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

net/publication/226288294

Forest fire detection system based on ZigBee wireless sensor network

Article  in  Frontiers of Forestry in China · September 2008

CITATIONS READS

105 9,182

4 authors, including:

Junguo Zhang Wenbin li

SEE PROFILE SEE PROFILE

Jiangming Kan
Beijing Forestry University
95 PUBLICATIONS   355 CITATIONS   

SEE PROFILE

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

Motion deblurring for color images of standing trees in natural scene View project

Structural Engineering View project

All content following this page was uploaded by Junguo Zhang on 23 January 2014.

The user has requested enhancement of the downloaded file.

RESEARCH ARTICLE

Junguo ZHANG, Wenbin LI, Ning HAN, Jiangming KAN

Forest fire detection system based on a ZigBee wireless sensor

E Higher Education Press and Springer-Verlag 2008

Abstract Compared with the traditional techniques of several defects. In the first place, this method requires
forest fire detection, a wireless sensor network paradigm many financial and material resources and a trained labor
based on a ZigBee technique was proposed. The proposed force. Second, many problems with fire protection person-
technique is in real time, given the exigencies of forest nel abound, such as carelessness, absence from the post,
fires. The architecture of a wireless sensor network for inability for real-time monitoring and the limited area cov-
forest fire detection is described. The hardware circuitry erage. The scope of application of satellite detection sys-
of the network node is designed based on a CC2430 chip. tems is also restricted by a number of factors, which
The process of data transmission is discussed in detail. reduces its effectiveness in forest fire detection. For
Environmental parameters such as temperature and example, a satellite monitoring system has a long scanning
humidity in the forest region can be monitored in real cycle and the resolution of its saturated pixel dots of images
time. From the information collected by the system, deci- is low. Another problem is cloud layers may mask images
sions for fire fighting or fire prevention can be made more during the scanning period and the real-time mathematical
quickly by the relevant government departments. quantification of fire parameters is very difficult to achieve
(Shu et al., 2005; Yu et al., 2005; Calle et al., 2006). Given
Keywords wireless sensor network, fire detection, net- these shortcomings of traditional monitoring, we suggest
work system, network node, data transmission the ZigBee wireless sensor network technology and explain
its application as a monitoring system. This system can
monitor real-time related parameters, e.g., temperature,
1 Introduction relative humidity, and send the data immediately to the
computer of the monitoring center. The collected data will
Forests are part of the important and indispensable be analyzed and managed by the computer. Compared
resources for human survival and social development that with the normal meteorological information and basic for-
protect the balance of the earth ecology. However, because est resource data, the system can make a quick assessment
of some uncontrolled anthropogenic activities and abnor- of a potential fire danger. The analytical results will then be
mal natural conditions, forest fires occur frequently. These sent to the relevant department as the policy-making basis
fires are among the most serious disasters to forest by which the department will make the decision of fire
resources and the human environment. In recent years, fighting or fire prevention.
the frequency of forest fires has increased considerably
due to climate change, human activities and other factors.
The prevention and monitoring of forest fires has become a 2 Main parameters of forest fire monitor
global concern in forest fire prevention organizations.
Currently, forest fire prevention methods largely consist Three factors compose the basis of a forest fire: the fire
of patrols, observation from watch towers and lately sat- source, environmental elements and combustible mater-
ellite monitoring (Lai, 2004; Huang et al., 2005). Although ial. A forest fire usually occurs as the result of their com-
observation from watch towers is easy and feasible, it has bined effects (Song et al., 2006). According to the Canada
Fire Weather Index Forecast Model, the moisture content
Translated from Journal of Beijing Forestry University, 2007, 29(4):
41–45 [译自: 北京林业大学学报] of the combustible material plays an important role in
forest fires, which means the probability of forest fires
Junguo ZHANG (*), Wenbin LI, Ning HAN, Jiangming KAN
School of Technology, Beijing Forestry University, Beijing 100083, depends on the moisture content (Tian et al., 2006).
China Therefore, the moisture content of combustible materials
E-mail: [email protected] is a major point of assessment and predicts whether a fire
370 Junguo ZHANG, et al.

