Smart Fall Detection System

for

Field Project
of
Second Year (Semester-IV)
Bachelors in Engineering
by
Varunraj Bhirud D9A (06)
Chirag Pedamkar D9A (36)
Aadit Pradhan D9A (40)
Aditya Rathi D9A (43)

Supervisor
Mrs. Mugdha Jogalekar
Designation: Assistant professor

Department of Electronics and Telecommunication Engineering

Vivekanand Education Society's Institute of Technology
An Autonomous Institute Affiliated to University of Mumbai
AY-2024-2025

3
 Field Project Approval
Project titled by Smart Fall Detection System by Varunraj Bhirud, Chirag
Pedamkar, Aadit Pradhan, Aditya Rathi is approved for the Second year of
Engineering.

Examiners

1.

2.

Supervisors

1.

Date:
Place:

4
 Certificate

This is to certify that Varunraj Bhirud, Chirag Pedamkar, Aadit Pradhan, Aditya
Rathi have completed the project report on the topic satisfactorily in partial
fulfillment of the requirements for Field Project of Second Year, (Semester-IV) in
Electronics and Telecommunication under the guidance of Mrs. Mugdha Jogalekar
during the year 2024-2025.

Supervisor Head of Department

Name Dr. Chandansingh Rawat

Principal
Dr. (Mrs.) Jayalekshmi Nair

Examiner 1 Examiner 2

5
 Declaration

We declare that this written submission represents our ideas in our words and where
others’ ideas or words have been included, we have adequately cited and referenced
the original sources. We also declare that we have adhered to all principles of
academic honesty and integrity and have not misinterpreted or fabricated or falsified
any idea/data/fact/source in my submission. We understand that any violation of the
above will be cause for disciplinary action by the Institute and can also evoke penal
action from the sources which have thus not been properly cited or from whom
proper permission has not been taken when needed.

Varunraj Bhirud
Chirag Pedamkar
Aadit Pradhan
Aditya Rathi

Date:

6
 Table of Contents

Chapter No. Title Page no.

Abstract 1
List of Figures 2

Chapter 1: Introduction 3
1.1 Field Visit 3
1.2 Problem Identification 4

Chapter 2: Literature Review 5

2.1 Survey 5
2.2 Research Gap 7

Chapter 3: Project Description 8

3.1 Problem Definition 8
3.2 Steps Involved 8
3.3 Block Diagram 9
3.4 Component Description/ Hardware & Software 10
3.4.1
3.5 Working of the project 11
3.5.1

Chapter 4: Implementation 12
4.1 Hardware 12
4.2 Software (Flowchart/ Algorithms) 14

Chapter 5: Results and conclusion 15

5.1 Results 15
5.2 Conclusion 17
5.3 Future Scope 18

References 19

6
 Abstract
Falls are among the most common and hazardous risks faced by elderly and disabled
individuals, often resulting in severe injuries, hospitalization, or the need for immediate
medical intervention. Recognizing the urgency of timely response, this project presents
wearable fall detection system designed to ensure real-time alerting for caregivers or
family members. The system is built using an ESP32 microcontroller integrated with
an MPU6050 accelerometer and gyroscope sensor, which continuously monitors the
user's motion. By detecting a pattern of sudden acceleration followed by a period of
inactivity, the system can accurately identify potential incidents without relying on
rigid threshold values or complex machine learning models. This pattern-based
approach enhances reliability while maintaining simplicity and efficiency.

Once a fall is detected, the ESP32 utilizes Wi-Fi connectivity to automatically send an
alert via email, enabling quick assistance even when the individual is alone. The system
emphasizes affordability, compactness, and ease of use, making it ideal for elderly users
in home settings or residents of assisted living facilities. Unlike commercially available
solutions that may be expensive or overly complicated, this project prioritizes
accessibility and practicality. Moreover, the decision to use lightweight components
and wireless communication ensures the device is non-intrusive and comfortable to
wear for extended periods. This project demonstrates how embedded systems and basic
IoT principles can be leveraged to improve quality of life and safety for vulnerable
population.

