Wireless Sensor Network, 2010, 2, 698-702
doi:10.4236/wsn.2010.29084 Published Online September 2010 (http://www.SciRP.org/journal/wsn)
Sensors Dynamic Energy Management in WSN
Xianghui Fan, Shining Li, Zhigang Li, Jingyuan Li
College of Computer Science, Northwestern Polytechnical University, Xi’an, Shaanxi, China
Received July 2, 2010; revised August 16, 2010; accepted September 12, 2010
Abstract
A wireless sensor node is typically battery operated and energy constrained. Therefore, it is apparent that
optimal energy management is one of the most important challenges in WSN development. However, energy
management requires in-depth knowledge and detailed insight concerning specific scenarios. After Carrying
out a large number of experiments in precision agriculture, we find that it is the sensors that have never been
concerned consuming the most energy of the node. In order to conserve energy and prolong the lifetime of
WSN, we design and carry out a dynamic energy management strategy for sensors. The basic idea is to shut
down all sensors’ power when not needed and wake them up when necessary. Valuable conclusions are
extracted and analyzed.
Keywords: WSN (Wireless Sensor Network), Precision Agriculture, Dynamic Energy Management, TinyOS
1. Introduction
A large number of intelligent micro-sensor nodes with
sensing, processing and wireless communicating capa-
bilities form wireless sensor network (WSN), which
completes complicated tasks in some specific field, such
as precision agriculture. Compared with old methods,
WSN has significantly drawn extensive attention. It does
not rely on fixed infrastructure and has many characteris-
tics such as fast setup, strong survivability and so on [1].
It has been considered as a good scheme to conduct pre-
cision agriculture data collection and processing. In 2002,
Intel has a project looking at how WSN can be used to
improve grape production. They worked with agricul-
tural scientists on a long-term deployment of WSN in a
wine grape vineyard. By densely monitoring and analyz-
ing they found the relationship between grape quality
and climatic conditions. It has been proved that WSN
could play a role in precision agriculture.
Just the same as other applications, energy constraint
of sensor nodes is the major problem for precision agri-
culture. Data aggregation [2] and low power listening [3]
algorithms are effective method to reduce energy con-
sumption in normal wireless sensor networks. However,
after a sufficient number of experiments we found that
energy consumption in precision agriculture has some
special issues. Generally speaking, in order to monitor
the growth conditions of crops, one node has to connect
with many sensors, such as Co2 sensor, air temperature
sensor, air humidity sensor, light sensor, soil tempera-
ture/moisture sensor and so on. Although the sensors
consume a large portion of the energy, we never pay any
attention to this issue in our previous research. It is
necessary to reduce the energy consumption of sensors.
We design and carry out a sensor dynamic energy man-
agement (SDEM) to reduce energy consumption of sen-
sors and extend network lifetime. The basic idea is to
shut down sensors when not needed and wake them up
when necessary [4,5]. The experimental results indicate
that SDEM is an effective technique in reducing node
energy consumption without significantly degrading
performance.
The remainder of this paper is organized as follows.
Section 2 gives the energy consumption of all parts of
the sensors in precision agriculture. And we get a con-
clusion that the sensors consume most of the energy. The
architecture of the SDEM is described in Section 3, and
Section 4 reports the hardware and software design and
some considerations about implementation. In Section 5,
the actual deployment is described in detail, and finally
Section 6 reports our conclusion and gives some direc-
tions on the ongoing work.
2. Sensors Energy Consumption Analysis
A sensor node has several major components: processor,
memory, A/D converter, sensing unit and radio. Each
node sleep mode corresponds to a particular combination

X. H. FAN ET AL.
699
of component power modes. In general, if there are N
components labeled (1, 2, … , N) each with i
K
sleep
modes, the total number of node sleep modes is
1
n
i
ik 
 [6]. However, from a practical point of view not
all sleep modes are useful. Optimizing the key issue
could achieve a noticed effect. In our case, the NPUnode
used in precision agriculture consists of a processor
Atmel2561, which has rich resource such as 256K bytes
in-system programmable flash, 4K bytes EEPROM, 8K
bytes SRAM and so on, a RF230 chip acts as the radio
unit, an AT45DB041B chip acts as an extern memory.
And the NPUnode has been equipped with six sensors,
just as showing in Figure 1. Each component is con-
trolled by the micro-operating system (TinyOS). Tabl e 1
enumerates the characteristics of all devices of the
NPUnode.
Sensor transducers translate physical phenomena to
electrical signals, and are classified as either analog or
digital devices depending on the type of their output.
There are several sources of power consumption in a
sensor, including 1) signal sampling and conversion of
physical signals to electrical ones, 2) signal conditioning,
and 3) analog to digital conversion. Given the diversity
of sensors there is no typical power consumption number.
In general, however, passive sensors such as temperature
sensors etc., consume negligible power relative to other
components of sensors, array sensors such as Co2 sen-
sors can be large consumers of power. As shown in Ta-
ble 1, the current value of Co2 sensor is 30mA, some-
times even reaches the max value 150mA. And other
sensors also have power dissipation.
