2009 Third International Conference on Sensor Technologies and Applications

Real-time Wireless Sensor Network for Landslide Detection

Maneesha V. Ramesh
Department of Computer Science, Amrita School of Engineering
Amrita Vishwa Vidyapeetham (AMRITA University)
Clappana. P.O, Vallikkavu, Kollam, Kerala, India-690525
[email protected]

Abstract—Wireless sensor networks are one of the emerging This paper discusses the design and deployment of a
areas which have equipped scientists with the capability of landslide detection system using a WSN system at Anthoniar
developing real-time monitoring systems. This paper discusses Colony, Munnar, Idukki (Dist), Kerala (State), India. The
the development of a wireless sensor network(WSN) to detect deployment site has historically experienced several
landslides, which includes the design, development and landslides, with the latest one occurring in the year 2005,
implementation of a WSN for real time monitoring, the which caused a death toll of 10 (people).
development of the algorithms needed that will enable efficient The remainder of the paper is organized as follows.
data collection and data aggregation, and the network Section II describes related work in WSN systems, and other
requirements of the deployed landslide detection system. The
methods for landslide prediction. In Section III, we describe
actual deployment of the testbed is in the Idukki district of the
Southern state of Kerala, India, a region known for its heavy
the sensors for landslide detection. Section IV details the
rainfall, steep slopes, and frequent landslides. different wireless sensor network algorithms implemented in
the landslide detection network. Section V presents our
Keywords- wireless sensor network, distributed algorithms, wireless sensor test bed. Field deployment, its design
heterogeneous networks, landslide concerns, and experiences are described in Section VI.
Finally we conclude in Section VII and discuss the future
work also in the same section.
I. INTRODUCTION
Environmental disasters are largely unpredictable and
occur within very short spans of time. Therefore technology II. RELATED WORK
has to be developed to capture relevant signals with Wireless sensor technology has generated enthusiasm in
minimum monitoring delay. Wireless sensors are one of the computer scientists to learn and understand other domain
cutting edge technologies that can quickly respond to rapid areas which have helped them to propose or develop real-
changes of data and send the sensed data to a data analysis time deployments. One of the major areas of focus is
center in areas where cabling is inappropriate. environmental monitoring, detection and prediction.
Wireless sensor network(WSN) technology has the The Drought Forecast and Alert System (DFAS) has
capability of quick capturing, processing, and transmission been proposed and developed in [2]; it uses mobile
of critical data in real-time with high resolution. However, it communication to alert the users, whereas the deployed
has its own limitations such as relatively low amounts of system uses real time data collection and transmission using
battery power and low memory availability compared to the wireless sensor nodes, WiFi, satellite network and also
many existing technologies. It does, though, have the through internet. The real streaming of data through
advantage of deploying sensors in hostile environments with broadband connectivity provides connectivity to wider
a bare minimum of maintenance. This fulfills a very audience.
important need for any real time monitoring, especially in An experimental soil monitoring network using a WSN is
hazardous or remote scenarios. presented in reference [3], which explores real-time
We aim to use the wireless sensor networks in the measurements at temporal and spatial granularities. Paper [4]
landslide scenario1 for estimating the chance occurrence of describes a state-of-the-art system that combines multiple
landslides. India faces landslides every year with a large sensor types to provide measurements to perform
threat to human life causing annual loss of US $400 million deformation monitoring. Reference [5] discusses the topic of
[1]. The main goal of this effort is to detect rainfall induced slip surface localization in wireless sensor networks, which
landslides which occur commonly in India. can be used for landslide prediction. A durable wireless
sensor node has been developed [6], which can be employed
in expandable wireless sensor networks for remote
1
This work has been partially funded by the WINSOC project, a Specific monitoring of soil conditions in areas conducive to slope
Targeted Research Project (Contact Number 003914) co-funded by the stability failures.
INFSO DG of the European Commission within the RTD activities of the In this paper, real time deployment of a heterogeneous
Thematic Priority Information Society Technologies and also by Wireless network in India for landslide detection has been discussed.
Sensor Network for Real-time Landslide Monitoring project funded by
Department of Information Technology (DIT), India. This study incorporates both theoretical and practical