1
List of figures: -

Figure number Content Page number

1 Field Visit (Clinic) 3
2 Field Visit (Doctor) 3
3 Block Diagram 9
4 Pin diagram of ESP32 10
5 Pin diagram of MPU6050 10
6 Circuit Diagram 12
7 Outer View 13
8 Inner View 13
9 Mounted device 13
10 Mounted device (Full Body) 13
11 Flow Chart 14
12 Graph for Walking 15
13 Graph for Sitting 16
14 Graph for Disturbance 16
15 Graph for Fall 17
16 Email Alert 17

2
 Chapter 1
Introduction

1.1 Field Visit: -

As part of our field project work, our group conducted a field visit to gain real-world
insights into the needs and challenges faced by our target audience, primarily elderly
individuals living alone, disabled people, and residents of nursing homes, hospitals, or
rehabilitation centers. We visited a local clinic operated by Dr. Hemant Tadwalkar, a
well-respected adult vaccinator and fitness expert. The visit aimed to help us understand
the real-life scenarios where a fall detection system would be most beneficial and to
receive professional advice from someone experienced in dealing with our target
demographic on a regular basis.

During our interaction, Dr. Tadwalkar provided us with valuable information based on
his extensive experience in preventive healthcare and elderly fitness. He shared several
insights into the challenges faced by the current fall detection systems, especially in
terms of their reliability, affordability, and ease of use. One of the key issues he
highlighted was that many existing systems are either too expensive for widespread
adoption or too complicated for elderly users to operate comfortably. He also
emphasized the importance of designing a system that is a lightweight wearable and
capable of functioning with minimal user interaction.

Furthermore, Dr. Tadwalkar emphasized the need for accurate detection methods that
minimize false alarms while ensuring that real falls are never missed. According to him,
the ability of a system to promptly alert caregivers or family members is crucial in
improving the chances of providing timely medical assistance. This industrial visit was
very helpful and informative and helped us refine our hardware selection, select
communication method and select the fall detection method as well

Figure 1: Field Visit (Clinic) Figure 2: Field Visit (With Doctor )

3
1.2 Problem identification

Falls are one of the most common and dangerous risks faced by elderly individuals,
people with disabilities, and patients recovering from surgery. Through careful
observation, user feedback, and research into real-world scenarios, we recognized a
recurring problem: many fall incidents go unnoticed, especially when the individual is
alone. Although various fall detection devices already exist, such as smartwatches,
pendants, and other wearables, these systems often fall short in several key areas. Many
are priced beyond the reach of the average user, making them inaccessible to those who
need them most. Others require regular charging and maintenance, which can be
challenging for elderly users. Some devices only work within a limited range or depend
heavily on mobile apps and internet connectivity, reducing their reliability in remote or
low-network areas. Moreover, many of these products either raise frequent false alarms
due to inaccurate motion sensing or fail to detect real falls, which could lead to serious
consequences.

In response to these challenges, we developed a more practical and efficient fall

detection system using the ESP32 microcontroller and the MPU6050 sensor. This
combination provides precise tracking of body movement and orientation without
relying on expensive or overly complex technologies. The system processes motion
data in real time and, when a fall is detected, immediately sends an email alert to
designated contacts using the SMTP protocol. Unlike many commercial products, our
solution is cost-effective, easy to set up, and doesn’t rely on mobile applications or
subscriptions. It runs independently, requiring minimal interaction from the user,
making it especially suitable for elderly individuals, patients in home care, and workers
in high-risk environments. By focusing on affordability, reliability, and simplicity, our
system aims to provide a more inclusive and dependable safety net for those most at
risk.

4
 Chapter 2
Review of Literature

2.1 Survey

2.1.1 A Real-Time IoT Based Fall Detection and Alert System for Elderly

In this paper, the authors present a wearable system designed for real-time fall
detection and alerting, specifically targeting elderly individuals. The system uses
the MPU6050 accelerometer and gyroscope module to capture motion data,
which is processed using a microcontroller. An ESP8266 Wi-Fi module is
utilized to send alerts to caretakers through both email and SMS using IoT-based
communication. Although the detection mechanism in this system is based on
threshold values of acceleration and orientation, the implementation of real-time
notifications through the internet aligns closely with the proposed system in this
project. This paper serves as a valuable reference for understanding how low-
cost hardware and basic communication protocols can be integrated to create an
efficient alert mechanism.