will take place. The moisture content has much to do with an easy and flexible network configuration, low equip-
relative humidity in the atmosphere, air temperature, ment costs and long-lasting batteries. Thus, it has great
wind and similar factors (Shu et al., 2003; Zhang, 2004). development potential and a promising market applica-
Water evaporation can be directly affected by relative tion in the field of industrial control.
humidity. At the same time, the physical properties of By applying a wireless sensor network based on ZigBee
combustible materials can be changed indirectly by air to a forest fire monitoring system, information such as
temperature. Thus, relative humidity and air temperature temperature and humidity at any part of the forest covered
are regarded as the two main factors which affect the by the network could easily be collected, dealt with and
moisture content of the fuel. Therefore, to reflect the analyzed at any time. In addition, the system can be
moisture content indirectly, these two parameters are extended significantly, the cost of equipment maintenance
the main objects of our investigation, which should pro- could be reduced and the whole system could be optimized.
vide an important basis for the prediction and monitoring
of forest fires. Certainly, forest fires are also caused by 3.1 Forest fire monitoring system based on ZigBee
other factors, such as the active degree of thunder and wireless sensor network
lightning above the forest, human factors, wind speed,
and condition of area vegetation. However, these factors A ZigBee wireless sensor network system includes sensor
will be ignored in our discussion. nodes, gateways (routers) and a monitoring host computer.
To decrease the loss of energy and data packets, a cluster-
tree network topology structure (Tillett et al., 2004) (shown
3 Application of ZigBee wireless sensor as Fig. 1) is applied in this design. Sensor nodes fitted with
network in a forest fire monitoring system microprocessors of low processing capacity are distributed
randomly in the forest and nearby areas to collect fire
A wireless sensor network, which combines computer and monitoring parameters such as relative humidity and
communication technology with the technology of a sen- atmospheric temperature (Zenon and Fady, 2005).
sor network, is considered to be one of the ten emerging Depending on the part the different sensor nodes play in
technologies that will affect the future of human civiliza- the whole network, they are divided into three categories:
tion. This network is composed of numerous and ubiquit- ordinary bottom nodes, cluster heads and network coordi-
ous micro sensor nodes which have the ability to nators. Data collected is transmitted to its own cluster head
communicate and calculate. These nodes can monitor, by an ordinary bottom node. A cluster head mainly han-
sense and collect information of different environments dles data fusion and data packet transmission. Via the
and various monitoring objects cooperatively. cluster head, data collected by ordinary bottom nodes in
ZigBee is a low-rate, low-cost and low-power kind of the cluster can be fused and transmitted to the nearest
short range wireless network communication protocol. network coordinator and data packets transmitted by the
Compared with other wireless technologies, ZigBee has network coordinator can be broadcast to related clusters.
unique advantages of safe and reliable data transmission, A network coordinator mainly deals with basic network

Fig. 1 Structure of a wireless sensor network for forest fire detection based on ZigBee technique
Forest fire detection system based on a ZigBee wireless sensor network 371

