2017 International Conference on Computational Intelligence in Data Science(ICCIDS)

Fire-Fighting Robot
Jayanth Suresh
[email protected]
M.S Electrical Engineering
Rochester Institute of technology
New York, US

Abstract— with the advent of technology, humans are replaced Firefighting robot is designed to be an unnamed ground
with robots in life-threatening situations. We aim to design a robot vehicle, implemented for finding and fighting the fire. Few
capable of detecting and suppressing fires. By designing and types of robot’s vehicles are fighting the home fire and fighting
implementing an autonomous robot capable of detecting and forest fires [1]. The fire event may involve more dangerous in
extinguishing flames, disasters can be avoided with minimal risk life. One of the most important systems in the fire detection
to human life. In this research, we illustrate an autonomous robot function system is an intelligent home [2]. The system can have
capable of detecting flames indoors and maneuvering towards the
designed an intelligent multisensor based security that contains
flame to extinguish it with the help of carbon dioxide.
a firefighting robot in our daily life [5]. Security and
Keywords— C02; Arduino UNO; Ping))); IoT;IDE. Firefighting advanced robot which is used in the UK is very low
in cost and have high performance of detecting fire and
I. GOAL extinguishing them [4]. For the novel fire around the surroundings
using image processing and device controlling algorithms to
Using Four-wheeled differential drive robot, flame sensors,
detecting fire fastly and accurately [7]. In present days’ problem
ultrasonic sensors and Arduino IDE, an autonomous was
designed to detect fires, extinguish and avoid obstacles. The of safety on road and railways tunnels considered to more risk
scope of the project in the industrial sector is vast, especially in connected to fires, this robotic system can be installed on the
the fire department. The main operation of the robot is to detect existing tunnels without requiring significant modifications of
and extinguish the fire source with the input from the flame the existing infrastructures [8]. The designing of an
sensors and extinguishing flames with CO2. It has a field of autonomous mobile robot that navigates through a maze
view of 180º to detect flames with obstacle avoidance to searching for a fire in a room (burning candle), it detects the
maneuver in the surroundings. This is achieved by integrating three candle's flame through sensors, and then extinguishes the flame,
flame sensors capable of detecting flames into the robot 60° apart in and returns to a starting location of the maze. This fire-fighting
order to achieve an 180° field of view. Since each flame sensor has a design interdisciplinary design in colleges [9]. Using Fire
field of view of 60°. By using three sensors, a general direction of extinguisher with gas such as CO2 and N2 has advantages
flame can be computed by the data read from each sensor. In order to compared ones with water. For example, they provide electrical
extinguish the flame, a 16g CO2 cartridge mounted on the robot is insulation, they avoid water damages to constructions, electrical
discharged. A CO2 Bike tire inflator is modified as a fire extinguisher equipment’s, paper materials, etc., it may useful in spaces
by attaching a servo motor to the lever to discharge the cartridge. The hidden from extinguishers [10]. In this robotic system, obstacle
firefighting robot is built upon an obstacle avoidance robot. A Ping)))
avoidance and detection using ultrasonic sensors in large fire
ultrasonic sensor is used in order to design an obstacle avoidance robot.
The distance between the objects ahead of the distance sensor and the fields under large smoke at higher temperature situations, the
robot can be computed by reading the digital values from the ping))) transducer, anti-jamming processing is designed [11]. In [12],
ultrasonic sensor from an Arduino UNO. The ultrasonic sensor is authors proposed a PID controller based on back-propagation
mounted on a servo motor to pan the surrounding in order to choose an (BP) neural networks which are used only in PID controller. To
optimum path when it encounters an obstacle. reduce the error rate parameters of PID controller are adjusted
concurrently in real time. In [6], an intellectual PID control,
II. LITERATURE REVIEW which determines system dynamics and states using error and
Robotics motion control can be divided into two categories error rate information as an input of the controller, of the robotic
sensor based system and vision based system. The Sensor-based system. The ‘size and weight’ and ‘cost and performance’ of
system will be controlled by the feedback from the different firefighting robots are problems in present conditions [3]. In this
sensors such as obstacle sensor, IR sensors, flame sensor etc., paper, we are designing a firefighting robot with obstacle
while vision based system uses the cameras and the image avoidance and detecting the fire flame and extinguishing the
processing techniques to find the target position. flame.

