Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

FALL DETECTION IN THE ELDERLY USING

NODE MCU

Abstract—Falling down, for the elderly is always a risk. A


minor mishap, such as tripping over a rug or slipping on a most of the research work in fall detection deals with
damp floor, might drastically alter their life. Many elderly applying supervised classification methods. Most of
people get bone fractures each year as a result of falls. A these techniques need training data for falls and
broken bone might trigger more significant health issues in employ either domain knowledge or data analysis
elderly adults, which can result in long-term impairment. If techniques to compute threshold(s) for their
they fall, we must assist them as soon as we can to prevent any identification. The problem with thresholding
such mishaps. The issue is that we can't always monitor their techniques for fall detection is that it is very difficult to
behaviour. The significant increase in the older population and adapt thresholds to new types of falls and generalize
the high risk of falls make the automatic fall detection system
them across different people. If sufficient fall data are
crucial. In order to distinguish falls from regular activities,
existing fall detection systems that use accelerometers as the
available, one-class classifiers can be trained on falls to
detector are frequently built using an empirical acceleration reject normal activities. However, in realistic settings
threshold. It is possible to determine an ideal detection such a strategy is highly unlikely because the
threshold that achieves the desired false alarm rate while availability of sufficient real falldata is difficult . Zhang
enhancing the detection likelihood. To address the et al] train OSVM from falls and outliers from non-fall
aforementioned issue, we will develop a fall detection ADL and s h o w t h a t f a l l s c a n b e d e t e c t e d
prototypein this project that will send an alert to your mobile effectively.They test four OCC methods; K- center,
phone when the fall is detected. Knearest neighbor, OSVM and single class minimax
probability machine (SCMPM) and find that SCMPM
Keywords—component, formatting, style, styling, insert achieves the overall best performance among them. Han
(key words) et al propose to use wireless signal propagation by
employing the time variability and special diversity of
Channel State Information as the indicator of human
I.INTRODUCTION activities. The problem of fall detection is a good
A wearable fall-detection device has been created to example for imbalanced learning because the data for
detect falls using a proposed fall-detection algorithm. normal activities are easy to collect and sufficiently
The growing aging population, especially those over 65 available in comparison to falls. Traditional supervised
years old,are at risk of falls due to declining physical algorithms expect balanced datasets with equal
abilities.If an elder person falls and becomes unconscious misclassification costs for different classes. However,
or unable to move, they may be harmed by the fall's these algorithms fail to represent characteristics of the
injuries. Thus, the development of a system to data and provide unfavorable accuracies when
automatically detect falls in the elderly and other patients presented with imbalanced dataset.
is actively being studied. TheInternet of Things (IoT) is a
growing concern with the seamless interaction among
objects, sensors, and computing devices. Smart
embedded objects, such as fall-detection sensors with
wireless communication, will also become an important
part of the IoT. The identified fall-detection systems can
be divided into two categories: context-aware systems
and wearable devices. The wearable device-based
approach uses clothing with embedded sensors to detect
the body's motion and location. The benefits of wearable
devicesinclude cost efficiency, ease of installation, setup,
and operation. In this project, a fall-detection algorithm
using 3- axis acceleration is proposed, with fall-feature
parameters calculated from the 3-axis acceleration
applied to a simple threshold method.

II. LITERATURE REVIEW


A simple approach is analysed in the paper “Review of
Fall Detection Techniques: A Data Availability
Perspective”, by Shehroz SKhan and Jesse Hoey, where

