Impact of the temperature and humidity variations on link quality of xm1000 mote sensors

International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.5, No.6, December 2014
DOI : 10.5121/ijasuc.2014.5603 21
IMPACT OF THE TEMPERATURE AND
HUMIDITY VARIATIONS ON LINK QUALITY OF
XM1000 MOTE SENSORS
Samia Allaoua Chelloug
Networks and Communication Systems Department, College of Computer and
Information Sciences, Princess Nourah bint AbdulRahman University, Riyadh
Kingdom of Saudi Arabia
ABSTRACT
The core motivations of deploying a sensor network for a specific application come from the autonomy of
sensors, their reduced size, and their capabilities for computing and communicating in a short range.
However, many challenges for sensor networks still exist: minimizing energy consumptions, and ensuring
the performance of communication that may be affected by many parameters. The work described in this
paper covers mainly the analysis of the impact of the temperature and humidity variations on link quality of
XM1000 operating under TinyOS. Two-way ANOVA test has been applied and the obtained results show
that both the temperature and humidity variations impact RSSI.
KEYWORDS
sensor network, XM1000, TinyOS, ANOVA, temperature, humidity, RSSI, LQI.
1. INTRODUCTION
By definition, a sensor network is deployed either inside the phenomenon to be monitored or very
close to it. Each sensor consists of a microcontroller, a transceiver, a source of power, and a
sensing unit [1]. Sensor networks are being considered for use in: smart homes, E-health,
industrial automation,… Some applications have stringent requirements that concern the delay
and the reliability of communication [2]. However, recent studies show that signals fluctuate and
sometimes it is very hard to ensure a certain degree of quality of service. Hence, the aim of our
experimental study is to show how the temperature and humidity fluctuations may affect the link
quality of XM1000 sensor motes that have been manufactured by ADVENTICS and use an
MSP430 microcontroller. It provides a program flash memory and data program storage space.
XM1000 functions under TinyOS2.x operating system which is appropriate for programming
embedded systems. In particular, our experimental study focuses on understating the effect of the
two parameters (temperature and humidity) upon link quality.
This article provides a literature review of related works. Sections 3 and 4 explains the hardware
architecture of XM1000 and TinyOs. Sections 5 outlines the principles of ANOVA. Then, our
experimental results are illustrated and analyzed by two-way ANOVA.

22
2. IMPACT OF METEROLOGICAL CONDITIONS ON SENSOR NETWORKS
Designing appropriate protocols for sensor networks is a critical task due to their hardware
constraints. Many research works exploit some indicators that can be obtained via hardware to
process and communicate data in wireless sensor networks. In particular, the research community
faces the challenge of power control and power saving to extend the lifetime of wireless sensor
networks. Power control and power saving may be applied jointly with other sensor protocols to
efficiently localize sensors, route data packets, ovoid interference,… Extensive studies are based
on measuring the strength of the received signal (RSSI) or the link quality (LQI) that can be
calculated by the transceiver. Some results show that LQI is a good indicator for new sensor
platforms [3]. Unfortunately, meteorological conditions (temperature, humidity) may affect the
link quality. Hence, it is necessary to analyze the impact of such parameters to assure accuracy
while controlling or saving power. Some research works provided experimental studies for
understating the impact of meteorological conditions on the link quality of sensor networks.
The study presented in [4] supports greenhouse monitoring. It is based on the deployment of four
sensors to monitor temperature, humidity, light, and the level of carbon Dioxide in a green house.
The hardware platform sends compressed IPV6 packets over IEEE802.15.4 with a loss rate of
5%. Authors conclude through their experimental study that there is a relationship between the
relative temperature and humidity inside a greenhouse such that the temperature increases once
the humidity has decreased.
The focus of [5] is to understand signal propagation between fixed and a mobile sensor and
evaluate the impact of meteorological conditions on signal propagation. Memsic wireless sensors
with three integrated sensors( humidity, temperature, barometric pressure) have been used. Their
study demonstrates that the signal variation at each position is small and the variations in the
presence of large water bodies is influenced by changes in temperature, humidity, and the
refractive index of the medium.
[6] Proposed an idea for power control in wireless sensor networks that depends on the evaluated
temperature and humidity. They developed a controller that adapts power according to RSSI
feedback.
The main research interest of [7] is to study the impact of meteorological conditions on 802.15.4
links. Authors of the paper evaluated different conditions in Sweden. Their research conclude that
the temperature variations correlate the most with RSSI and PRR.
[8] Conducted experiments and designed a scheme that combines both temperature-aware link
quality compensation and a closed-loop feedback process to adapt to link quality.
In this paper, we investigate how the temperature and humidity variations may affect link quality
of XM1000 sensors. Our idea studies the influence of the two factors on RSSI and LQI and uses
ANOVA (variance analysis) to recognize the importance of each parameter.
3. XM1000 SENSOR MOTE
The consolidation of micro-electro mechanical systems (MEMS) and digital wireless
communications has led to the ability of deploying small devices that are application-specific. In

