You are on page 1of 7

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

net/publication/338799096

Design and Implementation of Fire Detection System Using Fuzzy Logic


Algorithm

Conference Paper · November 2019


DOI: 10.1109/APWiMob48441.2019.8964144

CITATIONS READS

3 323

3 authors, including:

Istikmal Istikmal Nyoman Karna


Telkom University Telkom University
37 PUBLICATIONS   104 CITATIONS    31 PUBLICATIONS   102 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Knowledge Model to support autonomous knowledge transfer between Knowledge-based Systems View project

Audio Band Analog Signal Measurement Instrument for Vocational School Practicum Aids View project

All content following this page was uploaded by Nyoman Karna on 16 November 2021.

The user has requested enhancement of the downloaded file.


2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

Design and Implementation of Fire Detection


System Using Fuzzy Logic Algorithm
Anak Agung Putu Bunga Surya Devi, Istikmal, Nyoman Karna
Faculty of Electrical Enginering, Multimedia-Big data and Cybersecurity laboratory, Telkom University, Bandung, Indonesia

Abstract— One of the features of a smart home is fire functions allows fuzzy logic to make objective observations of
detection. There have been many developments in previous values that are subjective. Furthermore, this membership
studies, but not many have implemented a detection system with function can be combined to make the expression of concepts
the fuzzy logic method. Therefore, in this research we have clearer.
developed a fire detection system that applies fuzzy logic
methods and algorithms. We used Raspberry pi 3 as the
embedded system, DHT-11 and MQ-2 sensors to detect fires. II. BASIC THEORY
Detection results will be processed using a fuzzy system and the
results will be notified through the WhatsApp application and A. Smart House
monitored through the web. The test results show that the Smart House is a place of residence wherein there are
system that has been developed is running well and can be used various devices that have been connected to the internet
as a fire detection system in smart homes. network by managing and controlling it can be done remotely
via an internet network [3].
Keywords— Fire detection, Raspberry pi 3, Fuzzy Logic In this system Smart House that will be created is a
Algorithm, DHT-11, MQ-2. solution in terms of fire prevention from an early age in the
house. Smart House which will be implemented using the
I. INTRODUCTION Fuzzy Logic Algorithm as the brain, where Smart House must
An accident is an event that cannot be predicted when it be connected directly to the surrounding Wi-Fi network, in
occurs, one of which is fire. Sometimes fire prevention cannot order to be able to send notifications to the user.
be done quickly due to various factors, including because B. Rasberry Pi 3
residents and homeowners are late in detecting fires so that
efforts to mitigate and reduce risk cannot be carried out on The Raspberry Pi 3 Model-B is the 3rd generation
time. Now, people who use fire sensors at home broadcast fire Raspberry Pi which is a single-board computer developed in
information through alarms, only then do people find out about the United Kingdom by the Raspberry Pi Foundation. The
the fire after hearing the sound of the fire alarm. This creates a Raspberry Pi 3 Model B is a mini computer with a 64-bit
weakness that is if the fire alarm cannot be heard by the 1.2GHz quad-core processor, 1GB RAM, Wi-Fi and
property owner who is not in the location and the community Bluetooth controller 4.1. It also has 4x USB 2.0 ports, 10/100
outside the radius is heard an alarm sound. So that the Ethernet, 40 GPIO pins, Full-size 1.3a HDMI port, Camera
mitigation of the potential for the spread of fires cannot be Interface (CSI), 3.5mm analog audio Combination and
done. composite video jack, Display interface (DSI), MicroSD slot
The previous research was carried out by [1]about a fire and Video Core IV multimedia or 3D core graphics @
warning system in a housing complex to the fire department 400MHz / 300MHz. [4].
using a GSM modem, which in this research has succeeded in C. The Sensor
designing a device that can detect fires and send information
The sensor is an electronic component that functions to
directly to homeowners in the event of a fire through voice
calls. Design and implementation of location-based home fire convert mechanical, magnetic, heat, light, and chemical
detectors using WSN [2]. In this case, the author will develop quantities into electrical quantities in the form of voltage,
the research using Raspberry Pi 3 as a Server and Fuzzy Logic resistance and electric current. Sensors are often used for
Algorithm as the method, besides that the author also uses detection when measuring or controlling [5].
WhatsApp as media to distribute fire information to DHT-11 is a temperature and humidity sensor, this sensor
homeowners and others (neighbor or neighborhood chief ) via has a digital signal that is calibrated with a complex
notifications also web monitoring. temperature and humidity sensor. This technology ensures
To detect fire, a sensor consisting of temperature and gas is high and excellent performance in the long run.
used. Based on previous research, the authors chose to use the Microcontroller connected at high performance of 8 bits.
fuzzy logic algorithm to be able to make decisions more These sensors include resistive elements and NTC
accurately when compared to other logic systems. Fuzzy logic temperature measuring devices. Has excellent quality,
is suitable because it uses non-linear linguistic values. Fuzzy responsiveness, anti-interference ability and high benefits.
can express concepts that are difficult to formulate, such as
"dangerous room temperature”. The use of membership