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

Camera-based methods can be divided into three classes


Several approaches have been proposed for abnormal based on the principles used: inactivity detection, 2D body
activity recognition using the computer vision sensors.This shape change analysis, and 3D head motion analysis. Each
approach is simple; however, choosing a threshold remains approach has its advantages and disadvantages and may be
challenging. However, it is not clear from their work how more suitable for specific situations or populations.
many iterations are needed to terminate the process of Many efforts have been made in the development of fall
outlier detection. Their model assumes that the normal detection technology due to the high demand and potential
activities data contains unusual events and guarantees one market value. Numerous products and literature exist on
outlier per iteration that could result in high false the topic, and further research is ongoing to improve the
alarm rate. The general principle to identify an unusual accuracy and reliability of fall detection methods. By
event is to vary the covariance of the model of normal continuing to develop and improve these technologies, we
events to determine the interval with the highest likelihood can reduce the harm caused by falls and improve the
where events can be classified as ‘not normal’. To model quality of life for those at risk.This project involves the use
dynamic detection of abnormal events, they add a new of a microcontroller( NodeMCU) which is connected to the
factor to the existing SLDS model by inflating the system node MPU6050 (accelerometer and gyroscope sensor).
noise covariance of the normal dynamics. The sensor data is When a person falls there is a sudden change in the
collected using various probes connected to each baby. The acceleration and orientation, there is an increase in the
computation of the factor related to increasing the acceleration. The heart rate for the same is monitored. If
covariance remains challenging and is critical in this thedetected heart rate is less than normal then a notification
application. The general principle to identify an unusual is sent on the app. The message is sent to the caretaker
event is to vary the covariance of the model of normal andthe buzzer goes off to inform people.
events to determine the interval with the highest likelihood
where events can be classified as ‘not normal’. To model
dynamic detection of abnormal events, they add a new
factor to the existing SLDS model by inflating the system
noise covariance of the normal dynamics. The sensor
data is collected using various probes connected to each
baby. The computation of the factor related to increasing
the covariance remains challenging and is critical in this
application
IV. TECH STACK
III. METHODOLOGY
A) Software
Falls are a common occurrence and can have serious The app was developed in Android Studio and was
consequences, especially for the elderly and patients. To supported with a database in firebase. It is an integrated
prevent or minimize the harm caused by falls, detecting development environment for developing new apps for
them as quickly as possible is crucial. There are four different purposes. And it supports many programming
types of falls: fall from sleeping, fall from sitting, fall languages eg: java, C++, and C. For the listed features, 3
from walking or standing on the floor, and fall from libraries additionally have been imported including the
standing on supports. Although these types of falls share mail and its authentication libraries, externally to make the
some common characteristics, they possess significantly working app more accurate and efficient. To save the data
different characteristics.To detect falls, various methods from hackers and other malware we are also using Data
have been developed based on the sensors and security protocols such as Encryption, Firewalls,
principles used. Three main approaches are wearable Cryptographic Agility, Vulnerability scanning. The android
devices, ambiance devices, and camera-based methods.
Wearable devices can be further divided into posture and app code was written in Java with XML involved in GUI
motion devices, while ambiance devices can be further designing. The above decision is made on its flexibility
divided into presence and posture devices. and cross-platform compatibility.We have also used two
external dependencies named TXTLocal which helps in
sending real-time sms to given phone no.s. We have
initiated the automated sms upon not meeting the
necessaryconditions followed by the buzzer alerting.

B) Hardware
This section discusses the basic theory of components used
for this work.

Node MCU
Node Micro Controller Unit is an open-source software
and hardware development environment that is built around
a very inexpensive System-on-a-Chip (SoC) called the
ESP8266. MCU stands for Micro Controller Unit - which
really means it is a computer on a single chip. A
microcontroller contains one or more CPUs (processor
cores) along with memory and programmable input/output
peripherals.

MPU6050
The MPU6050 has an embedded 3-axis MEMS gyroscope,
and a 3-axis MEMS accelerometer. It is very useful for
As soon as the heart rate of the patient goes low a notification is sent
some motion detection. This small module integrates the
on the app in the following manner
logic level converter circuit(which makes it compatible
with the 3.3V-5V voltage level) together with the
MPU6050 sensor.

Pulse sensor
This sensor has two surfaces, on the first surface, the light-
emitting diode & ambient light sensor is connected.
Similarly, on the second surface, the circuit is connected
which is accountable for the noise cancellation&
amplificationThe working of this sensor can be done by
connecting it from the fingertip or human ear to the
Arduino board. So that heart rate can be easily calculated.

As soon as the rate goes below the threshold it sends the notification
of fall being detected.

V. EXPERIMENTAL RESULTS:

There is also an option to call the Ambulance directly from the app in
case of emergency.

Also, during the forecast period of 2022 to 2029,


prominent players who offer smartwatches with fall
detection provide lucrative prospects to market participants.
Moreover, the market will grow due to the development of
fall detection technologies based on Wi-Fi and machine
learning.