978-1-5090-5595-1/17/$31.00 ©2017 IEEE

 2017 International Conference on Computational Intelligence in Data Science(ICCIDS)

III. HARDWARE IMPLEMENTATION between 3.3 and 5 Volts. Its detection range varies from 20 to
In order to obtain the desired mobility and speed, a four- 100 cm. And this distance increases with increase in flame
wheeled differential drive robot is used for this application. intensity. The output from the Ping))) ultrasonic sensor is used
Pololu Dual MC33926 motor controller motor board is used to to determine the distance between the robot and the obstacle by
control the four 12V DC motor on the robot. Arduino UNO is interfacing it with the Arduino UNO. This also provides us with
used to read the Analog and digital values from the sensors and sufficient data to not only avoid obstacles but also to maintain a
also to send control signals to the motor controller board. The safe distance between the fire source and then extinguish the
Arduino UNO generates a PWM signal to control a servo motor fire.
attached to the bike tire inflator in order to discharge a CO2.
This can be achieved by mounting a suitable servo motor’s arm
to control the inflator. A mount is 3D printed as shown in Fig IV. SOFTWARE IMPLEMENTATION
A. The Arduino UNO is programmed using the Arduino IDE.
Once the robot is powered ON, the Arduino will initialize a
digital I/O pin as an output in order to send control signals to the
Ping))) ultrasonic distance sensor. The ultrasonic sensor returns
a pulse whose duration is equal to the time taken for the
ultrasonic pulse to travel from the sensor to the object and back.
This pulse is sent back to the I/O pin of the sensor. Hence the
Arduino digital I/O pin should be made as input in order to read
this pulse. The 'pulseIn()' function of the Arduino library
provides us the pulse duration on the digital pin. We know that
the pulse duration received is for the wave to propagate from the
sensor to the object and back again. Hence, we divide the pulse
duration by 2. In order to calculate the distance, we use the speed
of sound and the pulse duration.

Fig A. 3D printed mount

The 3D printed part will mount the Bike tire inflator and the high
torque servo motor. This unit can then be mounted on an
existing differential drive robot. By turning the servo motor, the
CO2 can be discharged from the cartridge. It is to be noted that
a high torque servo motor such as HS-5685 should be used to
discharge the C02

Fig 1. Program Flow Chart

We assume the speed of sound to be 340m/s. The time taken for

the ultrasonic sound wave to travel 1cm is given by equation (1),

ଵ
ቀ ቁ ‫Ͳͳ כ‬ଷ ‫ ଺ିͲͳ כ‬ൌ ʹͻ‫ݏݑ‬ሺͳሻ
ଷସ଴

Fig B. Prototype
From the pulse duration and the time taken for the wave to travel
The flame sensor outputs a digital and an Analog signal. For 1cm, we can compute the distance of the obstacle from the
higher accuracy, we use the ADC on the Arduino to read Analog Ping))) ultrasonic sensor using equation (2).
signals instead of the digital values. The flame detector sensor
௉௎௅ௌா஽௎ோ஺்ூைே
has a detection angle of 60°, hence three sensors are used to ‫݁ܿ݊ܽݐݏ݅ܦ‬ሺܿ݉ሻൌሺ ሻȀʹͻሺʹሻ
ଶ
obtain a detection angle of 180°. The flame sensor operates
 2017 International Conference on Computational Intelligence in Data Science(ICCIDS)

Once the distance is computed, we compare it to a pre-set

threshold value. The pre-set threshold value is a safe distance
the robot maintains from an object. This value can be obtained
from practical testing. If the measured distance is more than the
threshold, the robot continues to move forward. If the measured
distance is less, the servo motor to which the Ping))) ultrasonic
sensor is mounted is controlled by sending PWM signals using
the 'servo.h' library functions. The sensor is moved left to right
to measure the least obstacle path. Once that is determined, the
robot makes a turn in that direction and continues to move
forward.
There are three flame sensors mounted on the robot. The analog
values from these sensors are read continuously to detect any
fire source. Once the flame sensor is detected, by comparing the
output analog values of each sensor, the direction of flame can
Fig 2. The robot discharges the CO2 to extinguish the fire
be established. Once the direction is established, the robot is
made to turn in small increments for the middle sensor to aim at
the fire source. Then the CO2 cartridge is discharged and the VI. RESULTS
robot makes a turn and continues to maneuver around the
The Fire-Fighting robot is capable of detecting flames and
building.
extinguishing them successfully. The motor controller and
V.SIMULATION
Arduino code work together to control the movement of the
robot with obstacle avoidance. It can detect the flame more
The flame sensor is sensitive to the wavelength of 760nm to effectively in the buildings and fixed lighting conditions. The
1100nm. Hence any source of light in that bandwidth can be robot is designed for the indoor application. Since the ambient
used for testing. Considering safety, a regular lighter or a candle daylight varies throughout the day, a dynamic threshold value
can be used to mimic as a source of the fire. A candle of suitable is necessary to compensate for the change in ambient light. Use
dimension can be kept lit in a room, the robot functioning as an of high torque servo motor was necessary to discharge the CO2
obstacle avoidance robot will detect the flame while Cartridge.
maneuvering across the room. Once the flame is detected, the
robot positions itself facing the flame source and stopping at a VII. CONCLUSION
safe distance and extinguishes the fire. It then automatically
The design was successfully implemented on a four-wheel drive
makes a turn in order to avoid going over the flame source.
robot. The 3D printed mount was mounted on the front of the
robot’s chassis. The battery is mounted inside the robot to
prevent any damage to the battery from external fire source. The
ultrasonic sensor was directly mounted on a servo motor. The
robot successfully detected multiple flame sources and
extinguished them from a safe distance. The speed of the robot
was reduced to the desired speed in order to increase the
operating time and efficient detection of the flame source. The
flame sensor threshold values need to be manually obtained.