2.1.2 Smart Wearable Socks Enabled with IoT for Fall Detection

This work introduces a novel approach to fall detection by integrating sensors

into wearable socks. The system includes pressure and motion sensors to monitor
the wearer’s posture and movement. The authors designed a cloud-based IoT
model which collects and processes the data, and sends alerts in case any
abnormal motion or possible falls are detected. The focus of the study is on
increasing the comfort and usability of the fall detection devices by embedding
the sensors in daily-wear items. While this system does not rely on a jerk-based
approach, it offers valuable insights into non-intrusive wearable designs and
emphasizes the importance of user convenience in assistive technologies. The
paper highlights how sensor data can be transmitted and managed over the cloud,
which is conceptually relevant to this project's email alert system.

5
2.1.3 Using Smart Devices for Fall Detection: Algorithms, Systems and
Applications

This paper is a comprehensive survey of various fall detection algorithms and

systems that have been developed using smart devices. The authors examine
both, the traditional threshold-based approaches and the more recent methods
which involve machine learning and sensor fusion. One important aspect that is
discussed in this study is the use of "jerk", which is defined as the rate of change
of acceleration, as a key metric for fall detection. This supports the approach
taken in the proposed project, which focuses on detecting sudden changes in
acceleration rather than relying solely on static thresholds. The survey provides
insights into common challenges such as false positives and battery efficiency,
which are essential considerations when developing a wearable system. This
paper helps in understanding the broader context of fall detection research and
justifies the need for alternative techniques like jerk-based detection.

2.1.4 A Real-Time Fall Detection System Using ToF Depth Images

Unlike wearable sensor-based approaches, this study explores fall detection using
Time-of-Flight (ToF) depth cameras. The system processes real-time depth
image data to analyze human posture and detect falls based on body orientation
and motion trajectories. Although this technique differs fundamentally from the
wearable system proposed in this project, it introduces another dimension to fall
detection: vision-based analysis. This paper is significant as it presents an
alternative to traditional sensors and highlights trade-offs between accuracy, user
privacy, and the complexity of data processing. The findings reinforce the value
of wearable devices, especially when cost, portability, and privacy are key
concerns.

2.1.5 A Smart Wearable Device for Tracking and Fall Detection for Geriatric
People, Alzheimer Patients and Children with Autism

This paper presents a multi-purpose wearable device that monitors individuals

who are at high risk of falling, including elderly people, Alzheimer patients, and
children with autism. The system uses accelerometers to detect falls and
incorporates wireless communication modules to notify caregivers. The design
emphasizes the importance of reliability and simplicity in wearable health
monitoring devices. The authors highlight the usefulness of real-time alerts and
continuous tracking, which is directly relevant to this project. The study also
confirms the feasibility of using microcontrollers such as the ESP32 to
implement compact and cost-effective monitoring systems. Moreover, the study
supports the idea of designing inclusive fall detection solutions that cater to a
wide range of users with different needs.
6
2.2 Research gaps

Despite the advancements in the fall detection systems over recent years, several critical
research gaps still exist, particularly in the area of wearable implementations. Most
existing systems rely on static threshold-based detection methods which utilize fixed
acceleration or orientation thresholds to determine falls. Although, it is a straightforward
approach that is easy to implement, it often results in false positives during normal
activities such as sitting or bending, and it can overlook actual falls that may occur more
gradually.

The rate of change of acceleration, also known as "jerk," is considered to be a more

accurate indicator of sudden motion. But there remains a lack of exploration into its
implementation on low-power microcontroller platforms like the ESP32. Research
efforts have primarily concentrated on smartphones or sophisticated laboratory
equipment, leaving a notable gap in real-world embedded applications. Another
significant limitation is the absence of hybrid detection approaches. Most devices utilize
either acceleration or orientation independently, rather than using a combination of
parameters such as jerk, post-fall stillness, and orientation to enhance overall accuracy
and reliability.