References

1. M. Alwan, P.J. Rajendran, S. Kell, D. Mack, S. Dalal, M. Wolfe, and


R.Felder. A smart and passive floor-vibration based fall detector for
elderly, ICTTA '06 ( 2nd Info. & Com. Tech.), vol. 1, pp. 1003 -
1007,24-28 April 2006.
2. D. Anderson, J.M. Keller, M. Skubic, X. Chen, and Z. He,
Recognizing falls from silhouettes, 28th EMBC (28" Int'l Conf.
of the IEEE Eng. in Medicine and Biology Society), pp. 6388 - 6391,
Aug. 2006.
3. A K Bourke, and G. M. Lyons. A threshold-based detection
algorithm using a bi-axial gyroscope sensor, Med. Eng. Phys., 2006.
4. A. K. Bourke, V. J O'Brien J and GM Lyons. Evaluation of a
threshold-based tri-axial accelerometer fall detection algorithm, Gait
Posture,2006
5. G Brown. An accelerometer based fall detector: Development,
experimentation, and analysis, Internal Report, University of
California in Berkeley, 2005.
The app also has the option of showing nearby hospitals for 6. G. Cambul, M.R. Sahin, and M.E. Karshgil. Automatic fall detection
assistance in real time video based applications, 2006 IEEE 14" Signal
Processing and Communications Applications, pp.1-4, April 17-19,
2006.
7. K. Cameron, K. Hughes, and K. Doughty. Reducing fall incidence in
community elders by telecare using predictive systems, 19th EMBC,
VI. CONCLUSION: vol.3, pp. 1036-1039, 30 Oct-2 Nov 1997.
Many cases of elderly individuals falling due to numerous 8. J. Y. Hwang , J.M. Kang, Y.W. Jang , H. C. Kim. Development of
causes, such as heart attacks, go undiscovered since there novel algorithm and real-time monitoring ambulatory system using
isn't a simple gadget available to them that can detect their bluetooth module for fall detection in the elderly, 26* EMBC, San
fall and send a notification to the persons concerned. A Francisco, CA, USA, Sept. 1-5, 2004
device like this is very necessary for today's culture 9. B. Jansen and R. Deklerck. Context aware inactivity recognition for
because most people are too busy at work to see their visual fall detection, Pervasive Health Conference and Workshops,
pp.1-4, Nov. 29-Dec. 1, 2006
elderly relatives, and it is nearly difficult to find out if
anything horrible happens because senior folks are not very 10. A. H. Khandoker, D. Lai, R.K. Begg, and M.Palaniswami. A
wavelet-based approach for screening falls risk in the elderly using
tech-savvy. The gadget is used to send an SMS to the support vector machines, ICISIP 2006 (4th Int'l Conf. Intelligent
designated users. It relies on the MPU6050 sensor, which Sensing & Inf Processing), pp: 184-189, Oct. 15 2006-Dec. 18 2006.
combines an accelerometer and gyroscope to record 11. J. Krauss, and J. Zhao. Activity sensor and fall detection,
acceleration values. manuscript,2008.
12. Y. Lee, J. Kim, M. Son, and M. Lee. Implementation of
accelerometer sensor module and fall detection monitoring system
VII.FUTURE SCOPE based on wireless sensor network, 29" EMBC, pp. 2315 -
The market is also being affected by the increase in 2318,22-26 Aug. 2007.
demand for fall detection systems based on 13. T. Lee, and A. Mihailidis, An intelligent emergency response system:
wearable and smartphone technology to provide preliminary development and testing of automated fall detection.
J.of Telemedicine and Telecare, 11, 194-198, 2005.
improved access to aid in the event of a fall. The
industry is also developing as a result of rising 14. L. J. Lin, F. T. Chiou and H. H. Cohen. Slip and fall accident
medical spending costs and rising demand for prevention: A review of research, practice, and regulations,
Journalof Safety Research, vol.26(4), pp.203-212.
multimodal technology.
15. C. S. Lin, H. C. Hsu, Y. L. Lay, C. C. Chiu and C. S. Chao, Wearable
Also, the market for fall detection systems is device for real-time monitoring of human falls, Measurement, vol.
positively impacted by rising urbanisation, lifestyle 40(9-10), pp. 831-840, Nov.-Dec. 2007.
changes, an increase in expenditures, and higher 16. C. S. Lin, H. C. Hsu, C. C. Chiu, S. L. Lin, and C. S. Chao, A PDA
consumer spending. based wearable system for real-time monitoring of human falls,
IETEJournal of Research, Vol 52 (6), 403-416, 2006.

17. C. W. Lin, and Z. H. Ling, Automatic fall incident detection in


compressed video for intelligent homecare, ICCCN 2007 ( Proc. of
16th Int'I Conf on Com. Comm. & Networks), pp. 1172-1177, Aug.
2007.
18. S. Luo, and Q. Hu. A dynamic motion pattern analysis approach to
fall detection, ISCAS 2004, Vol 1, pp:5-8, 1-3 Dec. 2004.
19. M. J. Mathie, J. Basilakis, and B. G. Celler. A system for monitoring
posture and physical activity using accelerometers, 23rd
EMBC,Istanbul, 25-28 Oct. 2001.
20. S. G. Miaou, F. C. Shih, and C.Y. Huang. A smart vision-based
human fall detection system for telehealth applications, 3rd IASTED
int'1 Conf. on Telehealth, pp.564, May 30-June 1, 2007, Montreal,
Quebec, Canada.

You might also like