VIII. FUTURE WORK

IoT can be implemented onto the robot to send an E-mail to the
user when the robot detects a flame. More sensors can be
mounted, to achieve a 360º field of view. This will reduce the
reaction time to the fire source. The addition of a camera to the
robot in order to distinguish between ambient sun light or fire
Fig 1. The robot detects the flame and halts source. This will help set a variable threshold for comparing the
sensor value to decide whether there is a fire or not. By adding
wheel encoders, we can make the robot maneuver with
precision and independent of battery voltage fluctuation.
Replacing the 16g canister with a portable fire extinguisher
would help suppress larger flame source. The motion of the
robot can my mode smooth by implementing PID control.
 2017 International Conference on Computational Intelligence in Data Science(ICCIDS)

REFERENCES
[1] E. Krasnov and D. Bagaev, "Conceptual analysis of firefighting
robots' control systems," 2012 IV International Conference "Problems
of Cybernetics and Informatics" (PCI), Baku, 2012, pp. 1-3.

[2] K. L. Su, "Automatic Fire Detection System Using Adaptive Fusion

Algorithm for Fire Fighting Robot," 2006 IEEE International
Conference on Systems, Man and Cybernetics, Taipei, 2006, pp. 966-
971.

[3] H. Amano, "Present status and problems of firefighting robots,"

Proceedings of the 41st SICE Annual Conference. SICE 2002., 2002,
pp. 880-885 vol.2.

[4] A. Bradshaw, "The UK Security and Fire Fighting Advanced Robot

project," IEE Colloquium on Advanced Robotic Initiatives in the UK,
London, 1991, pp. 1/1-1/4.

[5] T. L. Chien, H. Guo, K. L. Su and S. V. Shiau, "Develop a Multiple

Interface Based Fire Fighting Robot," 2007 IEEE International
Conference on Mechatronics, Kumamoto, 2007, pp. 1-6.

[6] T. Rakib and M. A. R. Sarkar, "Design and fabrication of an

autonomous firefighting robot with multisensor fire detection using
PID controller," 2016 5th International Conference on Informatics,
Electronics and Vision (ICIEV), Dhaka, Bangladesh, 2016, pp. 909-
914.

[7] J. H. Hwang, S. Jun, S. H. Kim, D. Cha, K. Jeon and J. Lee, "Novel

fire detection device for robotic fire fighting," ICCAS 2010, Gyeonggi-
do, 2010, pp. 96-100.

[8] L. Celentano, B. Siciliano and L. Villani, "A robotic system for

firefighting in tunnels," IEEE International Safety, Security and
Rescue Rototics, Workshop, 2005., Kobe, 2005, pp. 253-258.

[9] D. J. Pack, R. Avanzato, D. J. Ahlgren and I. M. Verner, "Fire-

fighting mobile robotics and interdisciplinary design-comparative
perspectives," in IEEE Transactions on Education, vol. 47, no. 3, pp.
369-376, Aug. 2004.

[10] M. Sato, H. Torikai and Y. Iwatani, "Flame extinguishment by a

prototype of an aerial extinguisher with an inert gas capsule," The
SICE Annual Conference 2013, Nagoya, Japan, 2013, pp. 2051-2056.

[11] Tong feng, Xu Lufeng and Tong Daoling, "An ultrasonic obstacle
avoidance system for firefighting robot," Proceedings of the 4th World
Congress on Intelligent Control and Automation (Cat. No.02EX527),
2002, pp. 1219-1222 vol.2.

[12] M. Li-xin, S. Dao-nian, C. Min-xuan, and W. Xiao-qin,

“Application of Intelligent PID Control for Robot”, IEEE Conference
on Cybernetics and Intelligent Systems, pp. 455-458, 21-24 Sept.
2008.