Anda di halaman 1dari 6

CNTA ’09 Université A.

MIRA BEJAIA

Information Technology for Enhancing Disaster


Management
Nadia Nouali, Nadir Bouchama, Ahcène Bendjoudi, Abdelaziz Babakhouya, Said Yahiaoui, Yacine Belhoul, Houda
Zeghilet, and Nabil Guellati
Division de Recherche Théorie et Ingénierie des Systèmes Informatiques (DRTISI)
Centre de Recherche sur l’Information Scientifique et Technique (CERIST)
03 Rue des frères Aissou, Ben Aknoun, Alger, Algérie
{nnouali, nbouchama, abendjoudi, babakhouya, syahiaoui, ybelhoul, hzeghilet, nguellati}@cerist.dz

Abstract— Responding to natural or man-made disasters, in a consequences. A large destruction of buildings and
timely and effective manner, can reduce deaths and injuries as infrastructures was caused by earthquakes and a large
well as economic losses. Predicated on the assumption that number of victims were observed (varying from 60 to 3000).
better information leads to efficient decision-making and more
effective performance of crisis response, research projects Disaster management is the discipline of developing
applying advanced information technology solutions to the strategies for reducing the impact of disasters and for giving
crisis management field have emerged. This paper provides an assistance to the affected population. Disaster management
overview of most recent projects, in this area, all over the encompasses mitigation, preparedness, response, and
world. Furthermore, the study highlights that using scalable recovery efforts undertaken to reduce disasters impact [13].
and robust IT solutions can drastically facilitate access to the
right information, by the right individuals and organizations, Mitigation is the efforts to reduce the physical and social
at the right time. impact of future disasters. It includes building structures that
resist the physical forces of disaster impacts and efforts to
Keywords: Disaster management systems, Information decrease the exposure of human populations to dangerous
technology, Wireless communications, Information situations. Preparedness includes development, deployment,
management. and testing of disaster management systems. Response is the
direct intervention in the disaster area for the immediate
I. INTRODUCTION protection of life and property and minimizing the effects of
A disaster is a tragedy that negatively affects society or the disaster. Finally, recovery is the process and activities
environment. It may be natural (tornadoes, hurricanes, intended to ensure operation continuation of vital systems.
tsunamis, floods, earthquakes, etc.) or human-made (riots, Recently, the use of IT in disaster management has
terrorist attacks, war, etc.). The most striking examples of emerged in several research and development projects (use
recent disasters are the tsunami that struck Thailand on of databases, GIS, wireless and mobile technologies, etc.).
December 2004 and the September 11 World Trade Center The survey presented in this paper is motivated by the
attacks. Disasters result in loss of life and property, and initiation of a disaster management project by our research
disrupt economic activity, besides causing immense misery team and it is a step of the first phase of our project which
to the affected population. All existing infrastructures are goal is to observe actual events, learn lessons from the
suspected to be destroyed by the disaster, including responders and domain experts as well as existing literature.
communication infrastructures. Thus, interventions on
disaster areas are obviously made difficult.
A considerable growth has been observed during the last
few decades in the number of disasters. Fig. 1 shows the
number of disasters that occurred during the last twenty
years in the entire world. These statistics are made by the
WHO1 Collaborating Centre for Research on the
Epidemiology of Disasters (CRED) [16]. According to the
statistics collected by the CRED, Algeria is classed ninth
among the most affected countries in 2007. In the last fifty
years our country was struck by about 10 disasters of severe

1
World Health Organization Fig. 1. Number of disasters in the world since 1988
CNTA ’09 Université A.MIRA BEJAIA

The paper is organized as follows: in section 2, we give • Sensor applications for measuring and identifying
an overview of disaster management projects using IT, while environmental and potentially harmful factors that may
a brief discussion is presented in section 3. In Section 4, we affect the rescue operation, such as: temperature,
introduce the CERIST’s project on disaster management. humidity, rainfall, wind speed, etc;
Finally, section 5 concludes the paper.
• Face recognition application that allows a rescuer to
II. OVERVIEW OF DISASTER MANAGEMENT PROJECTS compare face images captured from the site to a
database of known faces.
In this section, we present six projects concerned with
disaster management, namely: DUMBO, WISECOM, DUMBONET is at the first stage of deployment.
RESCUE, InfoWare, Sahana, and DDT. Experiments have been done only with two isolated disaster
sites and a simulated headquarter.
A. DUMBO project
DUMBO [6][7][10] (Digital Ubiquitous Mobile Broad- B. WISECOM project
band OLSR) is an emergency network platform developed WISECOM [3] (Wireless Infrastructure over Satellite for
by three main research groups: AIT’s intERLab2 laboratory, Emergency COMmunications) is an ongoing project created
INRIA3 institute, and the WIDE Project4 team. They where by the German Aerospace Center and funded by the
motivated by the tsunami event in 2004 which devastated European Commission. The WISECOM project aims at
several areas in countries along shores of Indian Ocean, and developing a complete telecommunication solution that can
caused the breakdown of telecommunications infrastructure. be rapidly deployed immediately after a disaster. The
DUMBO is developed to provide multimedia WISECOM system intends to restore local GSM
communication among field team members and with a infrastructures so that normal mobile phones can be used,
distant command headquarter. It is designed for collaborative and to provide wireless data access over WiFi and WiMAX
simultaneous emergency response operations deployed in a using satellite communication. The system incorporates also
number of disaster affected areas. location based services for the purposes of locating victims
The architecture of DUMBONET (Fig. 2) combines and rescue teams.
mobile ad hoc networks (MANET) and a satellite IP The WISECOM architecture includes one of the two
network. A MANET is deployed on each isolated disaster particular European portable satellite systems, namely:
site and satellite accesses allow multimedia communication Inmarsat, BGAN5 and DVB-RCS6. Wireless local access
between different sites and with the distant command points are also used for enabling the emergency personal
headquarter [1] [5]. and/or victims to access the network using standard WiFi
enabled devices (laptops, PDAs, WiFi phones, etc.). The
We distinguish three categories of bidirectional
local WiFi hotspots are deployed around the vehicles to
communications: intra-site, site to headquarter, and site to
provide wide area coverage, up to 1 kilometer, to the
site communications. Site to site traffic must pass through a
rescuers within a WiFi cell. The WiFi hotspots are in turn
terrestrial satellite gateway.
connected to the satellite access point using 802.16d
Three main applications are deployed on DUMBONET: WiMAX links over a radius of up to 10 kilometers. All of the
required equipments can be rapidly transported to the
• Multimedia applications including video, voice, and disaster site in a normal car or as standard luggage on a
short messages;