Additionally, many existing systems depend heavily on third-party platforms like IFTTT
or cloud services for alert notifications, which can introduce latency and may become
unreliable in areas with poor network connectivity. There are few systems that implement
direct, protocol-based alert mechanisms, such as SMTP, which would allow for
independent operation and precise control. Another critical issue that is often overlooked
is personalization; many fall detection systems utilize hard coded thresholds that do not
account for individual user characteristics, such as body weight, gait, physical deformity,
etc.

The lack of publicly available, real-world fall datasets further limits the ability to
effectively validate and compare fall detection algorithms. There aren’t many datasets
which are based on jerk. Lastly, energy optimization is rarely addressed in detail despite
the expectation that wearable devices will function continuously. Constant processing of
data can quickly drain the battery life of the device and yet, strategies like adaptive
sampling are rarely implemented. This project aims to tackle these challenges through
the employment of jerk-based detection approaches and direct SMTP communication.

7
 Chapter 3

Project Description

3.1 Problem Definition

Falls in the elderly and disabled people with impaired mobility are a major threat
to health and wellness, and it is important that they be detected to minimize the
injuries caused by such incidents. This project intends to create an affordable, real
time Smart Fall Detection System by an ESP32 microcontroller using the
assistance of an MPU6050 accelerometer and gyroscope sensor. The system is
meant to track movement and orientation, identifying falls due to sudden
acceleration changes followed by a period of inactivity. When a fall is detected,
the system will send an alert notification via email using SMTP protocol. With the
use of low-cost materials, the project intends to offer a reliable solution that is
affordable and scalable, easy to implement, hence enhancing safety for the elderly
and disabled individuals with impaired mobility.

3.2 Steps involved

• Requirement Analysis: The project begins by defining the needs of the system,
which includes the fall detection mechanism and alerting methods.
• Component Selection: After identifying the system’s requirements, the next step
is to select the necessary components which are ESP32 microcontroller, MPU6050
sensor and power supply.
• Designing the circuit: A circuit diagram is created for the system which represents
the connections of the ESP32, MPU6050 and power supply.
• Programming the Microcontroller: The ESP32 is programmed to read data from
the MPU6050 sensor, process the information, and detect falls based on
acceleration and inactivity and then send alert notification via email.
• Motion and Fall Detection Algorithm: A fall detection logic is developed, which
analyses the changes in acceleration, followed by inactivity. When such a pattern
is detected, the system confirms the fall and sends alerts via email.
• Testing: The system goes through rigorous testing in various scenarios to ensure
that it accurately detects falls and sends alerts.
• Enclosure: The components are assembled into a compact and protective
enclosure, ensuring they are securely housed while remaining portable. The
enclosure ensures durability of the whole apparatus.

8
 3.3 Block Diagram of proposed Project

Figure 3: Block Diagram

The above figure is of the block diagram that represents the overall process of our system.
The sensor starts reading the data which is sent to ESP32 via I2C protocol. This raw data
is first filtered by using a moving average filter in order to minimize the noise
interference so that we can get more accurate input readings. The filtered data is then fed
to the fall detection logic which checks if a fall has occurred or not. If a fall has been
detected, then the email alert is sent to a pre-determined recipient after which the system
resumes its operation and if there is no fall, then the alert notification is not sent and the
system simply continues to monitor.

3.4 Component Description

3.4.1 Hardware

1. ESP32:-

ESP32 is a powerful, low-cost microcontroller. It is used in a wide range of

applications like IoT, robotics & automation, embedded systems and edge
computing, etc. Some features and specifications of Esp32 include the following:

• Processor: Dual-core Tensilica Xtensa processor running up to 240 MHz.

• RAM: 520kB (Static).
• Flash memory: 4MB.
• Wi-Fi: 802.11 b/g/n protocol. Enables to connect to Wi-Fi, but only of 2.4
GHz frequency band.
• GPIO: 38 pins
• Supports multiple communication protocols like I2C, UART, SPI, etc.

9
 Figure 4: Pin diagram of ESP32

2. MPU6050:-

MPU6050 is a sensor that is used to detect orientation, motion, rotation and

stability of the device it is attached to. It is used in many devices such as mobile
phones, gesture recognition system, VR headsets, gaming controllers, fitness
trackers, flight control systems, etc. The specifications of this sensor are as
follows:-

• Sensor type: IMU (Inertial Measurement Unit).