978-0-7695-3669-9/09 $25.00 © 2009 IEEE 405

DOI 10.1109/SENSORCOMM.2009.67
knowledge from diverse domains such as landslides and that of each single sensor as discussed in research papers
geomechanics, wireless sensor, Wi-Fi, and satellite networks, [9][10][11]. Therefore, the wireless sensor network at the
power saving solutions, and electronic interface and design, deployment site uses the same concept for distributed
among others, which paved the design, development and detection, estimation and consensus to arrive at reliable
deployment of a real-time landslide detection system using a decisions.
wireless sensor network. Our study concentrates on the detection of rainfall
induced landslides, and so, the most relevant data has to be
III. SENSORS FOR LANDSLIDE DETECTION collected during rainy season. Therefore, we have developed
Under heavy rainfall conditions, rain infiltration on the a threshold based algorithm [8] that will influence the
slope causes instability, a reduction in the factor of safety; sampling rate of the geological sensors and the transmission
transient pore pressure responses, changes in water table of data to higher layers using rainfall and pore pressure based
height, a reduction in shear strength which holds the soil or alert levels. Along with these methods, state level transitions
rock, an increase in soil weight and a reduction in the angle have also been incorporated in the wireless sensor nodes.
of repose. When the rainfall intensity is larger than the slope Both methods reduce the energy consumption per node
saturated hydraulic conductivity, runoff occurs [7]. which will also contribute to reduced energy consumption
The geophysical sensors selected for the in-situ throughout the network. These requirements, however, lead
measurements are pore pressure transducers, soil moisture to the need of time synchronization, and the algorithm
sensors, geophones, stain gauges and tiltmeters, based on planned for implementation in our network is discussed in
their relevance in finding the causative geological factors for the research paper [12].
inducing landslides under heavy rainfall conditions.
However, commercially available wireless sensor nodes do V. WIRELESS SENSOR TESTBED
not have implanted geophysical sensors and this constraint The wireless sensor network follows a two-layer
has lead us to implement data acquisition boards to connect hierarchy, with lower layer wireless sensor nodes, sample
the external sensors to the wireless sensor nodes. The and collect the heterogeneous data from the sensor column
geological sensors were placed inside a sensor column and and the data packets are transmitted to the upper layer. The
they were connected to the wireless sensor node via a data upper layer aggregates the data and forwards it to the sink
acquisition board as shown in Figure 1. node (gateway) kept at the deployment site.
Data received at the gateway has to be transmitted to the
Field Management Center (FMC) which is approximately
500m away from the gateway. A Wi-Fi network is used
between the gateway and FMC to establish the connection.
The FMC incorporates facilities such as a VSAT (Very
Small Aperture Terminal) satellite earth station and a
broadband network for long distant data transmission. The
VSAT satellite earth station is used for data transmission
from the field deployment site at Munnar, Kerala, South
India to the Data Management Center (DMC), situated at our
university campus 300 km away.
The DMC consists of the database server and an analysis
station, which performs data analysis and landslide modeling
and simulation on the field data to determine the landslide
probability. The wireless sensor network architecture for
landslide detection is as shown in Figure 2.
The Munnar region experiences frequent landslides and
has several landslide prone areas within every 1 sq km,
which can be utilized as future extension sites for landslide
detection systems. The different deployment sites can
connect to the FMC via a Wi-Fi network.

Figure 1: Sensor Column [8]

IV. WIRELESS SENSOR NETWORK ALGORITHMS

The WINSOC project has developed a totally innovative
design methodology, where the high accuracy and reliability
of the whole sensor network is achieved through a proper
interaction among nearby, low cost, sensors. This local
interaction provides more accurate distributed detection than

406
 every five minutes and sent the data to upper level sensor
nodes in the network.
A. Design of Interfacing Circuits
Indigenous interfacing circuits have been developed for
data acquisition from the geophysical sensors by the wireless
sensor nodes. The analog signals from the geophysical
sensors are amplified; level is shifted; the signal conditioning
has been performed according to each sensors requirement.
The details are shown in Table I.

