Development of a Network-based Autonomous Firefighting Robot

Md. Hazrat Ali

, Sultan Shamishev and Aidos Aitmaganbayev

School of Engineering, Nazarbayev University, 010000 Astana, Kazakhstan

Keywords: Automated, Firefighter, Network, Sensor.

Abstract: This paper presents the design and development of a prototype of network-based automated firefighting

robot. It addresses the water-based extinguishing system for fire using the spray gun and pump. For this

purpose, a mini car washer is used. Gas sensors are integrated to create network system to navigate robot to

the target distance. The proper navigation depends on the algorithm of obstacle avoidance. Currently, fire

extinguish become challenging especially in a multi-storey building. This work gives a solution to

extinguish fire automatically to prevent danger in a residential premise. The prototype is built from the low-

cost materials available at the laboratory.

1 INTRODUCTION

Today, the rapid development of science and

technology makes it possible to apply new technical

approaches, and fire extinguishing means in

extinguishing fires, of varying complexity. Among

all the possible varieties of novelties in the field of

firefighting, it is necessary to single out fire robots

in a separate item. Fire robots are the technical

means that can radically distinguish fire. A fire robot

is a sophisticated technical device designed to

extinguish a fire. Depending on the ability to move,

there are stationary and mobile robots. Fire robots

are used in a situation where people cannot enter

near the fire as there is a direct danger to their health

and life. The use of fire robots can significantly

facilitate the work on extinguishing fires at various

industrial facilities, transport, and sometimes even

fire robots are irreplaceable, for example, fires on

radioactive, chemically hazardous objects. The focus

of this research is the design, assembly, and

development of a model of an intelligent fire robot

that can switch to a fire source offline and extinguish

it. This work project includes the electrical,

mechanical and computer programming. The sensor

network conducts the fire detection. The network of

the gas sensors divides the room into sections and

detects the presence of fire in them. By the wifi

communication between the network and robot, the

destination is calculated and navigates the robot

accordingly. By implementation of Bayesian

Algorithm and using three ultrasonic sensors, the

robot reaches the fire. After reaching the fire, the

robot starts the pump and water splashes toward the

fire, and as a result, fire is extinguished.

There are cases when the fire cannot be extinguished

by the conventional ways: the difficulty of reaching

by firefighters, the places where the conventional

splinker systems cannot be implemented. All these

generate the need for a new method of fire

extinguishing. The fire extinguishing contains

several ways to be performed; the work is dedicated

to decreasing the temperature of the fuel. To perform

that, the development of the autonomous firefighting

robot is crucial. The robot prototype detects the

presence of fire, and autonomously extinguish it.

The research focuses on existing robots and the

methods of elimination of fire.

There are many ideas for creating mobile robots that

can work, collect data and information, and look for

victims in a dangerous environment for humans.

These ideas are continuously enhanced and

upgraded. To date, there are models of robots that

are effectively used to rescue and evacuate people

caught in extreme conditions, such as a burning

building or environment of toxic gases. Robots are

used to extinguish the flames, both autonomous and

remote controlled. Controlled robots are often

equipped with a monitoring system and work

through a wireless communication system, while

autonomous robots are equipped with obstacles

avoidance system embedded into its autonomous

navigation system. Remote controlled robots are

equipped with microphones and acoustic systems to

Ali, M., Shamishev, S. and Aitmaganbayev, A.

Development of a Network-based Autonomous Fireﬁghting Robot.

DOI: 10.5220/0006928305250533

In Proceedings of the 15th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2018) - Volume 2, pages 525-533

ISBN: 978-989-758-321-6

525

communicate victims with the operator during

emergency situations. Also, these robots are often

equipped with cameras to capture the site of the fire

and sensors for measuring the temperature and

concentration of CO2 and O2 (Kim, 2009). It allows

data transfer, such as video and audio, and also

allows us to track the position of the robot itself.

