” using Support Vector
Machine (SVM) and Random Forest algorithms. A web-based platform facilitates coordination between
food donors (Informers) and NGOs (Responders), ensuring seamless and timely food distribution.
Additionally, an integrated route planner optimizes the delivery process by identifying the fastest and
most efficient transportation routes, providing estimated travel times for different modes of transport to
help NGOs reach the needy more effectively. Experimental results demonstrate high classification
accuracy (SVM: 92%, Random Forest: 95%), real-time monitoring capabilities, and effective alert
mechanisms, reducing spoilage and enhancing food donation efficiency. The proposed system is scalable,
cost-effective, and aligns with Sustainable Development Goal 12 (Responsible Consumption &
Production). Future enhancements include integrating blockchain for transparency, deep learning for
improved spoilage prediction, and a mobile application for better accessibility. AnnaMrita represents a
significant step toward building a sustainable, technology-driven food donation ecosystem.
Keywords: Food Waste, IoT, Sensor, AI, Machine Learning, Route Optimization, Food Safety,
Sustainable Development, Real-time Monitoring.
1. Introduction:
Food waste is a critical global challenge, with nearly one-third of all food produced annually going to
waste, leading to severe environmental, economic, and social consequences. In India, where food security
remains a persistent concern, millions face hunger despite the substantial amount of edible food discarded
daily. The inefficiencies in food donation systems, including food spoilage during storage and transit, lack
of real-time monitoring, and logistical challenges, further exacerbate the problem. A technology-driven
solution is essential to ensure timely food redistribution, enhanced food safety, and optimized delivery
mechanisms.
This paper introduces AnnaMrita, an IoT and AI-powered system that addresses food waste by detecting
spoilage in real time and facilitating efficient food donations. The system integrates IoT-based sensors
(MQ-135 and moisture sensors) to monitor volatile organic compounds (VOCs) and environmental
conditions affecting food freshness. Machine learning models, specifically Support Vector Machine
(SVM) and Random Forest, analyze sensor data to classify food as “fresh” or “spoiled”, ensuring that
only safe food is distributed.
A web-based platform enables coordination between food outlets (Informers) and NGOs/volunteers
(Responders), streamlining the donation process. To further optimize food distribution, AnnaMrita
incorporates an AI-powered route planner, which identifies the fastest and most efficient transportation
paths for NGO personnel, reducing delivery time and ensuring that food reaches its destination in optimal
condition.
Experimental results demonstrate high classification accuracy (SVM: 92%, Random Forest: 95%),
real-time spoilage alerts, and effective logistical coordination. Future enhancements include blockchain
integration for donation transparency, deep learning models for enhanced spoilage prediction, and a
mobile application for improved accessibility. By leveraging IoT, AI, and smart logistics, AnnaMrita
represents a scalable, cost-effective solution toward sustainable food donation and waste reduction,
aligning with Sustainable Development Goal 12 (Responsible Consumption and Production).
2. Problem Statement and Motivation:
2.1 Problem Statement
Food wastage remains a severe socio-economic and environmental challenge, with inefficiencies in food
donation and distribution systems contributing to substantial losses. In India, where over 194 million
people suffer from malnutrition, a significant portion of surplus food is discarded due to inadequate
storage conditions, lack of real-time monitoring, and inefficient logistical networks. Existing food
donation systems face several limitations:
● Food Safety Risks – No real-time monitoring mechanisms ensure the quality of donated food,
leading to concerns over spoilage and contamination.
● Logistical Inefficiencies – A lack of optimized transportation planning delays food deliveries,
increasing the likelihood of spoilage before reaching the beneficiaries.
● Data Gaps in Food Shelf Life – Current donation systems do not leverage predictive analytics to
assess food freshness, resulting in premature disposal of edible food.
2.2 Motivation
The increasing adoption of IoT and AI in smart agriculture and food supply chains presents an
opportunity to revolutionize food donation practices. AnnaMrita is designed to tackle food waste through
intelligent, real-time monitoring and efficient last-mile delivery optimization. The system aims to:
● Ensure Food Safety – Implement IoT sensors (MQ-135, moisture sensors) to monitor food
spoilage indicators, ensuring that only safe food is redistributed.
