A Water-saving Irrigation System Based on Fuzzy

Control Technology and Wireless Sensor Network

Peng Xiaohong
College of Information Technology
Guangdong Ocean University
Zhanjiang, P.R China
e-mail: [email protected]
Xiao Laisheng
College of Information Technology
Guangdong Ocean University
Zhanjiang, P.R China
e-mail: [email protected]
Mo Zhi
Technology Department
Zhanjiang New Zhongmei Chemical Industries Co.,Ltd.
Zhanjiang, P.R China
e-mail: [email protected]
Liu Guodong
Maoming Branch
China Construction Bank
Maoming, P.R China
e-mail: [email protected]


AbstractA scheme is promoted which builds a water-saving
irrigation system based on wireless sensor network and fuzzy
control technology. we design a wireless sensor network, which
consists of sensor node cluster, coordinator node and irrigation
controller nodes Here sensor node cluster is responsible for
gathering information such as soil moisture and regularly send it
to the coordinator node. Fuzzy controller embedded in the
coordinator node takes soil moisture error and error change rate
as its input and obtained water demand amount of crops under
certain soil moisture through fuzzy inference and fuzzy judge and
output it to irrigation controller node. Irrigation controller node
controls the implementation of automatic watering. In this
article, we describe the implementation scheme of sensor network
node and fuzzy controller in detail and design communication
protocol for sensor nodes. The experimental results show that the
system can quickly and accurately calculate water demand
amount of crops, which can provide a scientific basis for water-
saving irrigation. In this way, we have made an exploratory study
in applying wireless sensor networks and fuzzy control
technology to fine agriculture project.
Keywords-irrigation; fuzzy control technology; wireless sensor
network
I. INTRODUCTION
The implementation of precise control irrigation for crop
water demand information is one of the important ways to
improve the utilization of water. There is a large amount of
information of water demand of crops that should be collected
and it is more convenient to realize long-distance data
transmission by means of networks, Wireless sensor network
technology has a broad application in many areas because of its
advantages such as safe and reliable data transfer, simple and
flexible network, low-cost equipment, long battery life, etc.
[1,2] Therefore, nowadays wireless sensor networks are widely
used in crop irrigation monitoring[3,4].
Because of the non-linear feature of soil moisture sensors,
as well as its larger output delay, it is difficult to obtain
satisfactory results to use traditional method of feedback
control. Fuzzy control does not need to set up a precise
mathematical model, only relies on people's experience and
knowledge, and can imitate peoples thinking and experience to
control multi-parameter complex system and make decision-
making. Therefore, it shows a broad prospect in application of
intelligent water-saving irrigation [5]. In 1996, Zhang etc.
developed irrigation controller based on fuzzy control for
greenhouse ornamental plants. This system uses a resistance-
type soil moisture sensor as a detecting device and sensing
devices is used to measure soil water leakage rate. This fuzzy
controller was successfully running in the laboratory and
greenhouse for a few months [6].
From a comprehensive analysis about water-saving
irrigation research at home and abroad, we can see that, up to
now, there is seldom control scheme of water-saving irrigation
that use wireless sensor technology combined with fuzzy
control and the research on it is not perfect enough in theory
and application. Therefore, on the basis of research results at
home and abroad, according to the specific characteristics of
citrus seedlings watering in Guangdong province, combine
wireless sensor networks with fuzzy control technology ,we
have set up a low-cost, practical and intelligent water-saving
irrigation system .we design a wireless sensor network, which
consists of sensor node cluster, coordinator node and irrigation
controller nodes[7]. Here, wireless sensor network is
responsible for collection of information, such as positioning
and soil moisture, and completes wireless real-time
transmission of information of water demand of crops.
Meanwhile, we use fuzzy control technology in system control
scheme, we carry out intellectualized process to control
watering time and watering amount.
978-1-4244-3693-4/09/$25.00 2009 IEEE
Authorized licensed use limited to: SASTRA. Downloaded on March 01,2010 at 05:43:24 EST from IEEE Xplore. Restrictions apply.
II. DESIGN OF SYSTEM STRUCTURE
There are four parts in the system. They are sensor node
cluster, coordinator node and irrigation controller and irrigation
pipe network. Wireless sensor network consists of sensor node
cluster, coordinator node and controller node. Irrigation pipe
network is laid over the irrigated areas and electric control