Since both autonomous and controlled robots are

designed for extreme places, they usually withstand

high temperatures, are waterproof and have a shock

resistance function.

The shape and design of robots also have a wide

variety. Often, crawler mechanisms are used for

movement as they consist of metals and are steady to

fire, but non-standard types of robots are also found.

For example, the rescue robot BEAR (Battlefield

Extraction-Assist Robot) from Vecnarobotics is

created in the form of a humanoid. This type of

robot is used for prospecting since it can raise and

carry victims, but its cost is not comparable to its use

(Vecna Robotics Co., 2012).

In most of the cases, the movement of fire robots is

due to sensor systems or vision-based systems. A

sensor-based system is monitored by feedback from

various sensors, while a vision-based system uses

cameras and the image processing techniques to find

the target position. However, there are also robots

that use other systems for movement. For example,

Tehzeeb is a robot that uses a laser scanner module,

a manipulator and map generation algorithms for

localization and navigation (Kim, 2009). This

system is more effective than vision-based systems

in conditions of poor visibility, for example, in

dense smoke. The flame sensors detect the flame.

The analog output from the flame sensors was fed

into the input terminal of the analog converter

(ADC) with the microcontroller. A minimum value

was set based on the desired sensitivity, and its

gradual increase means that the robot moves towards

the source of the fire.

Obstacles avoidance system is also an important

measure that provides free movement without any

collisions during the navigation process in robot

creation. For this purpose, the robotic technology

uses ultrasonic sensors or range sensors (Harwayne-

Gidansky, 2007). In this project, this task was

realized with the help of ultrasonic sensors. The

obstacle avoidance module is implemented by

converting the analog output from an ultrasonic

sensor to a digital form (Scott Dearie, 2004). Since

the angle to which these sensors operate is limited,

this project uses three sensors that are located at a

certain angle to capture a sufficient distance for the

robot to travel. After the flame sensor detects the

source of the fire, the robot moves to the required

distance and the fire extinguishing system is

activated. For the test, a candle is used as a source of

the fire.

2 DESIGN AND DEVELOPMENT

2.1 Conceptual Design

The concept of the model is based on the primary

objective which is to develop an autonomous

firefighter robot by considering the following

criteria:

(a) Autonomous fire detection

(b) Navigate robot automatically to the fire

detected by the network-sensor

The network gas sensor is used to meet the first

criteria. The network is designed by using gas

sensors placed at different locations in a room and

connected to one Arduino board. The Arduino board

is connected to the laptop and Bluetooth module of

the robot.

The robot has to move towards the fire automatically

to meet the second criterion. In the accomplishment

of this task, the problem is in identifying the

destination toward which the robot should move,

and avoiding the obstacles towards the destination. It

is done by the Bluetooth module connected to the

Arduino board which receives the location of fire by

the sensor network. The obstacles are avoided by the

ultrasonic sensor which detects the object and avoids

it and finds the way towards the flame. The third

criterion is to extinguish fire autonomously. The

mini car washer is used to meet this criterion. The

car washer pump located in the water tank turns on

when the robot reaches the destination. It leads to

spray the water by the pistol of the car washer.

2.2 Microcontroller Selection

The Arduino Uno is the microcontroller, which

based on ATMega328P and is widely used in

robotics. The Arduino Uno board has a dimension of

length 68.6 mm and width 53.4 mm. The operating

voltage is 5V, and an input voltage varies from 7V

to 20V, for our purpose this value corresponds to

12V. The number of digital input and output pins of

14 is sufficient for this research. One Arduino board

ICINCO 2018 - 15th International Conference on Informatics in Control, Automation and Robotics

526

is used to communicate two DC motors by H-Bride,

H-bridge also controls three Ultrasonic sensors,

Bluetooth module, and pump motor. The second

Arduino Uno board is used to communicate with the

three gas sensors in the network systems with

Bluetooth. The inputs of each gas sensors translated

to specific outputs in Bluetooth. The Bluetooth

module establishes the wireless communication

between these two board module.