● Enable Proactive Decision-Making – Utilize AI-driven classification (SVM, Random Forest) to
predict spoilage trends and prevent waste.
● Optimize Logistics with Route Planning – Integrate an AI-based route optimization system to
help NGOs identify the fastest and most efficient transportation options, reducing transit time and
ensuring food reaches the needy before spoilage.
● Promote Sustainability – Align with Sustainable Development Goal (SDG) 12, emphasizing
responsible consumption, food waste reduction, and efficient resource allocation.
3. Related Work:
Numerous studies have explored IoT-based food monitoring, AI methodologies for spoilage detection,
and blockchain applications for transparency in food systems. However, the unique amalgamation of IoT,
AI/ML, and blockchain in AnnaMrita distinguishes it from existing solutions. This innovative platform
not only focuses on spoilage detection but also ensures secure and transparent food donation processes,
thus enhancing food safety and reducing waste more effectively than prior efforts.
Sr. Title Year Author(s) Technology Used Work Done Gap
No
01 Freshness of 2020 Nachiketa IoT: Node MCU, ✔ Integration of 1. Limited
Food Detection Hebbar oxygen sensor IoT Dataset
using IoT and Machine ✔ Integration of Dependency
Machine Learning: Machine 2. Less
Learning, 2020 Logistic Learning number of
International Regression, SVM, ✔ Real-Time sensors used
Conference on Linear Regression Monitoring
Emerging Trends
in Information
Technology and
Engineering
(ic-ETITE)
02 Spoilage 2022 Madhuri IoT: Arduino Uno, ✔ Integration of 1. Limited
detection and Borawake, Wi-Fi module IoT Parameters
shelf life Aradhana ESP8266, sensors ✔ 2. Machine
prediction of Sharma, (pH sensor, Implementation Learning
food using Ulfat oxygen sensor, of Machine Model
Internet of Shaikh, ammonia sensor, Learning model Training
Things and Sejal moisture sensor) with 98%
Machine Barkade, Machine accuracy
Learning, 2022 Putuja Learning: ✔ Real-Time
JETIR May Paturkar K-Nearest Monitoring
2022, Volume 9, Neighbors (KNN)
Issue 5 algorithm
03 Machine 2021 Sharif, A.; IoT: RFID Tag, ✔ RFID Tags: 1. Limited
Learning Abbasi, TagoPerformance Sticker-type Scope of
Enabled Food Q.H.; setup inkjet printed Contaminants
Contamination Arshad, K.; Machine RFID tags used. 2.
Detection Using Ansari, S.; Learning: These tags are Performance
RFID and Ali, M.Z.; XGBoost attached to food Metrics and
Internet of Kaur samples and Model
Things System, capable of Limitations
J. Sens. Actuator backscattering
Netw. signals when
interrogated by
an RFID reader.
✔ XGBoost
algorithm,
implemented in
Python, used for
training with 90%
accuracy
04 IoT and Machine 2022 Upendra IoT: Sensors and ✔ Utilization of 1. Incomplete
Learning based Singh, Dr. RFID IoT sensors Data
Model for Food Lokendra Machine ✔ Real-Time Gathering
Safety and Singh Learning: Monitoring 2. Lack of
Quality in Songera Convolutional-LS ✔ Immutable
Handling a TM Implementation Data Record
Pandemic of LSTM
Situation, Algorithm
International
Journal of Food
and Nutritional
Sciences Journal
Volume 11, Issue
12, Dec. 2022
4. Methodology:
IoT Setup:
The AnnaMrita project employs a simplified yet effective IoT setup to monitor and predict the shelf life
of perishable food. This setup integrates an ESP32 microcontroller with select sensors that provide key
data to detect spoilage trends, with real-time monitoring accessible via a personal website. Details of the
setup include:
Hardware Components
This circuit is designed to monitor environmental parameters including air quality, soil moisture, and
temperature. It utilizes a set of sensors interfaced with an ESP32 microcontroller, which is responsible for
reading sensor data and potentially sending it to a server. The sensors include the MQ-2 and MQ-135 for
air quality measurement, the SparkFun Soil Moisture Sensor for detecting soil moisture levels, and the
DS18B20 1-Wire Temperature Sensor for temperature readings.