In accordance with Table 1, we draw a figure of en-
ergy consumption of major components. Just as shown in
Figure 2, the sensors energy consumption is far more
than other components. According to electric power
formula WPtUIt , we know that reducing the run-
ning time is the only way to decrease energy consump-
tion while not change the sensors characters. T is the
data collection period, while woking Tis the working time
of sensors and idle T means the period that sensors is in
the idle state. And the relationship between them is
working idle
TT T and idle working TT . Data collection
period is a long time in agriculture application. For ex-
ample, the nodes we have deployed in greenhouse in
YangLing only need to report the sensor data once an
hour. This means that the sensors only need to work for
several minutes per hour, so they should be turned off
when they are idle to save energy.
3. Sensors Dynamic Energy Management
Nodes energy consumption is application specific. In
precision agriculture the sensors consume the greatest
portion of the node energy. The reason is that one node is
always connected with several sensors to monitor plant
growth environment. However, the most time of sensors
is in idle state and waste a lot of energy. For the purpose
of saving energy of nodes and extending the life of WSN,
we design the Sensors Dynamic Energy Management
(SDEM) strategy which turns on the sensors when the
node receives an acquisition command from root and
turns it off when the sensors are in the idle state. The
principle of SDEM is showed as Figure 3.
3.1. Hardware Design
SDEM needs independent hardware design supporting.
The NPUnode supports power with DC 5V or 3.6V
Nickel-Hydrogen Battery. And the sensors used to
monitor the plant growth environment need different
voltage. A power supply of DC3.0V is designed for light
sensor, soil temperature sensor, temperature and humid-
ity sensor. Co2 sensor and soil Moisture sensor power
are supplied with DC5.0V. The hardware design is de-
scribed as in Figure 4. XC6221B and TPS61202 chips
which are connected with ATmega2561 I/O pins trans-
late battery power to DC3.0V and DC5.0V to meet above
Figure 1. The NPUnode and equipped sensors.
Table 1.The devices of the node and their characteristics.
Device Type Characteristics
Processor ATmega
2561V
256K Bytes Flash, 4K Bytes EEPROM,
8K Bytes Internal SRAM,
Power :2.7~5.5V,10mA
Radio RF230
Power: Voltage: 1.8V~3.6V, SLEEP: 0.1
μA, RX: 16 mA, TX: 17 mA; Fast
Power-Up Time < 1 ms
Flash AT45DB0
41
Power Supply: 2.7V~3.6V, 4 mA Active
Read
Temp and
Humi
Sensor
SHT11
Temperature range:-40℃ to +123.8
℃;Humidity range: 0 to 100% RH;
Power : 3V, 0.5mA
Light
Sensor ISL29002 Range: 10,000lux~100,000lux;Power : 2.5V to 3.3V, <10mA
Soil Temp
Sensor PT1000 Range: -70°C to + 500°C;Power:3VDC, 3mA;Response time < 10s
Soil Mois
Sensor TDR-3
Range: 0～100%（m3/m3）；Power
Supply: 4.5~5.5V DC, 50～70mA；
Response time: < 1s
Co2 SensorGE/Telaire
6004
Range:0-2000ppm;Power: 4.3 VDC ~7.0
VDC, 30~150mA;Response time: < 2
min

X. H. FAN ET AL.
700
Figure 2. Energy consumption of node in precision agricul-
ture.
Figure 3. Principle of sensor dynamic energy management.
demand. Inputting L level signal for I/O pins we can
provide power for sensors. On the other hand, we can
turn off the sensors power supply by inputting H level
signal for I/O pins. It is easy to implement in TinyOS.
3.2. Software Design
The software system considered here is based on TinyOS
2.1 [7] which is an embedded operating system espe-
cially for WSN applications. The extreme power limita-
tion of nodes forces the operating system to take very
different approaches than traditional computing classes.
TinyOS is implemented using NesC programming lan-
guage—a dialect of C programming language. It inte-
grates with model of components/module and events
driven model. Over the past few years, TinyOS has
grown from a small research project to dominant operat-
ing system for low power wireless sensor networks.
At a high level, TinyOS provides three things to make
writing systems and applications easier [8]. 1) a compo-
nent model, which defines how you write small, reusable
pieces of code and compose them into larger abstractions;
2) a concurrent execution model, which defines how
components interleave their computations as well as how
interrupt and non-interrupt code interact; 3) application
programming interfaces (APIs), services, components
libraries and an overall component structure that simplify
writing new applications and services. The HplAtm256-
GenerallIOC is the most important components to design
the SDEM. It exposes the ATmega256’s 53 digital I/O
pins as 53 GeneralIO interfaces, hiding the slightly dif-
ferent instruction sequences needed to perform some
operations on some I/O pins. Some I/O pins can be set
atomically in a single assembly instruction. On the pur-
pose of turn on/off the sensors power, we need to use the
interface GeneralIO to enable/disenable the pins of At-
mel2561 which connect with XC6221B and TPS61202
by inputting Low/High level signal. The GeneralIO in-
terface offers commands to configure, read and write a
typical microcontroller digital I/O pin. Firstly, after the
OS initializes all needed components and booted suc-
cessfully, we call GeneralIO.makeoutput() to make the
pins as a controller. Then the NPUnode enters a
Figure 4. The hardware design of SDEM.