23
particular, sensors are provided with sensing and communicating capabilities that overcome the
limitations of traditional networks for remote real time monitoring. Many kinds of sensors are
available in the market and are designed and manufactured by leader industrial companies:
CrossBow, Atmel,… XM1000 mote module is based on TelosB specification which is the result
of the research conducted in UC Berkely. TelosB is equipped by MSP430 microcontroller which
is characterized by many features [9]:
Complete system on chip that includes LCD control, ADC, I/O ports, RAM(10 KB),
ROM, timer, UART,…
Low power consumption: 4.2 nW/instruction.
High speed: 300 ns/instruction a clock speed of 3.3MHZ.
Risc set of 27 core instructions that make the process of programming very easy.
48 KB of flash memory.
802.15.4 compliant mode and consumes low power for communications.
Current in active mode is 1.8 mA and the current in sleep mode is 5.1MA.
Three sensors are integrated on XM1000: temperature, humidity, and light sensors. Its dimensions
are 81.9mm*32.5mm*6.55mm including USB connector. It is provided by 116KB of program
flash, 8KB of data program and 1MB for logging. The radio frequency chip of XM1000 operates
in 2.4-2.485 GH with a rate of 250Kbps and can cover a range of 120m outdoor and between 20
to 30 m indoor [10].
Some research works have used XM1000 for carrying experiments:
ManOs has been tested on XM1000 to evaluate the potential to make sensor programming
accessible for a broader range of programmers [11].
[12] studied and compared a set of empirical models to estimate the path loss by means of
XM1000.
[13] demonstrated that device heterogeneity may improve energy by moving resource intensive
tasks to other nodes in the network. Their study has focused on imote2, Inga, Mulle v5-2,
Sunspot, TelosB, and XM1000.
4. TinyOS
XM1000 is compatible with TinyOs2.x which is open project that was developed at University of
California. It is based on Nesc language that is suited for programming embedded systems. Its
syntax is similar to C syntax but NesC is enhanced to support concurrency and allows to structure
a program into linked components to reduce the compile time. Unlike C, NesC locates all
functions at compile time. NesC compilers generate a C file after processing the NesC file. The C
file is then compiled to the desired microcontroller. Each NesC application program is divided
into linked modules to avoid any explicit pointer. The configuration file specifies the links
between the system modules however, the interface specifies the interaction between two
components: the user which implements a set of events and the provides that implements a set of
commands [14].

24
Figure1 . Component of NESC programming model.
The specification of each component (module or configuration) contains two parts: the signature
and the implementation. The signature describes if the component use or provide any interface.
5. ANOVA
ANOVA is a statistical technique that takes one or more populations as input and tests the
inequality among population means. More specifically, ANOVA analyses the effect of one or two
qualitative factors upon a quantitative variable. ANOVA is based on hypothesis test [15]:
• ‫:0ܪ‬ µଵ
= µଶ
= ⋯ µ௞
This means that all population means are equal. K is the total
number of groups.
• ‫:ܣܪ‬ At least two of the means differ.
ANOVA calculates the F-test such that:
‫ܨ‬ =
ெௌோ
ெௌா
(1)
‫ܴܵܯ‬ is the mean square of the treatment and ‫ܧܵܯ‬ is the mean square of error.
‫ܶܵܯ‬ =
ௌௌ்
௄ିଵ
(2)
‫ܧܵܯ‬ =
ௌௌா
௡ି௄
(3)
ܵܵܶ is the variation between however ܵܵ‫ܧ‬ is the variation within and n is the number of samples
in each group.
F is expected to be nearly equal to 1 if all population means are equals. Otherwise, the hypothesis
‫0ܪ‬ will be rejected.