valves are installed on pipelines. Deployment of humidity
sensor node is to form a sensor node cluster according to the
plant conditions and watering status of citrus seedlings. Each
node is responsible for monitoring a small area of soil moisture
conditions, real-time obtains soil moisture information, and
sends the information to coordinator node with a certain time
interval. When coordinator node receives the information of
soil moisture sent by humidity sensor node, the fuzzy controller
embedded in the coordinator node carries out fuzzy inference
and fuzzy decision to soil moisture information in order to
decide whether or not to conduct water and how long irrigation
time is. Coordinator node then sends the irrigation information
to irrigation controller node. Irrigation controller controls pump
valve at corresponding region to open or turn off. Thus a
closed-loop irrigation network control system is formed and it
can implement water-saving irrigation to crops. Its principle of
system structure is shown in Fig. 1.
III. IMPLEMENTATION OF SYSTEM HARDWARE
Sensor node in wireless sensor network periodically
collects information of soil moisture and sends it together with
the node location information to the coordinator node.
Coordinator connects to irrigation controller node through the
serial port and uploads data. Coordinator node and sensor node
have the same hardware, only different on software
configuration. The followings are the hardware designs for
these two types of nodes.
A. Sensor Node and Coordinator Node
Sensor node and coordinator node consist of data
acquisition module, data processing module, turn-off serial
transceiver and power supply modules, as shown in Fig. 2.
Data acquisition module uses digital humidity sensor SHT11,
which is responsible for collecting the information of soil
moisture on surveillance region. Data-processing module, we
select Chipcon's SoC chip CC2431 It is responsible for
controlling the operation of the entire sensor node, including
storage and processing data collected by itself, implementation
of wireless data transceiver. Single-chip microcomputer in
CC2431 communicates with sensor by its figure port. Self-
shutdown serial transceiver SP3223 is a transceiver that can
automatically work or shut down according to the state of serial
cable connection. When there is no serial cable connection or
without data communications, the chip is in shutdown state and
its consumption of current is only 1A. Coordinator node
connects to the irrigation controller through RS232 serial
interface in chip SP3223. Energy supply modules, CR2032 is
selected, is responsible for energy supply of entire node.
B. Irrigation Controller Node
Irrigation controller takes low-power ARM processor
S3C44B0X that is based on the ARM7TDMI-S core as its main
part, which is a cost-effective microcontroller solution scheme
designed for Samsung's handheld devices and general
applications. In the system, a lot of modules are used, such as
an 8-way 10-bit A/D converter module in the S3C44B0X, LCD
controller, 32-bit timers, UART, GPIO, PWM output module,
etc. Irrigation controller receives the information transmitted
from the serial port of coordinator and sends corresponding
irrigation command through I/O port to control the action of
electric control valve. Optocoupler isolation is set between
general-purpose I/O ports in microprocessor and control valves,
each I/O port corresponding to an electric control valve.
S3C44B0X has as many as 71 multi-function I/O ports, so the
system has a high expansibility. The composition of irrigation
controller node is shown in Fig. 3.
C. Communication Protocol
When designing a wireless sensor network, the first thing
needed to consider is how to get a low power consumption and
how to prolong the network life cycle as long as possible. In
this paper, IEEE802.15.4 standard and ZigBee wireless
network technology are introduced to design the wireless
sensor network. IEEE802.15.4 standard defines the physical
layer and MAC layer. ZigBee defines the network layer
protocol [8]. ZigBee is a close range, low-complexity, low
power, low data rate, low-cost two-way wireless
communications technology that is suitable for the system [9].
Because the primary energy consumption is at idle listening,
receiving unnecessary data and retransmission of collisions and
so on, combining the characteristics of self-localization and
RTC wake-up call in CC2431, in order to reduce power
consumption the network uses non-beacon access mode,
namely ALOHA CSMA/CA channel access mechanism.
Sensor node dialogues with coordination node only when
required for data transmission and it is in sleep the rest of the
time in order to save the energy consumption to maintain the
network connection. Since each sensor node needs to transmit
rare amount of information and the time of dormant state is
larger than its working hours, collision of information is almost
unlikely.