X. H. FAN ET AL.
701
ready state waiting for the command from root. Once the
node receives the sampling command, it then calls the
command GeneralIO.set() to turn on the sensors power
and waits for 2 minutes for warming up the sensors.
After reading all sensors data and sending it back suc-
cessfully, nodes calls GeneralIO.clr() to turn off the sen-
sors power. The Figure 5 and Figure 6 show the imple-
mentation of SDEM.
Figure 5.The module implementation of SDEM.
configuration AgriMonitorAppC{}
implementation{
Components AgriMonitorC as App;
Components MainC;
App.Boot -> MainC;
Components HplAtm256GeneralIOC;
App.PowerControl -> HplAtm256GeneralIOC.PortC0;
...
}
Figure 6.The configuration wired the needed components.
4. Experiment Results
We deployed NPUnodes in field crop production in
YangLing, LuoChuan, AnSai in ShaanXi province, just
as showing in Figure 7. The NPUnodes are equipped
with Co2 sensors, temperature and humidity sensors,
light sensors, soil temperature sensors and soil moisture
sensors to collect important data for plant growth envi-
ronment. The collected data is gathered at the edge of the
field by a field gateway and further transferred via GPRS
to a PC server for data analysis. Once something odd
happened the server will send message via MMS to
farmers. In addition to the agronomic experiment, we
expect to gather data and statistics on the behavior of
NPUnodes in real-world experiment. For energy-
efficiency considerations, we use SDEM strategy in
NPUnodes and they reported data only once per hour. To
get the exact value of sensors, we need to wait sensors
warm-up for two minutes. The rest of the time we turn
off the sensor power to save energy.
From collected voltage data of NPUnodes, we make a
comparison between the SDEM strategy and before
methods. Just showing in Figure 8, we can give the con-
clusion that the SDEM strategy has significantly saved
energy. After making in-depth research we find that the
Figure 7. NPUnode deployed in greenhouses with six sen-
sors.
2468 10 12 14 16 18 20
3.4
3.45
3.5
3.55
3.6
3.65
3.7
3.75
3.8
3.85
Time /h
Voltage/V
Us ing SDMP
Old m ethod
Figure 8. Lift time of sensors between SDEM and before
method.

X. H. FAN ET AL.
702
SDEM strategy almost has extended the NPUnodes life
time from 1 month to 3 months. However, SDEM should
save sensors energy 30 times than not using SDEM
nodes, theoretically on the condition that NPUnode re-
ports sensors data once per hour (all sensors only turn on
for 2 minutes). The reason is that not only the sensors
consume the energy, other components such as radio,
processor etc. also consume the energy.
5. Conclusions and On-Going Work
In precision agriculture, nodes are always equipped with
several sensors which consume a large portion of nodes
energy, especially active sensors. In order to save energy
and extend the lifetime of nodes, we design and imple-
ment the SDEM strategy based on our NPUnodes. The
basic idea is to shut down devices when not needed and
wake them up when necessary. And then we deployed
our nodes in greenhouse. From the voltage data collected
we give the conclusion that the SDEM strategy can sig-
nificantly save energy of nodes, and extend the life time
of nodes from one month to three month.
The nodes need continuous work for one year or more
in precision agriculture. So the nodes need an unfailing
supply of energy. Next time, we will make a serial re-
search and supply solar power for NPUnodes. Maybe it
is a way to solve the energy problem finally.
6. References
[1] S. M. Xiong, L. M. Wang, X. Q. Qu and Y. Z. Zhan,
“Application Research of WSN in Precise Agriculture Ir-
rigation,” International Conference on Environmental
Science and Information Application Technology, Wuhan,
July 2009, pp.297-330.
[2] B. Krishnamachari, D. Estrin and S. Wicker, “Impact of
Data Aggregation in Wireless Sensor Networks,”
DEBS’02.
[3] J. Polastre, J. Hill and D. Culler, “Versatile Low Power
Media Access for Wireless Sensor Networks,” Proceed-
ings of the 2nd International Conference on Embedded
Networked Sensor Systems, Maryland, November 2004.
[4] A. C. Sinha, “Dynamic Power Management in Wireless
Sensor Networks, Design & Test of Computers,” Pro-
ceedings of IEEE, Vol. 18, No. 2, 2001, pp. 62-74.
[5] R. C. Luo, L. C. Tu and O. Chen, “An Efficient Dynamic
Power Management Policy on Sensor Network,” Pro-
ceedings of the 19th International Conference on Ad-
vanced Information Networking and Applications, 2005,
pp. 341-344.
[6] C. Lin, N. Xiong, J. H. Park and T.-H. Kim, “Dynamic
Power Management in New Architecture of Wireless
Sensor Networks,” International Journal of Communica-
tion Systems, Vol. 22, No. 6, 2008, pp. 671-693.
[7] TinyOS Tutorials, Internet Available:
http://docs.tinyos.net/index.php/TinyOS_Tutorials
[8] P. Levis and D. Gay, “TinyOS Programming,” Cam-
bridge University Press, England, 2009, pp. 6-7.