See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/303404955

Design and Implementation of Fall Detection System Using MPU6050 Arduino

Conference Paper in Lecture Notes in Computer Science · May 2016

DOI: 10.1007/978-3-319-39601-9_16

CITATIONS READS

32 21,699

3 authors, including:

Prof. Ziad Tarik Al-dahan Nasseer Bachache

Al-Bayan University University of Alkafeel
106 PUBLICATIONS 334 CITATIONS 21 PUBLICATIONS 223 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Prof. Ziad Tarik Al-dahan on 16 July 2019.

The user has requested enhancement of the downloaded file.

 Design and Implementation of Fall Detection System
Using MPU6050 Arduino

Ziad Tarik Al-Dahan2, Nasseer K. Bachache1 ✉ , and Lina Nasseer Bachache2

( )

1
University College of Humanity Studies, Najaf, Iraq
[email protected]
2
Al-Nahrain University, Baghdad, Iraq

Abstract. Fall is the most significant causes of injury for Elderly or Epilepsy.
This has led to develop a many types of automatic fall-detection systems.
However, prevalent methods only use accelerometers to isolate falls from activ‐
ities of daily living (ADL). This paper proposes combination of a simple threshold
method and acceleration measurement to detect falls and fall-detection. To
demonstrate the activity of proposed scheme a device has been designed. We used
an Arduino-UNO, also we used MPU6050 as a sensor and we can measure the
velocity and acceleration by calculate the derivative for the phase this program
was C-language. Several fall-feature parameters and possible falls are calibrated
through an algorithm. The implementation of program built to read an analogue
variable from its port as an additional adjustment to fixed the upper and the lower.
The total sum acceleration vector ACC to distinguish between falling and ADL.
The results using the simple threshold, PMU, and combination of the simple
method and MPU were compared and analyzed. The proposed MPU reduced the
complexity of the hardware also the algorithm exhibited high accuracy.

Keywords: Fall Detection Systems (FDS) · Activities of Daily Living (ADL) ·

MPU6050-Arduino

1 Introduction

Fall is the most significant causes of injury for elderly. These falls are because many
disabling fractures that could eventually go in front to death due to complications. Most
elderly (over 75 years old) have fallen at least once a year, and 24 % of them have severe
injuries [1]. Among people affected by Alzheimer’s disease, the probability of a fall
increases by three times. Elderly care can be improved by using sensors that monitor
the vital signs and activities of patients, and remotely communicate this information to
their doctors and caregivers. For example, sensors installed in homes can alert caregivers
when a patient falls. Some fall detection algorithms also assume falls happen when the
body lies prone on the floor. But they are less effective when a person’s fall posture is
not horizontal as shown in Fig. 1.
The consequences of a fall can vary from scrapes to fractures and in some cases lead
to death. Even if there are no immediate consequences, the long-wait on the floor for
help increases the probability of death from the accident. For this reason, fall detection
is an active area of research. Most of the research on falls in which accelerometers is

© Springer International Publishing Switzerland 2016

C.K. Chang et al. (Eds.): ICOST 2016, LNCS 9677, pp. 180–187, 2016.
DOI: 10.1007/978-3-319-39601-9_16
 Design and Implementation of Fall Detection System 181

Fig. 1. Falls of patient’s body on the floor

used focus on determining the change in magnitude of acceleration. When the acceler‐
ation value exceeds a critical threshold, the fall is detected [2, 3]. A contribution is made
towards such standardization by collecting the most relevant parameters, data filtering
techniques and testing approaches from the studies done so far. State-of-the-art fall
detection techniques were surveyed, highlighting the differences in their effectiveness
at fall detection. A standard database structure was created for fall study that emphasizes
the most important elements of a fall detection system that must be considered for
designing a robust system [4], as well as addressing the constraints and challenges. In
addition, fall activity patterns are particularly difficult to obtain for training systems.
These systems successfully detect falls with sensitivities. However, focusing only on
large acceleration can result in many false positives from fall-like activities such as
sitting down quickly and running. Furthermore, previous studies used complex algo‐
rithms like support vector machine (SVM) [5] and Markov model [6] to detect the fall.
However, accuracy of these systems has not been proven to be highly effective. They
also use excessive amounts of computational resources and cannot respond in real time.
In this paper we propose a new device based on microcontroller (Arduino-UNO) and
the sensor is MPU6050 Accelerometer and Gyro Chip.

2 Fall Risk Factors

A person can be more or less prone to fall, depending on a number of risk factors and
hence a classification based on only age as a parameter is not enough. In fact, medical
studies have determined a set of so called risk factors:
• Intrinsic:
1. Age (over 75)
2. Chronic disease
3. Previous falls
4. Poor balance
5. Low mobility and bone fragility
182 Z.T. Al-Dahan et al.