Figure 1. Principle of system structure

Figure 2. Structure of sensor node
Authorized licensed use limited to: SASTRA. Downloaded on March 01,2010 at 05:43:24 EST from IEEE Xplore. Restrictions apply.

Figure 3. Composition of irrigation controller node
IV. DESIGN OF IRRIGATION FUZZY CONTROLLER
For the fuzzy controller embedded in coordinator node, its
input variables are E and E, its output variable is T, where E
is the error between the actual soil water amount and the soil
water amount which is most suitable for citrus seedlings
normally to grow, E is error change rate calculated from
current error and the previous one, T is the length of time for
irrigation. According water demand of crops at different
growth stages, combined with agricultural experts outcome of
long-term research and practical experience of the operators,
fuzzy control rule table can be summed up. Corresponding
control decision-making can be derived from control rules.
Then the corresponding control amount can be obtained due to
the principle of maximum membership degree. Through offline
and repeated calculation and debugging by computer, finally,
we can get fuzzy control query table for practical application
and store it in the coordinator node. When the system runs, the
first step is to calculate E and E according to the input data,
and quantify them to calculate the universes of discourse of
fuzzy amount. Then by searching fuzzy control table, we can
achieve the current water demand information of crops at any
area of irrigation.
A. Fuzzy Process of Input and Output Variables
In this paper, we set soil moisture error E between [-0.8, 0],
error change rate E between [-0.6, 0.6], irrigation time T
between [0, 15 min].
The universes of discourse of input variable E are {-3, -2, -
1, 0, 1, 2, 3}. Its fuzzy language values are {NL (very dry), NM
(dry), NS (relatively dry), ZO (moderate), PS (relatively wet),
PM (wet) and PB (very wet)}. The universes of discourse of
E are {-3, -2, -1, 0, 1, 2, 3}. Its fuzzy language values are
{NL (serious loss of water), NM (relatively serious loss of
water), NS (no serious loss of water), ZO (no change of water),
PS (little increase of water), PM (relatively more increase of
water) and PB(more increase of water)}. Membership function
of E and E is shown in Fig. 4.
The universes of discourse of T are {0, 5, 10, 15}. Its fuzzy
language values are {ZO (no watering), PS (watering for a
short time), PM (watering for a relatively long time) and PL
(watering for a long time}. Its membership function is shown in
Fig. 5.
B. Fuzzy Rule Table
In general, control rules of fuzzy controller can be summed
up from experts professional knowledge and operators
experience. According to water requirement of crops at
different growth stages, combined with agricultural experts
outcome of long-term research and on-site operator's practical
experience, we set up fuzzy control rule table which is shown
in TABLE I.
C. Foundation of Fuzzy Control Query Table
Two-input & single-output language control strategy shown
in TABLE 1 is composed of 49 fuzzy conditional statements.
We can obtain the precise amount of T through removing its
fuzzy value and according to the largest membership law and
calculation of its fuzzy relation matrix[10]. So we can get fuzzy
control query table which is shown in TABLE II.

Figure 4. Membership degree of input var. E & E
0 5 1 0 1 5
Z O P S P M P L

Figure 5. Membership degree of output var. T
TABLE I. FUZZY CONTROL RULE TABLE
NL NM NS ZO PS PL PM
E
NL
NM
NS
ZO
PS
PM
PL ZO ZO ZO ZO ZO ZO ZO
PS ZO ZO ZO ZO ZO ZO
PM PS
PL PM
PL PL
PL PL
PL PL PL PM PS PS ZO
PM PM PS ZO ZO
PM PS ZO ZO ZO
ZO ZO
ZO ZO ZO ZO
ZO PS ZO
ZO
E