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 99


2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

Each DHT-11 sensor has a very accurate calibration


feature of the calibration chamber humidity [6]. The (2)
calibration coefficient is transferred in the OTP memory
program, an internal sensor that transfers signals in the
2. Triangle Function
process, we have to change the calibration coefficient. DHT-
11 has integrated single-cable serial interface system, small The form of a triangle function is basically a combination of
size, low power, signal transmission distance up to 20 meters. two linear curve lines. Triangle Membership Function:
DHT-11 is a 4-pin single package line. Convenient
connection, special packages can be provided according to
user requirements.
MQ-2 is an electronic component to detect hydrocarbon (3)
gas levels such as isobutane (C4H10), propane (C3H8),
methane (CH4), ethanol (CH3CH2OH), hydrogen (H2),
smoke, and liquid petroleum gas (LPG)
This sensor output in the form of analog resistance can
3. Trapezoidal function
easily be converted into voltage by adding one ordinary
resistor (can also use a potentiometer so that the detection The function of the trapezoidal curve is basically almost the
sensitivity threshold can be adjusted as needed). By same as the triangle function, except that the trapezoidal
converting this impedance to voltage, the sensor readings can function has several points whose member values are equal to
be read by the ADC (Analog to Digital Converter) pin on the 1. Trapezoidal Membership Function:
microcontroller [7].
D. Fuzzy Logic Algorithm
The fuzzy word itself has several definitions namely, blur,
dim, and vague, while for a fuzzy system is a system built (4)
based on fuzzy logic theory. So that fuzzy logic is a
calculation method that uses language (linguistic) instead of
calculating using numbers or numbers. For example, a
measure of room temperature can be expressed in fuzzy logic
theory with cold, normal, or hot words. The form of language Inference process or fuzzy rule is a stage that does a
or words in fuzzy logic is not as detailed as using numbers, but reasoning using fuzzy input and fuzzy rules that have been
the use of fuzzy logic theory will be closer to human intuition. determined, so that it will produce a fuzzy output. In the fuzzy
Because fuzzy logic theory has a membership degree of more rules, have IF antecedent then consequent rules. At this stage
than two values, namely the value between 0 and 1, while in the fuzzy rule model used is the Mamdani Method. In this
digital logic, only has two values, namely the value of 0 or 1. process use "Min-Max Method".
In a classic set there is a very important term, namely crisp In the defuzzification stage a mapping of fuzzy output
set. Crisp set is a set that distinguishes members and members values is carried out resulting from the evaluation of the rules
with clear boundaries. In the fuzzy logic system there are three to the value of the quantitative output values in accordance
main components, namely fuzzyfication, inference, and
with the system applied. In this system the centroid method is
defuzzyfication.
Fuzzyfication process is to change inputs that have definite used:
truth values into a set of fuzzy variables. In the fuzzyfication
process in the form of membership function by creating a
(5)
curve that shows the mapping of each input value into
membership degrees that have interval values between 0 and 1.
There are many types used in determining fuzzy set
membership functions, namely [8]: III. DESIGN SYSTEM

