Advanced Materials Research Vols 846-847 (2014) pp 883-887 Online: 2013-11-21

© (2014) Trans Tech Publications, Switzerland

doi:10.4028/www.scientific.net/AMR.846-847.883

Information fusion technology based on wireless fire detection and

alarm system
NIU LINGHUI1,a,HU ZHUGE2,b
1
The Hong Kong Polytechnic University ,China
2
Shanghai Xian Dai Architecture,Engineering&Consulting Co. ,China
a
[email protected],[email protected]

Key words: Fire detection; STM32; ZigBee technology; Information fusion

Abstract. In order to solve the problem, which is "the traditional fire alarm system is only used to
detect a particular physical or chemical signal of the fire, moreover, false alarm and failure alarm
can occur easily.” We apply multi-sensor composite detection technology and wireless
communication technology in the fire detection and alarm system，and design an intelligence
distributed wireless fire detection alarm system, which is mainly based on STM32 control chip.

Introduction
The application of fire played an extremely important role in the development of human
civilization. While the fire always threaten human life and cause great losses of lives and property.
At present, automatic fire detection and alarm system is widely used, and it plays an important role
to protect our safety. But there are many problems in the practical application of fire detection and
alarm system. The quality of existing products in market is uneven. These products only detect a
certain kind of fire parameter. Many problems can be occured, such as false alarm and failure alarm.
With the continuous development of microprocessor technology, sensor technology, communication
technology, control technology and artificial intelligence technology, fire detection and alarm system
has beening developed to be a high reliablity, low false alarm and failure alarm, networking, and
intelligence system[2].
The purpose of this thesis is to propose a fire detection system solution based on multi-sensor
information fusion technology. It can collect a lot of fire parameters with different characteristics,
transfer data by wireless sensor networks, and process and judge data by intelligent algorithms, then
detect and alarm fires timely and accurately.

1 Structural Design
1.1 Design Ideas
1) The multi- sensor composite detection technology
Generally, fire process is divided into four stages, which are early stage, smoldering stage, flame
heat release stage and attenuation stage. The most of release is burning gas during the early stage of
the fire,smoke, flame and heat are very less. It will produce a lot of visible or invisible smoke during
smoldering stage. Flame and heat are not too much. The heat will be radiated out during fire flame
heat release stage. The ambient temperature will rise up rapidly[3]. Fire signals include a number of
characteristic parameters. The more fire parameters are used for system, the higher intelligence level
will be, the lower false alarm and failure alarm will be[4].
2) Distributed Fire Alarm System
Centralized control was used for early system ,the detector is just a absolute sensor, it will
transmit collected signals to controller at all times. The result can be got by processing and judging
those signals by controller. But when the scale of system is become bigger, negative factors will be
produced, such as: the load for controller will be over its ability and the response speed will lower.
Therefore, the distributed control mode will be gradually used instead of the centralized control
method [5].

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans
Tech Publications, www.ttp.net. (ID: 128.210.126.199, Purdue University Libraries, West Lafayette, USA-27/05/15,14:54:28)
884 Advances in Mechatronics, Automation and Applied Information
Technologies

3)Wireless Data Transmission

The energy supply and information transmission for traditional fire detectors was usually realized
by cable. For cable laying during installation, not only the cost will be very high and work load will
be very heavy, but also building structure will be damaged sometimes. Wireless Sensor Network
(Wireless Sensor Network, WSN), will use wireless communication technology and microsensor
with small size and low power consumption. It is composed by a large number of micro sensor nodes
located in the monitoring region. It is a multi-hop self-organizing network system formed by wireless
communication mode[6] [7]. Therefore, it will be a new important research direction of fire detection
by combining the fire detection technology and the wireless communication technology to realize a
wireless，network and intelligent fire detection system.
4) Intelligence Algorithm of Fire Signals
For different burning material and environment detected, the difference dynamic characteristic
parameters value collect by sensors is very big. During fire detection, the traditional thresholds
method for setting alarm delay time and alarm threshold is too simplify. Therefore, the application of
intelligent algorithm in fire detection system has an important positive role in reduce false alarm and
failure alarm for fire alarm system [8].
1.2 Wireless Composite Fire Detection Alarm System
For fire detection alarm system researched in this paper, it respectively designed the fire alarm
controller and composite fire detectors by using STM32 as processor to form fire alarm system
controlled by distributed control mode. The composite detection node can collect smoke
concentration, temperature and CO concentration signals. The processing and judgment of fire
information are realized in detecting node. the information will be transferred to fire alarm
controller through ZigBee wireless network. The storing and display of fire data and management of
detection node are done by fire alarm controller. Figure 1 shows the system's overall structure
diagram.