2.3 Base Platform Design

Figure 1: The assembly of the base platform.

Figure 1 shows the base of the firefighter robot. The

lower basement part has a dimension of 0.5m x 0.5m

and cut-outs of 0.1m x 0.05m. The cut-outs are made

to adjust the wheels insides of the robot. The upper

part has a dimension of 0.5m x 0.5m and cut-outs of

0.1m x 0.05m and a circular cut in the middle with a

diameter of 0.27m. The circular cut is made for the

purpose to set up a water tank. Figure 2 and 3 show

handles connected to the pistol as well as the base

connected to the handles and pistol.

Figure 2: The connection of handlers and pistol.

Figure 3: The base prototype.

2.4 Motor Selection

Motor’s key parameters are selected based on the

following calculations.

2.4.1 Required Speed

The speed of the robot is assumed to be 150 cm/s.

This assumption is made to find the RPM, which is

provided by the DC motor. The speed conversion

from cm/s to RPM is done in the following way:

V = r * ω (1)

Where V=velocity, r=radius of the tire, and

ω=angular velocity.

By substituting the values, ω=V/r, ω=(150cm/s)/5cm

ω=30rad/s , RPM=ω/2π (2)

RPM=(30rad/s)/2π*60

RPM=286.47rpm.

A minimum 286.47 RPM is needed to meet the

desired 150cm/s speed.

2.4.2 Required Torque

The required torque is the next property that needs to

be taken into account while selecting the DC motor.

The torque produced by the DC motor should be

sufficient to carry the total weight of the robot,

including the tank with the water. First, the force is

calculated as follows:

F = μ

* N (3)

Where, F=Force, μ

=Rotational friction co-efficient,

N=normal force. By substituting the values of the

rotational friction coefficient and normal force for an

assumed total weight of 27 kg, we get:

F = 0.015*(27kg*9.8m/s^2)

F = 3.969N

The equation of the torque:

T = D* F (4)

Development of a Network-based Autonomous Fireﬁghting Robot

527

Where, D=distance (in this case D is r=radius of the

tire), T = 5cm* 3.969N, T = 19.845 cmN, The

selected DC motor needs to provide at least 286.47

rpm and a torque of 19.845 cmN.

3 ALGORITHM DEVELOPMENT

3.1 Obstacle Avoidance Algorithms

In the productive activity and life of the human

society, automation becomes increasingly essential.

In recent years, due to the significant decrease in

resources, global climate change, and population

growth, the role of automation is amplified, and it is

being expanded. To date, there are practically no

technical devices without automation elements -

from simple devices used in everyday life to

sophisticated industrial installations. Modern

humanity uses various devices, without which today

it is impossible to imagine a comfortable and safe

daily life.

For an autonomous vehicle, the ability to detect and

avoid obstacles in real time is a significant

embodiment and a fundamental guarantee of a

ground mobile robot for performing various tasks.

Consequently, there are some studies and solutions

to this problem, but most of these solutions require a

significant computational load and are complicated,

some even completely impossible. In general, there

are two types of avoidance technology: the first is

based on the global map, and the other is based on

sensors. In this work, two algorithms for preventing

obstacles in the implementation of low-cost control

structures based on microcontrollers and sensors,

namely, Bug1 and Bug2 are used.

Bug1 and Bug2 are motion planning algorithms that

are applied to mobile robots. For these algorithms,

sensors of the sonar range are used as sensing

elements. Mobile robots create new traffic planning

in those cases when they meet an unknown obstacle

while moving towards the goal. Bug's algorithms

create a path for moving the robot to the target in a

straight line if the path to the target is clear;

otherwise, the robot follows the obstacle border with

which it encounters (Lumelsky, 1990). These

algorithms are based on three assumptions about the

mobile robot: i) the robot is a point, ii) it has an ideal

localization, and iii) its sensors are accurate (James,

2007).