1. Linear Representation Function A. Block Diagram


In linear representation in mapping input values into Before designing this program system, a block diagram is
membership levels they are described in a straight line. Linear needed whose function is to know the system description in
Membership Function Up: general. Each block has its own function and work method.
Here is a general description of the system.
Figure 1 shows the General Block Diagram System above
(1) for designing a fire detection system using the Fuzzy Logic
Algorithm method, which is first of all in the Smart House
must have a sensor that has 2 parameters (temperature and gas)
Linear Membership Function Down: namely DHT-11 and MQ-2 The sensors are connected to the
Raspberry Pi 3. Data detected from the sensor will be
processed into Fuzzy Logic, according to predetermined

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 100


2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

conditions, if it does enter into a dangerous criterion (fire) then C. Fuzzy Design of Membership Function
it will immediately send information via notification using 1. Temperature
Whatsapp to Users and can also be monitored through web
monitoring.
(6)

(7)

(8)

Figure1. Block Diagram System in General

Figure 3. Temperature membership function chart

In Figure 3. it is explained that the temperature


membership function consists of three conditions, namely
cold in the temperature range of 0-25 degrees Celsius, then in
normal conditions the temperature is in the range of 10-40
degrees Celsius and finally in cold conditions, the temperature
is in the range of 25- 50 degrees Celsius. These values are
based on previous research on fires by [9] and based on direct
evaluation by the author.
2. Gas

B. Flowchart Detector System


(9)
Figure 2. Flowchart Detector System

In Figure 2. the Detection Flowchart System above can be


seen from the start of the flow, then the sensor detects signs of
fire in accordance with 2 sensor parameters (temperature and
gas) then enter into Fuzzy Logic to determine a decision that (10)
can be monitored through web monitoring, if entered the
dangerous category, the server will send information via
notifications using Whatsappp to the user and if it is not
dangerous, the sensor will detect it again.

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 101


2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

previous research on fires, same like the values of


temperature and based on direct evaluation in this research.
(11)
Table 1 shows the temperature and gas rules.

Table 1. Temperature and gas rules

IV. RESULTS AND ANALYSIS


Figure 4. Gas membership function chart
A. Tool Parameter Testing
In figure 4, it is explained that the gas membership
function is divided into three conditions, namely thin In this test, only the data from the DHT 11 sensor was
conditions in which the gas range contained is 0-50 ppm. In tested with a digital room temperature thermometer. The
normal conditions, the gas contained therein is in the range of purpose of this test is to find out the error system of the sensor
25-75 ppm, whereas in thick conditions it has a range of 50- when compared to a digital room temperature thermometer.
Figure 6 shows DHT11 sensor test result. Temperature
100 ppm. These values are based on previous research on
reading is done by bringing the sensor closer to the fire source
fires and based on direct evaluation by the author.
to increase and optimize the temperature reading. The output
3. Detection Condition
of sensor readings can be seen through the monitor used.
Based on the tests carried out, the DHT11 sensor readings are
as follows:
(12)
Table 2. DHT11 Sensor Test Results

(13)

Figure 5. Gas membership function chart

In Figure 5 above, it can be explained that the membership


function condition is divided into two conditions, namely the
harmless condition enters the 0-80 range, while the dangerous
condition enters the 50-100 range. These values are based on Figure 6. DHT11sensor test

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 102


2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

B. Testing of Fuzzy Logic Parameters d) The trial is done in 8 times


In this test, the test is a Fuzzy Logic process. To find out
success in determining a decision on the condition of the room Table 3 shows the network throughput and delay for the
from the fire detection system made. The decisions made system.
based on the determination of Fuzzy Logic can be divided into Table 3. Results of Testing
2 conditions which are not dangerous and dangerous which
can be categorized manually using Fuzzy formulas and
automatically from the system.
1. Harmless Condition
Temp.= 27°C & Gas=0,0051 ppm

D. System Notification Display


The notification display of the system can be divided into
2, namely via WhatsApp and web monitor. Figure 7 and
Figure 8 shows the WhatsApp and web monitor notification.