6. Sight problems
7. Cognitive and dementia problems
8. Parkinson disease
9. Use of drugs that affect the mind
10. Incorrect lifestyle (inactivity, use of alcohol, obesity)
• Internal Environment:
1. Need to reach high objects
2. Slipping floors
3. Stairs
4. Incorrect use of shoes and clothes
• External Environment:
1. Damaged roads
2. Dangerous steps
3. Poor lighting
4. Crowded places.

3 Mpu-6050

ITG MPU-6050 is a sensor that contains MEMS accelerometer and a MEMS gyroscope
in one chip. Both accelerometer and gyroscope contains 3 axis that can captures x, y and z
with 16-bits analog to digital conversion hardware for each channel. Mpu-6050 uses I2C
for communication which is a multi-master, multislave, single-ended, serial computer bus
with low speed but very useful because uses only two wires: SCL (clock) and SDA (data)
lines. Although the breadboard and the wires are optional items, the two pull-up resistor are
essential. The diagram in Fig. 2 shows how to connect the sensor to a Arduino Uno. It is
also important connect all the sensor pins with the correct arduino pins. The pull-up resistor
will always keep a small amount of current flowing between VCC and the pin, in other
words, it will keep a valid logic level if it is not flowing current in the pin 5. Sensor VDD
- Arduino 3.3 v or 5 v. Sensor GND - Arduino GND. Sensor INT - Arduino digital pin 2.

Fig. 2. Source: RFduino, image by unknown

 Design and Implementation of Fall Detection System 183

Sensor SCL - Arduino SCL dedicated pin = A5. Sensor SDA - Arduino SDA dedicated
pin = A4.

4 Fall Detection Algorithm

The total sum acceleration vector Acc, contain both dynamic and static acceleration
components [3, 7], is calculated from sampled data as indicated in Eq. (1)
√
Acc = (Ax )2 + (Ay )2 + (Az )2 (1)

Where Ax, Ay, Az are the acceleration in the x, y, z axes, respectively.

Similarly to the acceleration, the angular velocity is calculated from sampled data
as indicated in Eq. (2)
√
w= (wx )2 + (wy )2 + (wz )2 (2)

Where wx, wy, wz the acceleration in the x, y, z axes, respectively.

When stationary, the acceleration magnitude, Acc, from tri-axial accelerometer is a
constant, and angular velocity is 0o/s. When the subject falls, the acceleration is rapidly
changing and the angular velocity produces a variety of signals along fall direction.
Since the Fall Index (Acc) requires high sampling frequency and fast acceleration
changes, it will miss falls that happen slowly. Hence, Acc is not used unless we want to
compare the performances of our systems with previous studies that have used the same
positions but with deferent speed and accelerations.
The lower and upper fall thresholds for the acceleration and angular velocity used
to identify the fall are derived as follows [8]:
1- Lower fall threshold (LFT): the negative peaks for the resultant of each recorded
activity are referred to as the signal lower peak values (LPVs). The LFT for the
acceleration signals are set at the level of the smallest magnitude lower fall peak
(LFP) recorded.
2- Upper fall threshold (UFT): the positive peaks for the recorded signals for each
recorded activity are referred to as the signal upper peak values (UPVs). The UFT
for each of the acceleration and the angular velocity signals were set at the level of
the smallest magnitude UPV recorded. The UFT is related to the peak impact force
experienced by the body segment during the impact phase of the fall.
Fall detection algorithms using thresholds are normally divided into two groups, one is
based on the LFT comparison and the other is based on UFT comparison of acceleration
data. Although past research has achieved some significant results, the accuracy is still
below desired levels. In this study adjust the UFT and LFT and found the performance
to be 83.33 % and 67.08 %, respectively [9]. Figure 3 shows the flowchart of our algo‐
rithm which implemented in the program of Arduino_UNO in C-language this program
built to read an analogue variable from its port as an additional adjustment to fixed the
upper and the lower ACC.
184 Z.T. Al-Dahan et al.

Fig. 3. Flowchart of fall detection schema.

5 Results and Discussion

Some existing acceleration-based fall detection systems are only used to distinguish falls
from ADL. However, some activities like sitting down fast also feature large vertical
acceleration. Figure 4 shows the acceleration and rotational rate of the trunk where the
cases as follows: (a) run on a damaged road, (b) walking fast (c) step up a Stair. In
Fig. 4 along three cases there is no fall then there is no alarm. Figure 5 shows the accel‐
eration and rotational rate of the trunk and thigh for sitting fast. Where the cases as
follows: (a) is a dangers fall detected, (b) is a fall posture. In Fig. 5.a the fall detected
very fast and give alarm in the first moment and continued along 30 s while in Fig. 5.b
give alarm after 10 s because there is no fall.
 Design and Implementation of Fall Detection System 185

1.5

0.5

-0.5

-1
0 10 20 30
Time (sec)

(a) Run on a damaged road

1.5

0.5

-0.5

-1
0 10 20 30
Time (sec)
(b) Walking fast
1.5

0.5

-0.5

-1
0 10 20 30
Time (sec)

(c) Step up a Stair.