• Number of axes: 6-axis (3-axis accelerometer + 3-axis gyroscope).
• Communication interface: I2C, allows easy integration with ESP32.
• Power consumption: designed for low power consumption.

Figure 5: Pin diagram of MPU6050

10
3.4.2 Software

1. Arduino IDE (Integrated development environment): -

It is an open-source software that is used to write, compile and upload code of

microcontrollers to Arduino and other compatible boards. It has a user-friendly
interface and also includes some examples for common asks like blinking LED.
It is used for simplifying embedded programming and enabling hardware
interaction. Below are some features of this software: -

• Language used: Embedded C++

• Has a serial plotter function to plot graphs.
• Has a serial monitor to send and receive data from development boards.
• Contains thousands of libraries for sensors, modules, etc.
• Allows for cross-platform development.

3.5 Working of proposed project

1. When the system is switched on, the ESP32 connects to Wi-Fi using the
SSID and password credentials stored in it. Then the MPU6050 sensor is
initialized by the I2C protocol which allows the ESP32 to begin reading the
incoming acceleration and gyroscope data from the sensor.

2. The incoming data from the sensor is fed to the moving average filter which
minimizes the noise interference in order to produce a more accurate reading. This
filtered data is then fed to the fall detection logic in which two things are tested,
impact and inactivity. If there is a sudden and significant change in the
acceleration along Z-axis, an impact is said to be detected. If there is no impact,
then the ESP32 keeps reading the filtered input data that is sent by the sensor.

3. If an impact is detected, then the logic checks for inactivity for a set period of
time. If the time of inactivity exceeds the set time period, then inactivity is
confirmed and as a result a fall is said to have occurred. This step helps in ruling
out certain motions and actions that lead to false positives like jumping, running,
etc.

4. If a fall is detected, then an email alert is sent to a list of recipients that is pre-
programmed into the device. This email is sent via SMTP servers using SMTP
protocol.

5. The system needs to be reset manually by pushing the assigned button. Once it
has been reset, the ESP32 starts reading the input data from MPU6050 again, thus
repeating the entire process of data acquisition, filtration and fall detection.
11
 Chapter 4

Implementation

4.1 Hardware

The Smart Fall Detection System is designed around the ESP32 microcontroller, which
functions as the central processing unit for the setup. To detect potential falls, an
MPU6050 sensor module is integrated with the ESP32. This sensor communicates with
ESP32 board via I2C protocol, with the SDA (Serial Data) pin connected to GPIO21
and the SCL (Serial Clock) pin connected to GPIO22.

Power for all the components is supplied through the 3.3V and GND pins of the ESP32.
Additionally, a push button serves as a reset mechanism which when pressed, resets the
fall detection state and stops the alert signals.

One of the LEDs of ESP32 which acts as a visual indicator when a fall is detected. The
LED lights up and fades continuously until the reset button is pressed. The ESP32 is
connected to Wi-Fi using which the system sends email notifications in the event of a
fall detection.

Figure 6: Circuit diagram

12
 Figure 7: Outer view Figure 8: Inner view

Figure 9: Mounted Device Figure 10: Mounted Device

(Full Body)

13
4.2 Software (Flowchart/Algorithms)

The software was developed using C++ based programming language in Arduino IDE.
This program begins with the initialization of the MPU6050 sensor for communicating
with ESP32 using I2C protocol and Wi-Fi connectivity for sending email alerts using
SMTP protocol. Then the sensor reads the data and then feeds it to ESP32 where the data
goes through filtration by moving average filter in order to remove noise interference.
The filtered data is then fed to the fall detection logic which determines if a fall has
occurred or not and then send the notification accordingly and then continues monitoring.

Figure 11: Flow chart

14
 Chapter 5

Results and Conclusion

5.1 Results

The smart fall detection system was successfully implemented using ESP32 and MPU6050
sensor and it was also tested in various scenarios. The system detected falls by observing
sudden changes in the jerk and acceleration of the body, followed by a period of inactivity.
This system did not rely on strict threshold-based approach and utilized a pattern-based
approach along with it. Upon detecting a fall, ESP32 sent an email alert using SMTP
protocol which ensured timely notification. This system is a cost-effective solution as it
was made by using only 2 components, ESP32 and MPU6050 and no third party IoT or
cloud platform was utilized so, there was no expenditure on their services. The testing
results in various scenarios yielded expected results which proves that this system shows
potential for real-world applications as well.