25
6. EXPERIMENTAL STUDY
Our experimental study involves two XM1000 sensors. The first one is considered as a base
station and sends periodically a message to the second one. Upon receiving a message, the target
sensor measures the temperature, the humidity, RSSI, and LQI. The distance between the two
sensors is around 20 Cm.
Figure 2 illustrates the hardware used in our experiment. The plotting of the variation of the
measured RSSI according to the temperature is shown in figure 3. Linear regression has been
applied and the color of the trendline is black. It is clear that there is a polynomial relation
between RSSI and the temperature. Moreover, we can conclude from the result of the regression
that RSSI is very high in the first group (12-270
) and decreases once the temperature exceeds 270
Figure 2. XM1000 sensors.
Figure 3. Temperature versus RSSI.

26
The plotting of the variation of the measured RSSI according to humidity is shown in figure 4.
Linear regression has been applied and the color of the trendline is black. It is clear that there is a
linear relation between RSSI and humidity. We can recognize two groups of RSSI measures after
applying linear regression and RSSI increases in a linear manner from the first group that
corresponds to humidity values ranging from 9 to 13% to the second one that corresponds to
humidity values ranging from 13 to 15%.
Figure 4. Humidity versus RSSI.
Figures 5 and 6 illustrate the variation of LQI according to temperature and humidity
respectively. From the plots of linear regression, it is clear that the variation of LQI versus
temperature and humidity is linear.
Figure 5. Temperature versus LQI.

International Journal of Ad hoc, Sensor &
We were asking the question about the important parameter that impacts RSSI as the function of
variation of RSSI according to temperature is different from its function of variation according to
humidity. To this end, we applied ANOVA after preparing three groups of tem
warm, low) and two groups of humidity (high and low).
Three comments should be highlighted form table 1.
-The variance of each group is not far away from 1.
-The values for F are greater of F crit.
-P-values are smaller than 0.5.
So, H0 hypothesis is maintained and both temperature and humidity influence RSSI.
7. CONCLUSION
Estimating link quality is an important factor for many sensor protocols.
concerned the estimation of both RSSI and LQI in a small XM1000
show that the impact of the temperature and humidity
analysis shows that RSSI is influenced by temperature as well as humidity. Hence it is necessary
to save energy according to those two para
and saving in XM1000 sensor networks.
Figure 6. Humidity versus LQI.
the question about the important parameter that impacts RSSI as the function of
humidity. To this end, we applied ANOVA after preparing three groups of temperature (high,
warm, low) and two groups of humidity (high and low).
Three comments should be highlighted form table 1.
he variance of each group is not far away from 1.
he values for F are greater of F crit.
H0 hypothesis is maintained and both temperature and humidity influence RSSI.
Estimating link quality is an important factor for many sensor protocols. Our experimental study
concerned the estimation of both RSSI and LQI in a small XM1000 sensor network.
temperature and humidity variations on LQI are linear. ANOVA
to save energy according to those two parameters. We will focus in the future on power control
and saving in XM1000 sensor networks.
Ubiquitous Computing (IJASUC) Vol.5, No.6, December 2014
27
the question about the important parameter that impacts RSSI as the function of
perature (high,
Our experimental study
sensor network. Our results
on LQI are linear. ANOVA
We will focus in the future on power control

28
Table 1. AVOVA results.
REFERENCES
[1] Akyildiz. I-F, Su. W, Sankarasubramaniam . Y, Cayirci.E, “Wireless sensor networks, a survey” , in
the Proceedings of the International Journal of Computer and Telecommunications Networking , Vol.
38, Issue 4 pp. 393-422, 2002
[2] Boano. C-A, Tsiftes.N, Voigt.T, Brown.J, Roedig.U, “The impact of temperature on outdoor
industrial sensornet applications” , In the Proceedings of IEEE transactions of Industrial Informatics ,
Vol. 6, No. 3, pp. 451-459, 2010.
[3] Srinivasan.K, Levis.P , “RSSI is Under Appreciated”, In Proceedings of the Third Workshop on
Embedded Networked Sensors (2006) .