2. Dangerous Condition
Temp.= 34°C & Gas=30,7063 ppm

Figure 7. Notification Display On Whatsapp

C. Testing of Network Parameters


In this parameter, the test carried out is sending data from
Throughput and Delay when sending data from the system to
the User via WhatsApp notification. Testing is done 8 times.
In this test use a reference based on:
a) The software used is Wireshark.
b) When retrieving data starting from activating the
recording button until it stops recording on the
Wireshark.
c) The system is set up as well as testing the system, then Figure 8. Web Display Monitor
the off button is canceled or canceled by the user.

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 103


2019 IEEE Asia Pacific Conference on Wireless and Mobile (APWiMob)

V. CONCLUSION and Cloud) (I-SMAC), Palladam, 2017, pp. 65-70.


The fire detection system has been designed and [4] T. V. G. O. E. Abdalla, Implementation of Spy Robot for
implemented with the Fuzzy Mamdani method successfully A Surveillance System using Internet Protocol of
determines a decision in a certain condition properly. The Raspberry Pi, India: 2nd IEEE International Conference
results of the tests show the optimal output and according to On Recent Trends In Electronics Information &
what the author expects. This system managed to detect the Communication Technology, 2017.
conditions around the room with DHT11 and MQ2 sensors, [5] P. Mekikis, G. Athanasiou and C. Fischione, "A Wireless
then the data obtained from sensor readings were processed by Sensor Network Testbed for Event Detection in Smart
the Fuzzy Mamdani method in Raspberry Pi 3, then produced Homes," 2013 IEEE International Conference on
an output if the situation was dangerous, in the form of Distributed Computing in Sensor Systems, Cambridge,
notification via WhatsApp. Whereas, if the output state is not MA, 2013, pp. 321-322.
dangerous, it will not send the notifications and will continue
to detect and can be seen through web monitoring. The system [6] T. Adiono, M. Y. Fathany, S. Fuada, I. G. Purwanda and
created by the author has succeeded in detecting the S. F. Anindya, "A portable node of humidity and
conditions around the room in Real Time and continuously. In temperature sensor for indoor environment monitoring,"
addition, data transmission is strongly influenced by internet 2018 3rd International Conference on Intelligent Green
(Wi-Fi) network connections around the room, this is because Building and Smart Grid (IGBSG), Yi-Lan, 2018, pp. 1-5.
notifications via WhatsApp and Web Monitoring are Real- [7] Fadhil Puri Himawan, Unang Sunarya, Dwi Andi
Time so that data is always updated continuously and Nurmantri, Design of Smoke Detector based on
periodically. Microcontroller, GSM Module, Smoke Sensor and
Temperautre Sensor, e-Proceeding of Applied Science :
REFERENCES Vol.3, No.3 Desember 2017.
[8] Abdillah Alda Rosi, Angga Rusdinar, Erwin Susanto,
[1] F. R. Fauzi, Fire warning system from housing complex to Design and Implementation of Control Exhaust Fan
fire department using GSM modem, 2012. Speed-based Arduino with Carbon Monoxide Smoke
https://openlibrary.telkomuniversity.ac.id/pustaka/141403. Censor, e-Proceeding of Engineering : Vol.2, No.2
[2] I. L. Muflikhah, Design and Implementation of House Agustus 2015.
Fire Detectors using Location based Wireless Sensor [9] R. N. Yanuar, Implementasi Sistem Pemadam Kebakaran
Network, e-Proceeding of Engineering Vol.5 No.1, 2018. Pada Ruang Tertutup Berbasis Arduino Menggunakan
[3] T. Malche and P. Maheshwary, "Internet of Things (IoT) Logika Fuzzy, Malang: Universitas Brawijaya, 2019.
for building smart home system," 2017 International
Conference on I-SMAC (IoT in Social, Mobile, Analytics

978-1-7281-4865-6/19/$31.00 ©2019 IEEE 104

View publication stats

You might also like