Component List
1. MQ-2 Sensor
● Description: A gas sensor used for detecting LPG, i-butane, propane, methane, alcohol,
hydrogen, and smoke.
● Pins: VCC, GND, A0 (Analog Output), D0 (Digital Output)
2. MQ-135 Sensor Air Quality
● Description: A sensor for monitoring air quality, detecting a wide range of gases,
including NH3, NOx, alcohol, benzene, smoke, and CO2.
● Pins: VCC, GND, A0 (Analog Output), D0 (Digital Output)
● MQ-135 Sensor: This sensor detects volatile organic compounds (VOCs) such as
ammonia, which increase as food decomposes. By capturing VOC levels, the MQ-135
provides critical indicators of spoilage.
3. SparkFun Soil Moisture Sensor
● Description: A sensor for measuring the volumetric content of water in soil.
● Pins: VCC, GND, SIG (Signal Output)
● Moisture Sensor: Measures the moisture level around the food items. High moisture often
accelerates spoilage, especially in perishable foods. Monitoring moisture levels helps
provide context for VOC readings.
4. ESP32 (30 pin)
● Description: A powerful microcontroller with Wi-Fi and Bluetooth capabilities, suitable
for a wide range of applications.
● Pins: EN, VP, VN, D34, D35, D32, D33, D25, D26, D27, D14, D12, D13, GND, Vin,
D23, D22, TX0, RX0, D21, D19, D18, D5, TX2, RX2, D4, D2, D15, 3V3
● ESP32 Microcontroller: The core of the system, ESP32 receives data from connected
sensors, processes it, and transmits it to a locally hosted XAMPP server. The ESP32’s
wireless capabilities facilitate seamless data communication for real-time monitoring.
5. DS18B20 1-Wire Temperature Sensor Probe Cable
● Description: A digital temperature sensor that provides temperature readings over a
1-Wire interface.
● Pins: Shield, GND, DQ (Data), VDD (Power Supply)
● DSB1820 Temperature Sensor: Measures temperature and humidity, environmental
conditions that significantly affect food freshness. Maintaining a record of these metrics
helps model food spoilage patterns.
Wiring Details
1. MQ-2 Sensor
● VCC: Connected to 3V3 on ESP32
● GND: Connected to GND on ESP32
● A0: Connected to GPIO 34 (D34) on ESP32
2. MQ-135 Sensor Air Quality
● VCC: Connected to 3V3 on ESP32
● GND: Connected to GND on ESP32
● A0: Connected to GPIO 32 (D32) on ESP32
3. SparkFun Soil Moisture Sensor
● VCC: Connected to 3V3 on ESP32
● GND: Connected to GND on ESP32
● SIG: Connected to GPIO 35 (D35) on ESP32
4. ESP32 (30 pin)
● 3V3: Power supply for sensors
● GND: Common ground for all components
● D35: Connected to SIG on SparkFun Soil Moisture Sensor
● D4: Connected to DQ on DS18B20 Temperature Sensor
● D32: Connected to A0 on MQ-135 Sensor Air Quality
● D34: Connected to A0 on MQ-2 Sensor
5. DS18B20 Wire Temperature Sensor Probe Cable
○ VDD: Connected to 3V3 on ESP32
○ GND: Connected to GND on ESP32
○ DQ: Connected to GPIO 4 (D4) on ESP32
Data Transmission and Processing:
○ The ESP32 microcontroller, programmed using Arduino IDE, collects sensor readings at
specified intervals. These readings are sent to a local XAMPP server, where they are
temporarily stored for easy access and further analysis.
○ PHP scripts on the XAMPP server manage the backend, storing data in a MySQL
database for consistency and reliability.
● Visualization and Monitoring
○ Sensor data is displayed on a custom website, enabling remote monitoring. Graphical
visualizations aid in understanding food quality trends, while threshold-based alerts
notify users if food conditions suggest spoilage.
○ The thresholds for "fresh" and "spoiled" states are pre-defined based on initial sensor data
analysis from fresh and decayed samples, enhancing detection accuracy.