Fig. 4. Shows the acceleration and rotational rate of the trunk where the cases as follows: (a) run
on a damaged road, (b) walking fast (c) step up a Stair.
186 Z.T. Al-Dahan et al.

1.2
1

0.8
0.6
0.4
0.2
0

-0.2
0 5 10 15 20 25 30
Time (sec)

(a) Dangers fall detected

1.5

0.5

-0.5

-1
0 5 10 15 20 25 30
Time (sec)

(b) Fall posture

Fig. 5. Shows the acceleration and rotational rate of the trunk and thigh for sitting fast. Where
the cases as follows: (a) is a dangers fall detected, (b) is a fall posture.

6 Conclusions

The several fall-feature parameters of the 6-axes acceleration were introduced and
applied according the algorithm. Possible falls were chosen through the simple threshold
and then applied to the MPU to solve the problems such as deviation of interpersonal
falling behavioral patterns and similar fall actions. The test of the proposed device
studied along a different 350 case studies. The parameters of upper and lower of accel‐
eration and velocity have adjusted to give best fall detection with sensitivity, specificity,
and accuracy which were over than 95 %. These results demonstrate the reduction of
the computing effort and resources, compared to those of using all the events applied.
Then the proposed algorithms were very simple because it depend on a simple sensors
(measure the angle) and the program calculate the angular velocity and acceleration.
They can be implemented into an embedded system such as an 8051-based microcon‐
troller with 128 Kbyte ROM. In the future, if the proposed algorithms are implemented
to the embedded system, its performance will be tested in a real time.
 Design and Implementation of Fall Detection System 187

References

1. Hwang, J.Y., Kang, J.M., Jang, Y.W., Kim, H.C.: Development of novel algorithm and real-
time monitoring ambulatory system using Bluetooth module for fall detection in the elderly.
In: Engineering in Medicine and Biology Society, IEMBS 2004, 26th Annual International
Conference of the IEEE, vol. 1, pp. 2204–2207, September 2004
2. Lindemann, U., Hock, A., Stuber, M., Keck, W., Becker, C.: Evaluation of a fall detector based
on accelerometers: a pilot study. Med. Biol. Eng. Comput. 43(5), 548–551 (2005)
3. Kangas, M., Konttila, A., Lindgren, P., Winblad, I., Jämsä, T.: Comparison of low-complexity
fall detection algorithms for body attached accelerometers. Gait Posture 28(2), 285–291 (2008)
4. Abbate, S., Avvenuti, M., Corsini, P., Light, J., Vecchio, A.: Monitoring of human movements
for fall detection and activities recognition in elderly care using wireless sensor network: a
survey, pp. 1–20. InTech (2010)
5. Zhang, T., Wang, J., Xu, L., Liu, P.: Fall detection by wearable sensor and one-class SVM
algorithm. In: Huang, D.-S., Li, K., Irwin, G.W. (eds.) ICIC 2006. LNCIS, vol. 345, pp. 858–863.
Springer, Heidelberg (2006)
6. Ganti, R.K., Jayachandran, P., Abdelzaher, T.F., Stankovic, J.A.: Satire: a software architecture
for smart attire. In: Proceedings of the 4th International Conference on Mobile Systems,
Applications and Services, pp. 110–123. ACM, June 2006
7. Klenk, J., Becker, C., Lieken, F., Nicolai, S., Maetzler, W., Alt, W., Zijlstra, W., Hausdorff,
J.M., van Lummel, R.C., Chiari, L., Lindemann, U.: Comparison of acceleration signals of
simulated and real-world backward falls. Med. Eng. Phys. 33(3), 368–373 (2011)
8. Bourke, A.K., O’donovan, K.J., Olaighin, G.: The identification of vertical velocity profiles
using an inertial sensor to investigate pre-impact detection of falls. Med. Eng. Phys. 30(7),
937–946 (2008)
9. Bourke, A.K., Van de Ven, P., Gamble, M., O’Connor, R., Murphy, K., Bogan, E., Nelson, J.:
Evaluation of waist-mounted tri-axial accelerometer based fall-detection algorithms during
scripted and continuous unscripted activities. J. Biomech. 43(15), 3051–3057 (2010)

View publication stats