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

net/publication/352311882

Implementation of Wireless Sensor Network For Detection of Forest Fire. Title

of Book: NCRISET 2017 Publisher: G.H. Raisoni Institute of Engineering and
Technology, Nagpur ISBN 978...

Conference Paper · March 2017

CITATIONS READS

0 214

4 authors, including:

Saloni ..
KIIT University
3 PUBLICATIONS 7 CITATIONS

SEE PROFILE

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

UderGraduate View project

All content following this page was uploaded by Saloni .. on 11 June 2021.

The user has requested enhancement of the downloaded file.

 2nd National Conference on
Recent Innovations in Science, Engineering & Technology -2017

Implementation of Wireless Sensor Network For

Detection of Forest Fire
Achal Gaddamwar, Mayuri Nandagawali Mr. Rahul Suryawanshi
Saloni, Suwrna Bandgar Assistant Professor, Computer Science and Engineering
Student, computer science and Engineering G. H Raisoni Institute of Engineering and technology,
Final Year, Nagpur-440028
G.H. Raisoni Institute of Engineering and Technology,
Nagpur-440028

Abstract- In this paper, we present a new approach for monitoring and early detection of forest fires is required
forest fire detection. As we know, forest fires becomes a for efficient prevention and protection against the loss.
serious problem to the biosphere. Therefore, it requires Systems employing charge-coupled device cameras and
such kind of monitoring system for forest fires which infrared detectors, satellite systems and images [5] and
should be able to make a real-time monitoring of the target wireless sensor networks [3,4] are some of the
region and the early detection of fire threats. Here, we are technologies and system which have been proposed to
presenting a new approach for forest fire detection which
detect forest fire. In this paper we state the detection of
is based on the integration of Data Mining techniques into
sensor nodes. The idea is very simple we are going to use a forest fire by using wireless sensor network. A wireless
clustered WSN where each sensor node will capable of sensor network(WSN) is a network that consist of
detecting fire using a classifier of Data Mining techniques. spatially distributed autonomous devices using sensors
Whenever a fire is detected by any of the nodes that node to monitor different physical or environmental
will going to send an alert through its cluster-head and conditions such as temperature, smoke, light, etc and
from that cluster-head an alert will pass through gateways cooperatively pass their data through the network to a
to other cluster-heads. In this way, alert will finally reach main location. A wireless sensor network is usually
to the base station to inform the firefighting crews. consist of few sinks and a large quantity of small sensor
nodes, which are able to detect, process and transfer
Keywords- Wireless sensor network, Data mining, data from one node to another. For detecting fire, a
1. INTRODUCTION: sensor node can be placed or installed in the forest so
A forest fire is an uncontrolled fire occurring in nature. that it collect data such as temperature or smoke, and
Sometimes, the forest fire is so large that it takes a long deliver these data to the base station (sink node) when
time for the fire fighting crews to gain control over the these data reach to base station, it will be manually
situation. This could result in massive destruction. In the processed and analysed. In our approach, we are using
era of global warming, rising temperature is one of the the integration of data mining techniques for solving the
main causes for forest fires. In 1995,around 3.75 lakh limitations e.g., for improving the quality of collected
hectares of area was affected in Uttrakhand. In data in an intelligent way. Data Mining is a process that
1999,around 80,000 hectares of forest was destroyed in extracts the hidden patterns from a large volume of data
Ganga-Yamuna doab region .In 2008,the forest fire in and a critical component consist in knowledge
Melghat in Maharashtra affected some 10,000 hectares discovery process [2]. We are also implementing this
of forest. In 2010, about 19,000 hectares forest were system in a real time for predictive analysis. For real
affected by fire in Himachal Pradesh. Sikkim, time monitoring, one of the challenges is to detect the
Uttrakhand and Himachal Pradesh are among the most forest fire as early as possible and therefore for the
forest fire affected states of india. Also in 2016 forest accomplishment of this aim, we have proposed a new
fire have been noted in numerous places across the approach which is based on Data Mining techniques and
Indian state of Uttarakhand. These fires set mainly in a wireless sensor network clustered architecture.
pine forest in the slopes of sub-Himalayan region, Individually Each sensor node is capable of predicting a
produced clouds of smoke. Moreover, forest fire is not forest fire by using a classifier of data mining technique.
only a serious problem in India but it is across the globe. Whenever a fire is detected by any of the sensors, the
On an account from January 2010 to October 2015 over corresponding sensor node will send an alert to its
8388 forest fires have been counted in the cluster-head which will pass through gateways and then
Mediterranean region in France. Forest fires can be that alert will pass to other cluster-heads until it reaches
deadly threats to the biosphere. It is dangerous for both to the base station to inform the firefighters. After
flora and fauna. In some of these fires, large areas of receiving the message from the base station in our
forests of more than 21954.61 hectares have been system, fire fighters will take further actions.
destroyed and many people or animals have died which
will result in massive destruction. Therefore, the 2. LITERATURE REVIEW:

ISBN 978-93-5268-408-3
88
 2nd National Conference on
Recent Innovations in Science, Engineering & Technology -2017

In past many years, several studies have been

done for forest fire detection using wireless sensor
networks. The authors Massinissa Saoudi, Ahcène
Bounceur, Reinhardt Euler, Tahar Kechadi of [1]
proposed Data Mining Techniques Applied to Wireless
Sensor Networks for Early Forest Fire Detection in
which the goal was early detection and high accuracy of
the localization of forest fire for a rapid intervention of
firefighting personnel at the correct place.
The authors, Mirjana Maksimovi´c and
Vladimir Vujovi´c of [2] proposed various data mining
techniques on WSN by using weka tools. Its goal was to
perform comparative anaysis of various data minining
techniques and to determine which one is best suited for
fire detecion.
The authors, Yunus Emre Aslan, Ibrahim
Korpeoglu, and Ozgur Ulusoy of [3] proposed the
various cluster network architecture for forest fire
detection that provide effective and efficient way to
consume less energy.
In [4], the authors, Hamdy Soliman, Komal Figure 2: Flowchart for implementation of wireless
Sudan, and Ashish Mishra proposed another sensor networks for forest fire detection
methodology for forest fire detection which combines
WSN with artificial neural network(ANN). In this 3.1 Hardware Implementation:
system they gathered the data from the particular region In real time implementation, we have designed a cluster
throgh WSN by sensors(like temperature and smoke). that consist of four nodes, a cluster head, a base station
Then these data are transmitted to ANN at base station. which is connected with GSM module.
Then the base station use that received data to test
whether it belongs to fire class or not.
In [5], the authors, Jaime Lloret, Miguel Garcia, Diana
Bri, and Sandra Sendra proposed a unique methodology
For forest fire detection. In this, they are conbining
wireless local are network with sensor node technology
for detction of forest fires. They use multi-sensor nodes
and IP cameras for detection and verification of fires.

3. PROPOSED METHODOLOGY:
When a fire is detected, the corresponding node will
send an alert through its cluster-head which will pass
through gateways and other cluster-heads until it will
reach the sink in order to inform the fire fighters.
Through extensive simulation experiments and its
hardware implementation, we show that our approach
can provide a fast reaction to forest fires while
consuming energy efficiently. Figure 3: Designing of cluster nodes and cluster head
In our approach,first we have designed a cluster as
shown in figure 3. For designing cluster nodes and its
head we have used PCB plates. Each node consist of
two sensors(i.e., temperature sensor and smoke sensor).
It also consist of IC 8051 microcontroller, Encoder IC
HT12E, Decoder IC HT12D, Regulator IC 7805,
Receiver, Transmitter, LED’s.

Figure 1: architecture of forest fire detection

ISBN 978-93-5268-408-3
89
 2nd National Conference on
Recent Innovations in Science, Engineering & Technology -2017

Figure 6: Implementation of wireless sensor network

for detecting fire
Figure 6 shows a successful implementation of
wireless sensor network for forest fire detection. Here,
when a smoke sensor of first node detect smoke, it send
an alert to cluster head and then cluster head will pass it
to the base station. When base station receives an alert it
will notify the alert by blinking its first LED and sends
the SMS to fire fighters using GSM service to take
further actions.

Our work is mainly divided into four parts: Clustering,

Routing, Fire Detection and Routing alert.

3.2 Clustering:
Figure 4: Base station with GSM module
Clustering is the grouping of a particular set of objects
based on their characteristics, aggregating them
Then, we have designed a base station and connect it
according to their similarities.In our work we have use
with a GSM module as shown in figure 4. Base station
cluster of wireless sensor network as it has various
receives an alert from the cluster head and has
advantages:
capability of detecting the node in which fire is
 It transmits aggregated data to the data sink,
detected. In our work, we have connected the base
reducing number of nodes taking part in
station with the GSM module for sending a message to
transmission.
fire fighters. Base station consist of IC MAX 232, IC
8051 microcontroller, Decoder IC HT12D, Regulator IC  It is useful in energy consumption.
7805 and LEDs.  It provides scalability for large number of
nodes.
 It also reduces communication overhead for
both single and multi-hop.
TABLE II.
3.3 Routing:
In our work, we are using routing algorithm based on
cluster network. At the beginning, after applying
clustering algorithm, the routing table is empty. When
the sink(base station) sends the discovery route message
containing its idi, then the corresponding gateways will
receive the message and store the identifier idi of the
base station in the routing table, then that gateway node
will pass that meassage which contain identifier idi to
the corresponding cluster-heads except the base station.
After receiving the meassage of route discovery, the
cluster head will store the gateway identifier idi in their
Figure 5: Connection setup of the hardware components routing table and this process will continue until all the
cluster-heads and gateways receive the discovery route
As shown in figure 5, we have done a setup of wireless message. When all the cluster-heads and gateways
sensor network with a base station and a GSM module. received the message, they send the meassage to the
sink(i.e., base station). For routing the meassage to the
sink, multiple paths can be used by cluster-heads and
gateways.