AI/ML Models:
1. Support Vector Machine (SVM)
○ Purpose: SVM is employed to classify food as “fresh” or “spoiled” based on VOC and
moisture levels. Given that spoilage is influenced by precise environmental thresholds,
SVM’s strength in binary classification makes it ideal for separating fresh and spoiled
states with high accuracy.
○ Advantages in AnnaMrita: SVM works well with high-dimensional data, allowing it to
effectively use features derived from sensor readings to delineate the freshness boundary.
The model identifies the hyperplane that best separates fresh and spoiled food based on
historical data, providing a reliable decision boundary even with limited data samples.
○ Application in Prediction: By training SVM on data labeled as fresh or spoiled, the
model learns to detect subtle shifts in VOC and moisture levels. This capability enables
the system to alert users of spoilage onset at an early stage.
2. Random Forest
○ Purpose: Random Forest is utilized for feature importance and to increase prediction
robustness by analyzing decision trees built on sensor data patterns. It serves as an
ensemble method that evaluates multiple decision pathways, increasing model stability
and accuracy.
○ Advantages in AnnaMrita: Given that food spoilage can vary with multiple interacting
factors, Random Forest’s ensemble of decision trees helps account for complex
relationships between VOC levels, moisture, and time. The model identifies critical
thresholds and feature importance for each sensor reading, highlighting which indicators
are most predictive of spoilage.
○ Application in Prediction: The model’s collective decision approach enhances
generalizability, reducing the risk of overfitting and making it versatile across various
food types. By implementing Random Forest, AnnaMrita can adapt its spoilage
predictions to account for variability in food properties and storage conditions.
➢ Why SVM and Random Forest?
The combination of SVM and Random Forest in AnnaMrita allows for both precise classification
(fresh vs. spoiled) and adaptive learning of spoilage indicators, maximizing model accuracy.
SVM’s robustness in classification complements Random Forest’s strength in capturing complex,
multi-feature relationships, offering a comprehensive approach to food spoilage prediction. This
hybrid approach enables AnnaMrita to provide consistent, reliable alerts for food quality, aiding
in timely interventions and reducing food waste effectively.
Web Application:
The AnnaMrita web application is a versatile platform that not only monitors food quality but also
facilitates timely food donations, making it an essential tool in reducing food waste. The application
serves two primary user roles: Informers (food outlets) and Responders (NGOs and volunteers). Together,
these roles ensure the seamless flow of information on food availability and quality, allowing for efficient
food redistribution.
1. User Interface Design
○ Dashboard View: The main dashboard provides Informers and Responders with
real-time sensor data, displaying the VOC levels from the MQ-135 sensor and moisture
readings, which indicate food freshness.
2. User Roles and Functionalities
○ Informers (Food Outlets): These users, primarily food donors, can log in to provide
details about available food and monitor spoilage status in real time. The application
helps food outlets manage donations by notifying them if food conditions are nearing
spoilage, promoting timely distribution.
○ Responders (NGOs and Volunteers): Responders receive alerts regarding food
availability and can view data on food freshness, allowing them to prioritize collection
efforts. This role’s main functionality is centered around receiving notifications and
acting on donation opportunities, creating an efficient response mechanism.
AI-Powered Route Optimization:
● The system integrates a route planner that calculates the fastest and most efficient path
for food pickup and delivery.
● It provides estimated travel times for different transportation modes (bike, car, public
transport) to help NGOs optimize their collection efforts.
● This feature reduces transit delays, ensuring food reaches its destination in a safe and
timely manner.
3. Backend and Data Management
○ Server and Database: A XAMPP server hosts the backend, where sensor data is stored
in a MySQL database managed by PHP scripts. Data from the ESP32 microcontroller is
stored and classified using machine learning models for spoilage prediction.
○ Integration of Machine Learning: SVM and Random Forest models process sensor data
to classify it as “fresh” or “spoiled,” based on learned patterns. Classification results are
saved in the database and visualized on the dashboard, aiding users in making timely
decisions.