TABLE II. FUZZY CONTROL QUERY TABLE
-3 -2 -1 0 1 3 2
E
-3
-2
-1
0
1
2
3 0 0 0 0 0 0 0
5 0 0 0 0 0 0
10 5
15 10
15 15
15 15
15 15 15 10 5 5 0
10 10 5 0 0
10 5 0 0 0
0 0
0 0 0 0
0 5 0
0
E

Authorized licensed use limited to: SASTRA. Downloaded on March 01,2010 at 05:43:24 EST from IEEE Xplore. Restrictions apply.
V. DESIGN OF SYSTEM SOFTWARE
Sensor node separately collects soil temperature at each
sampling cycle and transmits data to the coordinator node. The
fuzzy controller embedded in the coordinator node carries out
fuzzy reasoning and fuzzy decision according to soil moisture
and the actual water demand of crops in order to decide
whether or not to conduct water and how long irrigation time
is. Coordinator node then sends the irrigation information to
irrigation controller node. Irrigation controller controls pump
valve at corresponding region to open or turn off. Thus a
closed-loop irrigation network control system is formed and it
can implement water-saving irrigation to crops. After a
sampling period, each sensor node immediately comes into
state of hibernation to conserve energy until the next cycle.
Flow chart of software design for sensor node, coordinator
node and irrigation controller node are expressed separately in
Fig. 6, Fig. 7 and Fig. 8.

Figure 6. Software flow chart of sensor node

Figure 7. Software flow chart of coordinator node

Figure 8. Software flow chart of irrigation controller node
VI. CONCLUSIONS
In this paper, wireless sensor networks and fuzzy control
technology are introduced to design a water-saving irrigation
system. By means of the characteristics that wireless
transmission is conducive to fruit seedlings irrigation, we
design a wireless sensor network and solve the problem of real-
time transmission of irrigation information. We apply fuzzy
control technology to intelligent water-saving irrigation, which
can meet the requirement and development of intellectualized
and network-based agriculture. We believe that the exploratory
study applying wireless sensor networks combined with fuzzy
control technology to precise agriculture projects must have a
broad popularization value and application prospect in the
future.
REFERENCES
[1] Sun Limin, Li Jianzhong, Chen Yu, Wireless sensor networks,
Beijing: Tsinghua university press, 2005, pp.1-20.
[2] Cui li Jue Hailing etc., Overview of Wireless Sensor Networks,
Journal of Computer Research and Development, 2005,42(1), pp.163-
174.
[3] Fang Xujie, Zhou Yiming, Chen Wenliang, Ding Chaohong,Yang
Xianglong, The design of wireless intelligent irrigation systan based on
ZigBee technology, Research on Mechanization of Farming, 2009.1,
pp.114-118.
[4] Zeng Liancheng, Luo Zhixiang, Xie Zhijian, Auto system based on
WSN for water-saving irrigation of farmland, Farming Network
Information, 2008.11, pp.13-15.
[5] Guo Zhengqin, Wang Yiming, Yang Weizhong, Feng Lei, Yang
Shaohui, Intelligent irrigation control system based on fuzzy control,
Research on Mechanization of Farming, 2006.16, pp.103-108.
[6] Zhang Q, John, Ken T, Application of fuzzy logic in an irrigation
control system, Proceedings of the IEEE international conference on
industrial technology[C], 1996.
[7] Luo Huiqian, Zhang Qing, Qiao Xiaojun, Research and design of
wireless sensor networks hardware flat roof in agriculture,
Microcomputer Applications, 2006, 27(5), pp.534-537.
[8] IEEE 802.15.4 web site, http://www.ieee802.org/15/pub/TG4.html.
[9] Patrick Kinney, etc, ZigBee technology: Wireless control that simply
works, Communications design conference, Oct 2, 2003.
[10] Junpu Wang, Intelligent Control, Press of University of Science and
Technology of China, 1996, pp.90-135.
Authorized licensed use limited to: SASTRA. Downloaded on March 01,2010 at 05:43:24 EST from IEEE Xplore. Restrictions apply.