TABLE I. Interfacing Circuit Requirements

Sensor Output Type Signal Pre-

Processing
Strain gauge Dual wire analog Level shifting,
piezometer Amplification
Vibrating wire RS-232 from the None
piezometer data logger
Dielectric Single wire None
moisture sensor analog
Tiltmeter Single wire Voltage reduction,
analog Amplification
Geophone Dual wire analog Level shifting,
Amplification

B. Design of Power Solutions

Figure 2. Wireless Sensor Network Architecture for Landslide Detection
Power circuits have been developed that provide constant
power for the excitation of the geophysical sensors and also
for powering the interfacing circuits. The power circuits
consists of voltage regulator and negative voltage converter
VI. FIELD DEPLOYMENT IC'S and provide different voltages from a lead acid battery
to the geophysical sensors and their interfacing circuits. The
Extensive field investigations were conducted for
lead acid batteries are automatically recharged by the solar
identifying the possible locations for sensor column
recharging unit.
deployment. At the deployment site, Anthoniar Colony, an
initial twenty sensor column locations consisting of 150 C. Design of Heterogeneous Networks
sensors total, were identified with respect to their geological The sensor column is physically attached to a wireless
relevance. sensor node which is integrated with a data acquisition
The pilot deployment consists of two sensor columns, board. The distance between current sensor columns is
with ten sensors, are deployed in the field along with six approximately 50 meters, at a slope of about 70°. Due to the
wireless sensor nodes. One of the sensor columns is terrain structure and vegetation, the data from the sensor
deployed at the toe region where various water seepage lines columns are not able to reach the gateway. The major reason
converge. This fact lead to the installation of pore pressure for this is no line of sight path between the columns, between
transducers at different depths (2m, 5m) and a dielectric the first sensor column and the gateway, and between the
moisture sensor, both sampled at every five minutes. second sensor column and the gateway. The above
Geophone is also deployed at sensor column 1 and sampled observations, along with experimental tests, have led us to
at the rate of 10 samples/second. The MicaZ wireless sensor employ three relay nodes in between the sensor columns
node connected to the sensor column transmits the digitized themselves and the gateway. One of the relay nodes is a
data values to the upper layers of the network. cluster head for the first and second sensor column. The data
The other sensor column is attached with movement from the cluster head is transmitted to the gateway in the
sensors since the location of it is in an unstable region. This form of packets. At the gateway the received packets are
sensor column has three tilt meters (1m, 2m, 3.5m), and three time stamped and stored.
strain gauges (1.5m, 2.5m, 4m) to capture the earth The wireless senor nodes used for the test deployment are
movement from the sensor column bending. A dielectric 2.4 GHz MicaZ motes from Crossbow. The gateway is a
moisture sensor is also connected to the sensor column at 1 single board computer which has 64 MB RAM, 32 MB flash
feet depth. The wireless sensor nodes sample these sensors at