Figure 1 The overall system structure diagram

(1) Compound detection node: Collecting site environment signal through three kinds of sensors,
the fire information is processed by intelligent algorithm. The fire information is transferred to fire
alarm controller through ZigBee communication module. The alarm device will be started if there is
a fire, fire alarm information will be sent to the fire alarm controller at the same time. If there is no
fire, the verification signal will be sent to fire alarm controller for process at regular time. As Shown
in Figure 2.
(2) Fire alarm controller: Be responsible for receiving information from composite detection
node. Record the date and time of fire by clock chip for future query. Store historical information by
using SD card . Display related detection node information on the LCD in real time. Starting alarm
and actuators when alarm information is detected, it can regularly check each compound detection
node at the same time to determine whether it works properly or not. As shown in Figure 3.
 Advanced Materials Research Vols. 846-847 885

LCD display Button

Alarm circuit

Communication
Temperature Communication Clock circuit interface
sensor interface
Smoke sensor STM32 STM32
JTAG interface Serial port
CO sensor circuit Alarm circuit

Power supply SD storage Power supply

Figure 2 Composite detection node Figure 3 Fire alarm controller

2 Hardware Design for System

2.1 Circuit Design for Microcontroller
The system selected 32-bit processing chip STM32F103RBT6 as the master control chip for
fire alarm controller and composite detection node. STM32F103RBT6 is new 32-bit
microcontrollers created by ST by using advanced architecture ARM Cortex-M3 core. Flexible static
memory controller makes it very convenient to connect with many memories and peripheral
connections, meanwhile because of its wealth on-chip peripherals, the design of peripheral circuits
can be simplified.
2.2 Sensor Selection
LM35DZ is used for temperature sensor in this system., it connects temperature sensor and
amplifying circuit in a silicon slice to form an integrated temperature sensor. the sensitivity is 10 mV
/ ℃. Working temperature range is 0 ℃-100 ℃. Working voltage is 4-30 V. Precision is ±1 ℃. The
maximum linear error is ±0.5 ℃.the static current is 80 µA. Its output voltage is a linear relation with
Celsius temperature scale. conversion formula is as follows：
Vout_LM35(T)=10mV/℃×T℃
The smoke sensor and CO sensor will use common MQ-2 and MQ-7 gas sensors separately.
Those two sensors are metal semiconductor sensor based on stannic dioxide (Sn02). Here we take
MQ- 2 gas sensor as an example to introduce its working principle. MQ-2 gas sensor has different
resistance values for different gas types and concentrations. Power supply for heating end and
measurement input end for MQ-2 gas sensor is 5V DC. the output end V1 can be input to MCU after
the conditioning, the magnitude of V1 is directly related to smoke concentration.
2.3 ZigBee communication module
ZigBee is a close-range, low power consumption, low data transfer rate, low complexity and
low-cost wireless network technology. ZigBee wireless network is based on DSSS spread spectrum
technology. it will use CSMA / CA (Carrier Sense Multiple Access / Collision Avoidance) channel
access method. Communication distance between nodes is from 10 m to100m. The distance can be
reached to one thousand meters with PA module [9]. The wireless communication module of this
article will use Shun Zhou SZ05 series Z- BEE embedded wireless serial interface communication
module. Reinforced ZIGBEE wireless technology will be used for this module . It has many
characteristics and advantages such as long communicating distance, anti-jamming ability, flexible
networking. It can group mesh network structure.
3 Software design for system
Overall, the system software can be divided into two parts: the fire alarm controller software
program and the composite detection node software program. In this system, in order to ensure that
each node works properly, the alarm controller can determine whether each node works properly or
not according to received ID number of each node at regular time.
3.1 The fire algorithm design for composite detection node
The collected temperature, smoke and CO data are converted to actual temperature values, smoke
concentration and CO concentration. Related values are extracted to determine whether there is a fire
or not. The fire is determined based on following six values: temperature (T), smoke value (S), CO
886 Advances in Mechatronics, Automation and Applied Information
Technologies

value (C), △T of temperature rise, △S of smoke rise, △C of CO rise. Prewarning will be started
when temperature, smoke or CO value reaches to threshold value, And then, pay attention to
whether△ T / △ S or △ C reaches to threshold value. If it reaches, the fire can be determined,
meanwhile the alarm will be activated and the information will be sent to fire alarm controller,
otherwise it returns to prewarning condition. [11]. Flow diagram for fire determination is shown in
Figure 4.
3.2 Software design for fire alarm controller
The main program mainly includes communication program for STM32 chip, SD card store
programs, LCD display procedure, etc. After the information from detection node is received by fire
alarm controller, it will be stored in a circular queue and shown on the main interface. When unusual
condition is determined, it will be shown and saved. Alarm circuit will be activated to warn worker.
Software flow diagram for fire alarm controller is shown in figure 5