3.4 Fire Detection:

Our work is based on real time monitoring of
data in the targeted region. For detection of forest fire,
we use two different sensors, i.e., temperature sensors
and smoke sensors. When the temperature exceeds its
defined limits or smoke is detected then sensor will
predict the fire. After detecting the fire, it will send an
alert to its cluster-head. This cluster-head then pass that
alert to the next head through the gateway. The last
cluster head then finally pass the alert to the base station
where base station will determine the concerned node or

ISBN 978-93-5268-408-3
90
 2nd National Conference on
Recent Innovations in Science, Engineering & Technology -2017

cluster in which fire is detected and finally notify the

fire fighters to take further actions. REFERENCES
[1] Massinissa Saoudi, Ahcène Bounceur, Reinhardt Euler, Tahar
Kechadi. “Data Mining Techniques Applied to Wireless Sensor
4. EXPECTED OUTCOME: Networks for Early Forest Fire Detection” ICC16 international
The expected outcome in our work is that each conference of internet of things and cloud computing,
of the sensor node is capable of detecting the fire. When Cambridge, United Kingdom 26. ISBN:978-1-4503-4063-2
a fire is detected by a sensor node, it will send an alert March 22,23 2016
[2] Mirjana Maksimovi´c and Vladimir Vujovi´c. “Comparative
to a cluster head. Then this cluster head will pass that analysis of data mining techniques applied to wireless sensor
alert to the next cluster head through the gateway. In network data for fire detection”. JITA- Journal of Information
this way the alert will pass to the base station, then the Technology and Applications (Banja Luka)-APEIRON, vol 6
base station will send the message to nearest fire Issue 2,ISSN 2232-9625, 2013
[3] Yunus Emre Aslan, Ibrahim Korpeoglu, and Ozgur Ulusoy “A
fighters. framework for use of wireless sensor networks in forest fire
detection and monitoring”, Computers, Environment and Urban
5. FUTURE WORK: Systems, vol 36 Issue 6 pp 614–625, 2012
Our future work will be based on providing the [4] Hamdy Soliman, Komal Sudan, and Ashish Mishra. “A smart
best classification techniques for comparative analysis forest-fire early detection sensory system: Another approach of
utilizing wireless sensor and neural networks”. In Sensors, 2010
of various data mining techniques. Also we will work to IEEE, pages 1900–1904. IEEE, 1-4 November 2010
provide more efficient clustering algorithm for forming [5] Jaime Lloret, Miguel Garcia, Diana Bri, and Sandra Sendra. “A
WSN which require less response time. By using wireless sensor network deployment for rural and forest fire
efficient clustering algorithm we can make our WSN detection and verification”, MDEI Journal on Sensors, vol 9
Issue 11 pp 8722–8747, October 2009
less energy consuming. We also wants to provide more
secure exchange of data throughout the network.

6. RESULT:
As per our expectations, we have successfully
performed simulation experiments and implementation.
We have shown our work in simulation by using
MATLAB.In simulation, there are number of clusters.
Each cluster consist of a pairs of temperature sensor and
smoke sensor. Whenever the fire is detected by any of
the member node, it will send an alert to its cluster head.
The cluster head will pass that alert to the next cluster
head through gateway. In this way, alert will finally
reach to the last cluster head and it will send that alert to
the base station which has capability of detecting the
node in which fire is detected and finally the message
will send by the base station to the nearest fire fighter.
We have also done the real time implementation for
forest fire detection in which same approach is used. In
this, each cluster is using the shortest path to send an
alert to the base station. For notifying the fire fighters,
the base station sends message through GSM.

7. CONCLUSION:
In this paper, we proposed a new approach by using
wireless sensor networks for forest fire monitoring and
detection. This work is based on measuring and
combining real data from diff erent sensors
(temperature, smoke). We have taken into account all
the characteristics of wireless sensor networks that
include low energy capacity computing limitations, low
memory capacity of sensor nodes and various
environmental conditions which can affect detection of
fire and WSN performance. Our work is based on
monitoring the real data from different sensors (i.e.,
temperature and smoke) and applying classification of
data mining techniques for forest fire detection. A node
itself detect fire Then it abort the normal values and
sends the abnormal values to base stations for detecting
the location of fire and to inform the fire fighters.

ISBN 978-93-5268-408-3
91
 2nd National Conference on
Recent Innovations in Science, Engineering & Technology -2017

ISBN 978-93-5268-408-3
92

View publication stats