Bug1 is an algorithm by which a mobile robot

moves directly to the target unless it encounters an

obstacle. In the event of a collision with an obstacle,

the robot examines the outer lines of the obstacle

until it moves to the target is available

(Sankaranarayanan, 1990). As soon as it encounters

an obstacle, it bypasses the obstacle on one side, and

then determines the breakpoint, calculating the

distance between the current position and the

position of the target G while traveling around the

object. The remaining point is the nearest point

around the obstacle to the target. The robot then

determines the shortest path to this nearest point to

reach the breakpoint, and changes or maintains its

direction of the wall, following the shortest path to

return to the breakpoint. Then the robot goes to the

breakpoint and sends an obstacle in the direction G

along the new line. When it encounters the second

obstacle, the same procedure applies. This method is

inefficient but ensures that the mobile robot can

reach any possible goal point (Isabel Ribeiro, 2005).

Figure 4 shows the Bug algorithm.

Figure 4: Illustration of the Bug algorithm.

According to the algorithm Bug2 as shown in Fig. 5,

the mobile robot follows a constant slope, calculated

initially between the positions S and G. The mobile

robot maintains its motion to G if the path on the

slope is not interrupted by the obstacle. If the robot

is in front of it, the robot follows the edges of the

obstacle, using its sonar sensors in a clockwise

direction until it finds its original inclination again.

This property leads to shorter paths than Bug1, but

in some cases the path may take longer than Bug1,

for example, searching for a labyrinth (Magid,

2004).

Figure 5: Illustration of the Bug2 algorithm.

ICINCO 2018 - 15th International Conference on Informatics in Control, Automation and Robotics

528

3.2 Movement

One of the primary requirements of the robot is the

movement. The following components provide the

movement of the robot:

(a) Wheels (2 of the conventional and 1 of

omnidirectional)

(b) DC motors (Two 12V DC motor)

motor, as well as the connection of DC

motor and basement

The following Fig. 6 illustrates the positions of

wheels in black color (2 conventional and 1

omnidirectional wheels). All the three wheels

provide the forward movement. The turning is

resulted by the stop of one back conventional wheel

and rotation of another. The omnidirectional wheel

provides the turning of the front side, as it can turn

in any direction and provide movement.

Figure 6: Positions of the wheels.

Two standard wheels are connected to a DC motor

by the metallic part by welding. The easiest and

reliable way of connection is welding. It leads to

simultaneous rotation of the wheels and the DC

motor’s rotational output. Each DC motor with the

conventional wheel is connected to the basement by

the two metal strips by screwing them into the lower

basement. The omnidirectional wheel is connected

to the lower basement by four screws. The following

Fig. 7 illustrates those as mentioned earlier (3)

connections.

Figure 7: The connection of (1) wheel, (2) DC motor and

metal two strips.

3.3 Navigation

The navigation is done in the following way:

(a) The gas sensor gives the location of the fire

(b) Wi-Fi module transfers the destination to

the robot

avoiding the obstacles using ultrasonic

sensors. Figure 8 explains the location of

the ultrasonic sensors.

Figure 8: Ultrasonic sensors positioning.

3.4 Gas Sensor Network

The gas sensors are located in the room forming a

network to detect the smoke. The MQ-9 gas sensor

is used as a smoke detector. The amount of smoke is

directly proportional to the voltage output. The

supply voltage used in MQ-9 is 5V. Figure 9 shows

the gas sensor.

Development of a Network-based Autonomous Fireﬁghting Robot

529

Figure 9: MQ-9 gas sensor.

As the network the room is divided into three

divisions, the following positioning is determined.

The network follows the following logic: when the

fire is detected at the location of the first sensor, the

destination is to be at first 150 x 100cm division of

the room; if the second sensor detects the smoke the

location is to be a second division, and the third for

the third division.