29
[4] Ahonen.T , Vir ankoski.R , Elmusrati.M, “Greenhouse Monitoring with Wireless Sensor Networks:,
in the Proceedings of IEEE Mechtronic and Embeded Systems and Applications, pp.403-408, Beijing,
2008.
[5] Harun.A, Ndzi.D-L, Ramli.M-F, Shakaff.A-Y-M, Ahmad.M-N, Kamarudin.L-M, Zakaria.M, and
Yang.Y, "Signal propagation in aquaculture environment for wireless sensor network applications," in
the Proceedings of Progress In Electromagnetics Research, Vol. 131, pp.477-494, 2012
[6] Ortega-Corral.C, Palafox.L-E, García-Macías.J-A, Sanchez-Garcia.J, Aguilar.L, Van Nieto-
Hipólito.J-A, “Transmission Power Control based on temperature and relative humidity”, in the
proceedings of ISSNIP, pp.1-6, 2014.
[7] Wennerström.H, “ Meteorological Impact and Transmission Errors in Outdoor Wireless Sensor
Networks”, IT Licentiate Thesis, Uppsala, 2013.
[8] Lee.J, Chung.K, “An efficient transmission power control scheme for temperature variation in
wireless sensor networks” . in the proceedings of Sensors, vol. 11, no. 3, pp. 3078-3093, 2011.
[9] http://www.win.tue.nl/~johanl/educ/RTcourse/MSP430%20UART.pdf
[10] http://www.advanticsys.com/shop/asxm1000-p-24.html
[11] Elsts.A, Selavo.L, “Improving the Usability of Wireless Sensor Network Operating Systems” , in the
Proceedings of the Federated Conference on Computer Science and Information Systems, pp. 89-94,
2013.
[12] Donmez.R-G , “A study on path loss for indoor wireless communication networks”, Master Thesis,
2013.
[13] Reinhardt.A, Burgstahler.D, "Exploiting Platform Heterogeneity in Wireless Sensor Networks by
Shifting Resource-Intensive Tasks to Dedicated Processing Nodes", In the proceedings of IEEE 14th
International Symposium on "A World of Wireless, Mobile and Multimedia Networks", WoWMoM
2013, pp. 1-9, Spain, 2013.
[14] http://www.tinyos.net/
[15] http://www.cimt.plymouth.ac.uk/projects/mepres/alevel/fstats_ch7.pdf
Biography
Samia Allaoua Chelloug received the Engineer degree (Computer Science) in 2003, Magister degree
(Computer Science) in 2006, and the Phd degree (Networking) in 2013 (all from University of Constantine,
Algeria).
From 2006 to 2013 she was an Assistant Professor in the Faculty of Computer Science in Constantine
University.
In August 2013, she joined the Department of Networks and Communication Systems at Princess Nourah
University (Riyadh, Kingdom of Saudi Arabia) where she is presently an Assistant Professor.
Her current research interests include: wireless sensor networks, cognitive radio, Vanets, and Fourth
Generation communication standards.

Impact of the temperature and humidity variations on link quality of xm1000 mote sensors

Related slideshows

Recommended for you

Recommended for you

More Related Content

What's hot

What's hot (19)

Viewers also liked

Viewers also liked (16)

Similar to Impact of the temperature and humidity variations on link quality of xm1000 mote sensors

Similar to Impact of the temperature and humidity variations on link quality of xm1000 mote sensors (20)

Recently uploaded

Recently uploaded (20)

Impact of the temperature and humidity variations on link quality of xm1000 mote sensors