Scenario 1: Walking

Figure 12: Graphs for walking

15
Scenario 2: Sitting

Figure 13: Graphs for Sitting

Scenario 3: Disturbance

Figure 14: Graphs for Disturbance

16
Scenario 4: Fall

Figure 15: Graphs for Fall

Mail:

Figure 16: Email alert

17
5.2 Conclusion

The Smart Fall Detection System presents an innovative and effective approach to monitor
and alert certain individuals about fall incidents. Utilizing the ESP32 and MPU6050 sensor,
a pattern-based detection method was adopted, which can recognize fall events without
depending on solely fixed thresholds or complicated machine learning algorithms. One of
the key features of the system is its capability to send real-time alerts via email using SMTP
protocol, which facilitates a prompt response during emergencies. In addition to that, the
inclusion of onboard hardware of a push button used as reset button contributes to the
overall simplicity and efficiency of the design. This project shows the potential of an
affordable and accessible technology to enhance safety particularly for vulnerable
populations like the elderly and disabled individuals with mobility challenges. Overall, the
Smart Fall Detection System exemplifies how technology can play a critical role in
improving the quality of life for those at risk of falls and how it can help our target audience.

5.3 Future Scope

• There are many things we can add to this project and make it more reliable and suitable for
real world deployment. We can integrate a GSM module for sending SMS over cellular
network if there is no Wi-Fi connectivity, GPS module to send the location of the patient
if they have wandered off, etc.
• We could also develop a mobile application or web dashboard in order to provide the
caregivers and nurses with real-time monitoring of the patient’s condition and also stores
historical data for future references for the doctors.
• Some biometric sensors like heart rate sensor, SpO2 sensor to assess the patient’s condition
post-fall. Adding a manual alert button would also be a very beneficial addition, just in
case if the patient is suffering with some other ailment.
• By training a lightweight ML model on real-world motion data collected from the
MPU6050 sensor, the system can be made capable of differentiating between actual falls
and normal daily activities. This would significantly reduce false positives and improve
personalization for different users.

18
 References

1. G. Mahesh and M. Kalidas, "A Real-Time IoT Based Fall Detection and Alert System
for Elderly," in Proc. Int. Conf. on Advances in Computation, Communication and
Information Technology (ICAICCIT), Hyderabad, India, 2023, pp. 327–329, doi:
10.1109/ICAICCIT60255.2023.10465914.

2. J. A. Shirley, R. Jansi, and R. Vyas, "Smart Wearable Socks Enabled with IoT for Fall
Detection," in Proc. Int. Conf. on Electronics and Renewable Systems (ICEARS),
Kattankulathur, India, 2025, pp. 497–499, doi: 10.1109/ICEARS64219.2025.10940604.

3. V. M. Ionescu and F.-M. Enescu, "Using Smart Devices for Fall Detection:
Algorithms, Systems and Applications," in Proc. 14th Int. Conf. on Electronics,
Computers and Artificial Intelligence (ECAI), Pitesti, Romania, 2022, pp. 1–6, doi:
10.1109/ECAI54874.2022.9847475.

4. R. Li, D. Li, and M. Zhang, "A Real-Time Fall Detection System Using ToF Depth
Images," in Proc. 10th IEEE Int. Conf. on Smart City and Informatization (iSCI),
Hangzhou, China, 2022, pp. 41–44, doi: 10.1109/iSCI57775.2022.00016.

5. M. M. J. Glenchrist, A. P. Subash, A. Anto, G. Kattukaran and V. P. Madhanmohan,

"A smart wearable device for tracking and fall detection for geriatric people,
Alzheimer patients and children with autism," in Pro c. 2021 5th Int. Conf. on
Electrical, Electronics, Communication, Computer Technologies and Optimization
Techniques (ICEECCOT), Mysuru, India, Dec. 2021, pp. 1–6, doi:
10.1109/ICEECCOT52851.2021.9708064.

19