3.5 Obstacle Avoidance

When the robot moves toward the fire, detection of

the obstacles on the path

should be considered. In

the literature review, the logic of the obstacle

avoidance is suggested. As the way to avoid an

obstacle, the Ultrasonic sensor SRF05 is used. The

operating voltage of the sensor is 5V, and the

measurement angle is 30 degree.

Figure 10: SRF05 ultrasonic sensor.

Figure 10 shows the ultrasonic sensor used for this

research.

4 ASSEMBLY AND TEST

RESULTS

4.1 Fire Extinguishing

Water is used to extinguish the fire in this research.

The following components and technique are

incorporated into the system.

4.1.1 Gun and Water Pump

The last action for the robot is to extinguish fire

autonomously. The fire is extinguished by the

elimination of heat from the fuel, one of the

consistent of the fire triangle, by decreasing its

temperature by adding water into it. The water

stored in the water tank needs to be splashed into the

fire. For this purpose, the "minicar washer" is used.

The mini car washer system consists of the

following components.

(a) Pump

(b) Water Hose

(d) Waterproof power wire

Figure 11: Mini car washer.

Figure 12: Isometric view of the prototype.

The total weight of the "minicar washer" system is

800g. The operating voltage is 12V. Figure 11

illustrates the "minicar washer". Figure 12

highlights the isometric view of the developed

prototype.

ICINCO 2018 - 15th International Conference on Informatics in Control, Automation and Robotics

530

4.2 Network Room Set up

The network room is built in the laboratory by

dividing the 1m x 3m area into three 1m x1m

sections. Then, three gas sensors are placed above

each section and connected to an Arduino Uno. The

Wi-Fi module connected to this Arduino sends the

reading of each sensor to the robot. Figures 13 and

14 show the sensor networking.

Figure 13: Network room.

Figure 14: Gas sensor network.

4.3 Prototype Development and

Testing

Figures 15 and 16 show the front and rear view of

the robot. Figures 17 and 18 demonstrate the robot's

action during fire detection. The testing started with

the adjustment of the network room sensors. The

height for placement of each sensor and reading

range of each sensor configured simultaneously to

cover the sections. The robot’s navigation is done by

changing the Arduino code according to the path of

each sector. Then, by implementing Bayesian

Algorithm, the code is modified based on the

presence of an obstacle. The testing is done using a

candle as a source of the fire. The candle is placed in

each section and at different places with respect to

the sensor location.

Figure 15: Front view of the robot.

Figure 16: Rear view of the robot.

Figure 17: Robot in action.

Development of a Network-based Autonomous Fireﬁghting Robot

531

Figure 18: Side view of the robot during fire extinguish.

5 CONCLUSIONS

The focus of this research was the design, assembly,

and development of an intelligent firefighter robot

that would be able to extinguish a fire source

automatically. This work includes electrical,

mechanical and mechatronics systems. The sensor

network is needed for the fire detection. The

network of the gas sensors divide the room into three

sections and then detect the presence of fire within

the range. By the Wi-Fi communication between the

network and robot, the destination signal is

transmitted to the robot. By the implementation of

Bayesian Algorithm and using three ultrasonic

sensors to avoid an obstacle, the robot reaches the

fire. After reaching the fire, the robot starts the pump

and water splashes toward the fire, and as a result,

fire is extinguished. The test results were obtained

after several trials to reach the desired destination.

Overall, the primary objective of the research; such

as autonomous fire detection, robot navigation, and

autonomous fire extinguishing is met.

REFERENCES

Moore, R., Lopes, J., 1999. Paper templates. In

TEMPLATE’06, 1st International Conference on

Template Production. SCITEPRESS.

Smith, J., 1998. The book, The publishing company.

London, 2

edition.

B. Li “Analysis of the Fire Fighting Robot”. 2011 “Child

strength first”. 2000 Product Safety and Testing

Group, Institute for Occupational Ergonomics and

Division of Manufacturing Engineering and

Operations Management, University of Nottingham,

University Park, Nottingham.