○ Data History: The application stores historical data, allowing both Informers and
Responders to analyze spoilage trends. This feature also supports machine learning model
retraining, further improving prediction accuracy over time
5. Experimental Setup and Results:
The experimental phase of AnnaMrita aimed to validate the effectiveness of the IoT-based system in
monitoring food quality and accurately predicting spoilage. This involved setting up sensor
configurations, calibrating machine learning models, and conducting real-time tests on various food items
to observe spoilage patterns. The collected data provided insights into the system’s reliability, enabling
refinements and improvements for practical applications.
1. Experimental Setup
● Sensor Configuration: The ESP32 microcontroller was connected to two key sensors: the
MQ-135 sensor for monitoring VOC levels, particularly ammonia and other gases emitted during
food decomposition, and a moisture sensor to assess environmental moisture levels around the
food. Each sensor was calibrated based on data from fresh and spoiled food samples to establish
baseline and threshold values for accurate classification.
● Data Collection: The setup included an isolated testing environment where common perishable
items, such as dairy products, fruits, and grains, were placed. Sensor data was recorded at regular
intervals over several days, capturing VOC and moisture changes as spoilage progressed. This
data was then sent to a local XAMPP server and stored in a MySQL database for processing and
analysis.
● Machine Learning Model Calibration: The SVM and Random Forest models were trained on a
dataset comprising labeled instances of fresh and spoiled states based on sensor readings. The
models were fine-tuned to classify data from the IoT sensors accurately. SVM, chosen for its
binary classification capability, was particularly effective in drawing a boundary between fresh
and spoiled states, while Random Forest added robustness by analyzing feature importance,
making the system adaptable to variations in food spoilage behavior.
2. Results
The experimental setup yielded promising results, validating AnnaMrita’s ability to reliably detect food
spoilage and alert users in a timely manner. Key findings from the testing phase are summarized below:
● Accuracy of Classification:
○ SVM Model: Achieved a classification accuracy of approximately 92% in distinguishing
fresh from spoiled states. The SVM model performed consistently in identifying spoilage
based on VOC levels alone, which correlated well with early stages of decomposition.
○ Random Forest Model: Demonstrated an accuracy of 95%, benefiting from its ensemble
approach and the inclusion of moisture data. By ranking feature importance, Random
Forest enabled the system to adapt classification criteria based on the food type and
environmental conditions, enhancing the model’s overall accuracy.
● Response Time: The setup effectively monitored sensor readings in real time, updating the web
dashboard with a delay of under 5 seconds. This rapid response allowed users to view real-time
data and receive alerts promptly, critical for food that is nearing spoilage.
● Effectiveness of Alerts: Alerts generated by the system were tested with Responders (NGOs and
volunteers) under simulated conditions. The system successfully sent notifications when readings
neared spoilage thresholds, facilitating timely response by alerting NGOs about food availability
before spoilage occurred.
6. Use Cases:
1. Reducing Excess Food Waste
AnnaMrita enables food outlets, restaurants, and organizations to proactively manage and donate
excess food items before they reach spoilage. By continuously monitoring food quality through
VOC and moisture levels, the system alerts food handlers when food is nearing spoilage, allowing
them to initiate donation processes. This proactive approach minimizes waste and promotes
efficient use of excess food resources.
2. Food Quality Assurance
Ensuring food safety is crucial throughout the donation cycle, from the moment food is offered
for donation to the point it reaches beneficiaries. AnnaMrita supports food quality assurance by
continuously monitoring and classifying food status as “fresh” or “spoiled.” This feature ensures
that only safe, high-quality food is distributed, protecting recipients from potential health hazards
and building trust in food donation efforts.
3. Data-Driven Decision Making
By utilizing real-time sensor data, AnnaMrita provides actionable insights for food donation
organizations and handlers. The system’s machine learning models predict spoilage patterns
based on collected data, helping users make informed decisions about food handling, storage, and
distribution timing. This data-driven approach optimizes the entire donation process by
prioritizing food that requires immediate attention, ensuring timely action.
4. Real-time Food Safety Monitoring
AnnaMrita empowers NGOs and volunteers with instant access to the status of donated food.
Through the web application, Responders (NGOs and volunteers) can view real-time VOC and
moisture readings and receive immediate alerts if food approaches spoilage thresholds. This
functionality is crucial for maintaining food safety, especially in situations where timely
redistribution is essential.