management functions such as network configuration, A CC2430 chip, recently launched by the Chipcon
equipment registration and access control. Data informa- Company, is the core chip used in the hardware design
tion can be transmitted to routers by wireless communica- of nodes. It is an SoC CMOS chip, embedded with a high-
tion. When receiving data, routers establish a local performance and low-power microcontroller chip 8051,
database and then transmit the data to the monitoring host an integrated ADC of 14-bit analog-digital conversion
computer via internet, which provides a decision-making and a 2.4 GHz RF wireless transceiver that conforms to
basis for forestry or fire prevention departments. the IEEE802.15.4 standard. This chip has good wireless
receiving sensitivity and excellent anti-interference per-
3.2 Structure of the sensor node formance. When under a receiving and emitting mode,
the loss of current is less than 27 or 25 mA respectively.
The sensor node is a basic unit and platform of the wire- The CC2430 chip requires only a very short time in chan-
less sensor network. A sensor node is commonly com- ging its mode from sleep to an active mode and is espe-
posed of a sensor module, a processing module, a cially suited for forest fire monitoring applications which
wireless communication module and a power module. require long-lasting batteries.
Figure 2 shows the structure of the sensor node. SHT11, an integrated digital sensor for relative humidity
The sensor module is responsible for data analog- and temperature with a I2C bus made by the Swiss
digital conversion and collecting parameters such as rela- Sensirion Company, is used in the sensor node. This sensor
tive humidity of the atmosphere and air temperature. The has some of the most advanced functions such as digital
processing module is responsible for controlling the output, is alignment- and calibration-free, has auto-dorm-
operation of the whole sensor node and saving and coping ancy, and can be completely submerged in water. Its size is
with data collected by its own node and the binary extremely small (7.5 mm 6 5 mm 6 2.5 mm) similar to a
information transmitted from other nodes. The wireless match head. The SHT11 requires a supply voltage of 2.4–
communication module is responsible for communication 5.5 V. The chip also has four free NC pins and a 10 kV
with other nodes and exchanging control information and pull-up resistor, access to the DATA bus and a 100 nF
filter capacitor connected to a VDD and GND.
receiving or transmitting data. The power module supplies
A DS2401 chip is treated as a unique identifier memory
power for the other three modules and drives the nodes,
for hardware nodes. Before use, the unique 64-bit regis-
making it the key factor for the effective operation of the
tration code is marked on a DS2401 ROM, including an 8-
network (Ren et al., 2003).
bit family code, a 48-bit sequence code and an 8-bit CRC.
Any DS2401 chip, produced by the Dallas Company, has
a different sequence code. Except for the grand pin, the
4 System hardware design DS2401 chip has only one function pin which can supply
power and deal with input and output.
The hardware design of network nodes, including sensor The power supply for the system is provided by two
nodes and router nodes, is the basis of the wireless sensor large-capacity, high-energy alkaline batteries. To assure
network structure. a stable working state and maintain working characteris-
tics, a MAX1724EZK30 is used in the nodes to keep the
4.1 Sensor node design working voltage steady.
A TA-XPQ2400-3dBi rubber antenna produced by the
Figure 3 shows the circuit principle of a sensor node. ShiDaiChuangXing Antenna Factory is applied in the

Fig. 2 Structure of a sensor node

Fig. 3 Circuit diagram of a network node

system. The antenna is 5 cm long, weighs 20 g and is easy are achieved, data from nodes can be received and control
to install. Its typical emission frequency is 2.4 GHz. orders can be sent to nodes. A CMS91 module made by
the Cellon Company is used as GPRS module. As the
4.2 Gateway design interface between the gateway and the internet, this
GPRS module deals with sending data to the Internet
The hardware configuration of the gateway (shown as and receiving control information from the Internet.
Fig. 4) includes the main processor, a memory unit, an
RF receiver and transmission module, a GPRS commun-
ication module, and I/O and Ethernet and extension inter- 5 Data transmission process
faces. Considering the large number of control functions
and the large data flow of the gateway, processors with To deliver data transmission inside the ZigBee network in
strong processing ability are required. Since the 8051 this design, a system of active requests for information by
micro-controller embedded in the CC2430 cannot meet the monitoring host computer and passive responses by
these requirements, a PXA255 processor chip of the the sensor nodes is used. Figure 5 shows the data trans-
Intel Company is installed. It is an RISC processor which mission process.
integrates a complicated circuit in one chip with high per- When the monitoring computer, operated by a telecom-
formance and low cost, thus combining powerful func- munications worker, sends an order to inquire about the
tions with the advantages of embedded processors. In state of forest temperatures and humidity, the order is
the RF receiver module, a CC2430 is still used, through transmitted to the router via the internet. The router then
which bi-directional communications with sensor nodes scans the routing tables according to the order and decides
the target coordinator, which then broadcasts in the
attached cluster branch to activate the target cluster head.
The cluster head broadcasts towards its member nodes to
activate the dormant nodes to carry out data communica-
tion. After receiving the data collected and sent by the
nodes, the cluster head integrates and returns the data
to the monitoring host computer along the original line.
If the target network is not found or not connected, the
cluster head will desert the data packet and generate a
Fig. 4 Structure of the gateway system report to the monitoring host computer. Most nodes in
Forest fire detection system based on a ZigBee wireless sensor network 373