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. doi:

10.1109/ICPCI.2012.6486328

Evgeni Magid and Ehud Rivlin, CAUTIOUSBUG:

ACompetitive Algorithm for Sensory-Based Robot

Navigation, Proceedings of 2004 IEEE/RSJ

International Conference on Intelligent Robots and

Systems, Sendal - Japan, September 28 – October 2,

2004.

“Green seal report”, 2003. Green Seal Inc. url:

https://www.sandiego.gov/sites/default/files/legacy/en

vironmental-services/pdf/ep3/cgr_tires_rr.pdf.

James Ng and Thomas Bräunl, Performance Comparison

of Bug Navigation Algorithms, J Intell Robot Syst,

Springer Science 50:73–84 DOI 10.1007/s10846-007-

9157-6, 2007.

Jared Harwayne-Gidansky; Michael Sudano; “A Low-

Complexity Navigation Algorithm for a Scalable

Autonomous Firefighting Vehicle”, Research and

Development, the 5th Student Conference-SCOReD

2007, 11-12 December 2007.

M. Engin, "Embedded and real-time system design: A case

study firefighting robot," 2016 5th Mediterranean

Conference on Embedded Computing (MECO), Bar,

2016, pp. 18-21. doi: 10.1109/MECO.2016.7525712

Maria Isabel Ribeiro, Obstacle Avoidance, 2005.

M. K. Rangan, S. M. Rakesh, G. S. P. Sandeep and C. S.

Suttur, "A computer vision based approach for

detection of fire and direction control for enhanced

operation of firefighting robot," 2013 International

Conference on Control, Automation, Robotics and

Embedded Systems (CARE), Jabalpur, 2013, pp. 1-6.

doi: 10.1109/CARE.2013.6733740

S. C. Purbarani, Q. A'yunina MOA, G. Jati, M. A.

Ma'sum, H. A. Wisesa and W. Jatmiko,

"Implementation of grid mapped robot planning

algorithm in a continuous map for firefighting robot,"

2015 International Symposium on Micro-

NanoMechatronics and Human Science (MHS),

Nagoya, 2015, pp. 1-7. doi: 10.1109/MHS.2015.74

38292

Sankaranarayanan, A., Vidyasagar, M., A new path

planning algorithm for moving a point object amidst

unknown obstacles in a plane, Proc. of the IEEE Int.

Conf. Robot. Autom. 3, 1930–1936, 1990.

Scott Dearie; Kevin Fisher; Brian Rajala; Steven Wasson;

“Design and Construction of a Fully Autonomous Fire

Fighting Robot”, 2004 IEEE, Pages: 303-310.

Tawfiqur Rakib, M. A. Rashid Sarkar, "Design and

fabrication of an autonomous firefighting robot with

multisensor fire detection using PID controller", 2016

International Conference on Informatics, Electronics

and Vision (ICIEV), vol. 00, no. , pp. 909-914, 2016,

doi:10.1109/ICIEV.2016.7760132

“Ultrasonic ranging module: HC-SR04" URL:

http://www.micropik.com/PDF/HCSR04.pdf

V. Lumelsky and T. Skewis, Incorporating range sensing

in the robot navigation function, IEEE Transactions on

ICINCO 2018 - 15th International Conference on Informatics in Control, Automation and Robotics

532

Systems Man and Cybernetics, vol. 20, pp. 1058 –

1068, 1990.

Vecna Robotics Co.; “Battlefield Extraction-Assist Robot

(BEAR)”, Available: http://www.vecnarobotics.com/

robotics/product-services/bear-robot/ [Accessed: June

27, 2012]

Young-Duk Kim; Yoon-Gu Kim; Seung-Hyun Lee;Jeong-

Ho Kang; Jinung An; “Portable Fire Evacuation Guide

Robot System”, Intelligent Robots and Systems,

IEEE/RSJ International Conference, 11-15 October

2009. Pages: 2789-2794.

Development of a Network-based Autonomous Fireﬁghting Robot

533