407
and a fixed base mote that is used to send and receive the E. Design of Feedback System
messages through the transceiver. A feedback loop is incorporated for remotely
The Wi-Fi Network is used to transfer the data from the administering the sampling rate of the geological sensors
gateway to the FMC and it uses an external antenna and an with respect to real-time climatic variations, monitor the
access point for the same. The Wi-Fi network allows us to level of battery charges, monitor the level of solar charging
install the gateway in any scalable distance from the FMC. rate, indicate faulty wireless sensor nodes or geological
The Network has been tested with WLAN standards sensors etc.
802.11a/b/g.
The basic satellite communication network in the VII. CONCLUSION AND FUTURE WORK
landslide scenario is based on VSAT. The geological data
Real time monitoring of landslides is one of the
collected at the landslide deployment site is transmitted from
challenging research areas available today in the field of
the FMC at the deployment site to the DMC at Amrita
geophysical research. This paper discusses the development
University's Amritapuri campus, using the VSAT earth
of an actual field deployment of a wireless sensor network
station. The data is transmitted using UDP Protocol which
based landslide detection system. This system uses a
includes recovery of lost packets, corrupted packets, secure
heterogeneous network composed of wireless sensor nodes,
transmission, route via broadband during unavailability of
Wi-Fi, and satellite terminals for efficient delivery of real
VSAT, buffering the data to disk in case both networks are
time data to the data management center, to enable
unavailable and sending the data as soon as the network is
sophisticated analysis of the data and to provide landslide
connected, etc.
warnings and risk assessments to the inhabitants of the
D. Design of Data Analysis Software region. A test setup of this design has been already deployed
Data received at the DMC is being been analyzed using at Anthoniar Colony, Munnar, Idukki, Kerala, India. In the
the integrated landslide modeling software and the in-house future, this work will be extended to a full deployment by
designed data visualization software, which has the using the lessons learned from the existing network. This
capability to determine factor of safety of the mountain and network will be used for understanding the capability and
probability of landslide occurrence with respect to the signals usability of wireless sensor network for critical and
received from the deployed sensors. It also has the capability emergency application.
to compare and analyze data from different sensor columns, ACKNOWLEDGMENT
different sensors, selective comparison, etc. Data is
successfully received from the deployment site with minimal I would like to thank Mr. Sangeeth Kumar, Mr. Vijayan,
data packet loss and analysis of data has been performed. Mr. Joshua Freeman, Dr. Nirmala Vasudevan, Ms. Rehna
Data received from two pore pressure transducers and a rain Raj, Mr. Mukundan, and Mr. Kailash for their valuable
gauge is shown in Figure 3. contribution to this work.
During 2008 monsoon season, the sensors were able to
REFERENCES
capture the expansion and contraction of soil mass during
and after heavy rainfall. The data analysis software showed
respective variations in each of the deployed sensors. [1] P. K. Thampi, John Mathai, G Sankar, and S. Sidharthan, “
The data analysis software has been integrated with the Landslides: Causes, Control and Mitigation”, (based on the
investigations carried out by the Centre for Earth Science Studies,
capability of real streaming of data over internet. This Trivandrum)
facility has been tested and will be uploaded soon in the [2] H. Kung, J. Hua, and C. Chen,, “Drought forecast model and
webpage www.winsoc.org. The scientists around the world framework using wireless sensor networks, Journal of Information
can analyze the data with very minimal delay and effective Science and Engineering, vol. 22, 2006, pp. 751-769.
warning can be issued on time. [3] E. R. Musaloiu, A. Terzis, K. Szlavecz, A. Szalay, J. Cogan, and J.
Gray, “Life under your feet: A wireless soil ecology sensor network”,
2006.
[4] C. Hill and K. Sippel, “Modern deformation monitoring: A multi
sensor approach”.
[5] A. Terzis, A. Anandarajah, K. Moore, and I. J. Wang, “Slip surface
localization in wireless sensor networks for landslide prediction},
Proc. Information Processing in Senor Networks (IPSN'06), April 19-
21, 2006.
[6] E. A. Garich, “Wireless, automated monitoring for potential landslide
hazards”, Master Thesis,Texas A\& M University, May 2007.
[7] H. Lan, C. Zhou, C. F. Lee, S. Wang, and W. U. Faquan, “Rainfall-
induced landslide stability analysis in response to transient pore
pressure - A case study of natural terrain landslide in Hong Kong”.
[8] M. V. Ramesh, and P. Ushakumari, “Threshold based data
aggregation algorithm to detect rainfall induced landslides”, Proc.
The 2008 International Conference on Wireless Networks
Figure 3. Real-time Field Data (ICWN'08), CSREA Press, July, 2008, vol. 1, pp. 255-261,.

408
[9] S. Barbarossa and G. Scutari, “Decentralized maximum-likelihood [12] M. Maréchal, J. Pierrot, and J. Gorce, “Fine synchronization for
estimation for sensor networks composed of nonlinearly coupled wireless sensor networks using gossip averaging algorithms”, Proc.
dynamical systems”, IEEE Transactions on Signal Processing, July ICC 2008.
2007, vol. 55, no. 7. [13] R. M. Iverson, “Landslide triggering by rain infiltration”,Water
[10] S. Barbarossa, G. Scutari, and A. Swami, “Achieving consensus in Resource Research, July 2000, vol. 36, pp. 1897-1910.
self-organizing wireless sensor networks: The impact of network [14] G. Wang and K. Sassa, “Pore-pressure generation and movement of
topology on energy consumption”, IEEE, 2007. rainfall-induced landslide: Effect of grain size and fine-particle
[11] G. Scutari and S. Barbarossa, “Distributed consensus over wireless content”, Engineering Geology, 2003, vol. 69, pp. 109-125.
sensor networks affected by multipath fading”, IEEE Transactions on
Signal Processing, August 2008, vol. 56, no. 8.

409