Start
T/S/C≥ N
threshold?
System
initialization
Y
Start
Fire warning Accept information of composite
detection node

System
ΔT / ΔS /ΔC N
initialization No fire occurred
≥ threshold? If there is a node N
in a fire
Acquisition of Y
signal Show the fire Y
and alarm
Alarm
Signal
processing Through the
Display information
Zigbee module
through the LCD

Store
Fire alarm controller
information

Figure 4 Flow diagram for fire determination Figure 5 Software flow diagram for fire alarm
controller

4 Experiment and Conclusion

In order to verify the reliability for fire detection and communication of this system, we test
composite fire detecting node and fire alarm controller. Parts of environment data detected by this
system is shown in table 1

Table 1 Environment data of composite fire detection node

Species Materials Temperature Smoke CO △T △S △C
output output output

Towel 2.30 2.61 2.17 3.5 0.5 0.6

Flame
Bubble 3.30 2.55 2.71 2.5 1.1 1.6
Plastic 2.65 3.20 3.00 2.7 0.7 1.4
Smoldering fire Paper 0.66 2.85 2.99 1.0 3.4 4.1
Timber 0.68 2.79 2.95 0.8 4.1 4.0
Plywood 0.32 2.57 2.96 0.9 3.8 3.5
Interference Candle 0.36 2.41 2.63 0.4 1.3 0.6
Cigarette 0.14 2.67 2.58 0.2 2.4 1.5
Liquefied 0.11 2.38 2.60 0.0 0.4 0.7
gas
 Advanced Materials Research Vols. 846-847 887

By the data of composite detection node， We can see that lots of heat is released in open
flame stage, and the temperature changes very quickly. During the smoldering stage， it mainly
release gas and smoke, and the temperature slightly changes. In the case of interference, Although
the smoke and CO output value over the threshold value, the change in value is very small. it can be
determined as fire interference.
Experiments show that by using of embedded technology, ZigBee wireless communication
technology and compound detection technology, fire detection and alarm system can be a simple
structure , fire algorithm is easy to realize, transmission of information between the system nodes is
stable and reliable, fire signal can be effectively detectd and fire can be found in time.

Reference
[1] Wu Yuze, Yang xingrong. Technology Development of Fire Alarm System [J]. Fire Fighting
Technology and Product Information, 2004 (1):16-19
[2] Xiao Jing. Development Trend of Fire Detection Alarm Technology[J]. Anhui Fire Fighting,
2007, 36 (16): 103-104.
[3] Du Jianhua, Zhang rencheng. Fire Early Stage Feature Information Research Based on Fourier
Transform Infrared Spectroscopy Detection [J]. Spectroscopy and Spectral Analysis, 2008, 28 (3):
559-563.
[4] RESTAS A. Forest Fire Management at Aggtelek National Park Integrated Vegetation Fire
Management Program from Hungary [C] / /The World Congress on Intelligent Control and
Automation (W CICA), 2008: 5014�5018.
[5] Bai Zhuowen. Application Status and Development Trend of Fire Monitoring System in
Building Automation[J]. Security Technology, 2003, (2): 24-26.
[6] Foreign New Fire Detection Product Introduction [J]. Fire control technology and product
information, 2006 (9): 77 - 78.
[7] Reng Fengyuan, Huang Haining, Lin Chuang. Wireless sensor network [J]. Journal of Software,
2003, 14 (7) 1282-290.
[8] Chen Tao, Yuan Shuihong, Fan Weicheng. Forecast for Fire Detection Technology Research
[J]. Fire Science, 2001, 2 (10): 108 a 110
[9] Lv Zhian. Principles and Application Development of ZigBee Network[M]. Beijing: Beijing
University of Aeronautics and Astronautics Press, 2008.
[10] Zhu Ming. Research on control system of distribution intelligent fire alarm [D].wuhan:
Huazhong University of Science and Technology, 2004:-41.
[11] Liu Xuan, Liu Shixing. Bus Type Smoke and Temperature Composite Fire Detector[J]. Hefei
University of Technology, 2010 (7), and (7): 23 a 24.
Advances in Mechatronics, Automation and Applied Information Technologies
10.4028/www.scientific.net/AMR.846-847

Information Fusion Technology Based on Wireless Fire Detection and Alarm System
10.4028/www.scientific.net/AMR.846-847.883