5. Smart Food Donation and Redistribution
The system streamlines food donation processes by connecting Informers (food outlets) with
Responders (NGOs and volunteers) through a centralized web application. Informers are
promptly notified of food nearing spoilage, and Responders receive alerts regarding food
availability, enabling them to respond efficiently. This seamless communication improves
engagement between food outlets and NGOs, making food redistribution faster and more reliable,
and ultimately maximizing the impact of donated food.
7. Feasibility and Viability Analysis:
An operational and economic feasibility study indicates that AnnaMrita can be implemented effectively
using cost-effective IoT components and open-source AI tools. Addressing operational challenges, such
as technology integration and user training, will be essential. Strategies for overcoming these hurdles
include continuous system testing, user education programs, and diversification of funding sources.
8. Potential Challenges and Risks:
● Technological Integration: Ensuring reliable data transmission and AI prediction accuracy.
● Operational Resistance: Educating stakeholders about the effective use of the technology.
● Funding and Scaling: Acquiring necessary resources to expand the initiative.
9. Conclusion:
The AnnaMrita project presents a transformative approach to addressing the critical issue of food waste
while enhancing food safety and security. By integrating IoT technology with advanced machine learning
algorithms, AnnaMrita presents a novel IoT and AI-driven approach to minimizing food waste by
integrating real-time spoilage detection, AI-powered classification, and smart logistics optimization for
food donations. The project has demonstrated high accuracy in spoilage detection, facilitating timely
alerts that empower NGOs and volunteers to respond efficiently. Through its innovative web application,
AnnaMrita not only streamlines the food donation process but also fosters collaboration between food
outlets and charitable organizations along with route planner which significantly improves food donation
logistics, ensuring that surplus food reaches the needy before spoilage occurs. While challenges remain in
terms of scalability, user adoption, the potential impact of AnnaMrita on reducing food waste and
supporting vulnerable communities is significant. Future enhancements, including the integration of
additional sensor technologies and mobile accessibility, promise to further strengthen the system's
effectiveness. Ultimately, AnnaMrita serves as a crucial step toward building a more sustainable food
ecosystem, promoting responsible consumption, and contributing to global efforts in combating food
insecurity.
10. Future Work :
Looking ahead, the AnnaMrita project aims to enhance its capabilities and expand its impact through
several key initiatives. One primary focus will be on improving sensor accuracy and reliability by
exploring advanced sensor technologies and integrating additional sensors to monitor temperature and
humidity levels, further refining spoilage predictions. Additionally, expanding the machine learning
models to incorporate more diverse datasets will enhance the system's predictive analytics, allowing for
more accurate classifications across various food types. The project will also explore mobile application
development to improve accessibility for users in the field, ensuring that both Informers and Responders
can access real-time data on their smartphones. Collaboration with local NGOs and food outlets will be
pursued to facilitate broader adoption of the system and gather user feedback for continuous
improvement. Finally, addressing scalability challenges is essential; thus, AnnaMrita will work on
optimizing backend processes to manage larger datasets and a growing user base efficiently. Through
these efforts, AnnaMrita seeks to not only minimize food waste but also foster a more sustainable and
responsive food donation ecosystem.
11. References :
[1] Nachiketa Hebbar, "Freshness of Food Detection using IoT and Machine Learning" International
Conference on Emerging Trends in Information Technology Engineering IC- ETITE), 2020.
[2] Madhuri Borawake, Aradhana Sharma, Ulfat Shaikh Sejal Barkade, Rutuja Paturkar, "Spoilage
detection and shelf life prediction of food using Internet of Things and Machine Learning",JETIR Volume
9. Issue 5,2022.
[3] Sharif, A.,Abbas,QHArshad, KAnsari, S.,Ali, M.Z., Kaur,"Machine Learning Enabled Contamination
Food Detection Using RFID and Internet of Things System", J. Sens. Actuator Netw. ,2021.
[4] Upendra Singh, Dr. Lokendra Singh Songarea, "IoT and Machine Learning based Model for Food
Safety and Quality in Handling a Pandemic Situation", International Journal of Food and Nutritional
Sciences Journal Volume 11,Iss 12, Dec 2022.