Fig. 5 Flowchart of data transmission

the system are in a dormant state to save energy and same time, the chain structure needs to be stable and its
extend the lifetime of the network. scale is limited, which needs to be improved in future
investigations. In other words, we propose this system
as a first attempt and complement to existing forest fire
6 Conclusions monitoring and prevention methods. It provides a solid
basis in terms of hardware for the application of advanced
Wireless sensor networks are increasingly applied in the wireless sensor network technology. To extend the poten-
field of environmental and ecological monitoring. tial of the system and improve forest fire monitoring tech-
Especially in difficult and harsh environments, it has nology, the problems of energy consumption, nodes
advantages that traditional monitor systems lack. In addi- location and clock synchronization need to be addressed
tion, wireless sensor technology has a broad application in the future. These are some of the remaining problem
background in the field of real-time forest fire monitoring. areas to be considered, before the level of forest fire mon-
But given the complexity and peculiar features of the for- itoring can be improved.
est, the system has not been extensively applied in prac-
tical forest fire monitoring. To monitor temperature and
humidity in the forest in a more timely and precise way, we References
pointed out unique advantages of safety in data transmis-
Calle A, Sanz J, Moclan C, Casanova J L, Goldammer J G, Li Z,
sion, flexibility in building the network, and low cost and Quin Z (2006). Detection and monitoring of forest fires in China
energy requirements for a forest fire monitoring system through the ENVISAT-AATSR sensor. In: Proceedings of the
based on a ZigBee wireless sensor technology that we 2005 Dragon Symposium–Dragon Programme Mid-term
designed. The topology structure of the system is an Results. Paris: European Space Agency, 283–291
Huang S L, Florian S (2005). ENVISAT ASAR wide swath back-
adaptation of a cluster-tree. Compared with a reticular scatter dynamics of the Siberial boreal forest fire scar. In:
structure, a cluster-tree structure can be built more easily Proceedings of the 2004 Envisat and ERS Symposium. Paris:
and the information path takes less memory space. At the European Space Agency, 1305–1313
374 Junguo ZHANG, et al.

Lai R W (2004). Studies on the forest resources monitored by 3S Tillett J, Yang S J, Rao R, Sahin F (2004). Optimal topologies for
technologies. Dissertation for the Doctoral Degree. Fuzhou: wireless sensor networks. In: Proceedings of SPIE–the
Fujian Agriculture and Forestry University (in Chinese) International Society for Optical Engineering, Unmanned/
Ren F Y, Huang H N, Lin C (2003). Wireless sensor networks. J Unattended Sensors and Sensor Networks. London: SPIE, 192–
Software, 14(7): 1282–1291 (in Chinese) 203
Shu L F, Zhang X L, Dai X A, Tian X R, Wang M Y (2003). Forest Yu L Y, Wang N, Meng X Q (2005). Real-time forest fire
fire research (II): Fire forecast. World For Res, 16(4): 34–37 (in detection with wireless sensor networks. In: Proceedings of
Chinese) 2005 International Conference on Wireless Communications,
Shu L F, Wang M Y, Zhao F J, Li H, Tian X R (2005). Comparison Networking and Mobile Computing. Wuhan: WCNM, 1214–1217
and application of satellites in forest fire monitoring. World For Zenon C, Fady A (2005). Wireless sensor network based system for
Res, 18(6): 49–53 (in Chinese) fire endangered areas. In: Proceedings of the 3rd International
Song W G, Ma J, Satoh K, Wang J (2006). An analysis of multi- Conference on Information Technology and Applications.
correlation between forest fire risk and weather parameters. Eng Sydney: ICITA, 203–207
Sci, 8(2): 61–66 (in Chinese) Zhang G (2004). Study on forest fire dynamic monitoring in
Tian X R, McRae D J, Zhang Y H (2006). Assessment of forest fire Guangzhou City. Dissertation for the Doctoral Degree.
danger rating systems. World For Res, 19(2): 39–46 (in Chinese) Zhuzhou: Central South Forestry University (in Chinese)

View publication stats