Fig. 2. Architecture of DUMBONET Fig. 3. Architecture of WISECOM


2
http://www.interlab.ait.ac.th/ (Thailand)
3 5
http://www.inria.fr/index.en.html (France) http://broadband.inmarsat.com
4 6
http://www.wide.ad.jp/ (Japan) http://www.dvb.org
CNTA ’09 Université A.MIRA BEJAIA

plane. The WISECOM architecture is illustrated in Fig. 3.


The first experimental test of the WISECOM system has
been performed in March 2008. This test validated the
network functionalities, and allowed to verify the correct
operation of applications such as file transfer, web browsing,
VoIP, videoconferencing and video streaming.
C. RESCUE project
The goal of the RESCUE project [8] (RESponding to
Crises and Unexpected Events) is to radically transform the
ability of responding organizations to gather, manage, use,
and disseminate information both within emergency
response networks and to the general public. With more Fig. 4. Architecture of Disaster Portal System
robust information systems, response can be prioritized, and
focused to activities that have the highest potential to save • Privacy is concerned with determining "best practices"
lives and property. (minimal data collection, limiting information
disclosure/inference, establishing clear policies for
Such a radical transformation requires a multidisciplinary information collection/use/sharing, etc.), exploring how
approach that recognizes that while information technology such practices can be realized technologically (policy
is paramount. A thorough understanding of how emergency languages, enforcement mechanisms, information hiding
organizations form and work together in crisis situations is techniques such as data perturbation, anonymization,
vital to the solution. Based on this understanding, the team etc.), and studying how technological innovation
proposes a research program that consists of five major influences technology adoption.
multidisciplinary research projects:
Test beds have been created for evaluating RESCUE
• Situation awareness (SAMI) which goal is to design and research and several products have been already developed
develop technologies that can create actionable such as: the Crisis Alert System to disseminate information
situational awareness from the heterogeneous multi to schools and other organizations in case of disasters, the
modal data streams (audio, speech, text, video, etc.) Disaster Web Portal that provides a wide range of real-time
including human-generated input (e.g., first responders’ information in disaster situations, such as situation
communications, field reports, etc.) during or after a summaries, announcements, shelter information, and
disaster. SAMI undertakes an event-oriented approach aggregated services such as family reunification and
to building situational awareness. donation management (Fig.4). A first version of this portal
• Information sharing (PISA) which objective is to was deployed by the City of Ontario in September 2007. The
understand data sharing and privacy policies of current system serves as a base to develop and refine results
organizations and individuals; and design, develop, and from several areas of research which are being incorporated
evaluate a flexible, customizable, robust, scalable, into the existing system to provide additional or advanced
policy-driven architecture for information sharing. capabilities [20].

• Robust communications (ENS) which objective is to D. Ad-Hoc InfoWare project


develop systems that provide computing, The Ad-hoc InfoWare (Middleware Services for
communication, and higher layer services at a crisis site. Information Sharing in Ad-hoc Networks) project is created
The goal is to develop a system that can operate under by the Research Council of Norway7 and the Faculty of
extreme conditions by consolidating and enhancing Mathematics and Natural Sciences8 at the University of Oslo.
available systems and seamlessly extending new The project [11][15] implements a Middleware for
capabilities to all end users and devices as information sharing in sparse ad hoc networks which use
communication services get incrementally restored. efficiently the available infrastructure in a rescue operation.
It gives analysis of the different organizations structures and
• Information dissemination focuses on information that is the various interactions during a rescue operation. Also, it
disseminated to the public at large specifically to identifies the technical challenges imposed by a such highly
encourage self-protective actions, such as evacuation dynamic environment. The Middleware framework is
from endangered areas, sheltering-in-place, and other composed of six components:
actions designed to reduce exposure to natural and
human-induced threats. The grand challenges that are • Data Management: is a distributed database;
addressed include building highly scalable, reliable and
timely dissemination services from unstable and • Knowledge Management: to support the dissemination,
unreliable resources using a peer-based architecture for sharing and interpretation of ontologies;
both wired and wireless dissemination. 7
http://www.forskningsradet.no
8
http://www.matnat.uio.no
CNTA ’09 Université A.MIRA BEJAIA

• Context Management: to manage context models,


context sharing, profiling and personalization;
• Communication Infrastructure: to supporting distributed
event notification, publish and subscribe services, and
message mediation;
• Resource Management: to enable resource sharing
among the devices involved in the network.
• Security Management: to secure all the other
components using some mechanisms such as: access
control, message signing and encryption; Fig. 7. Components Diagram of Sahana [4]
The Middleware has mechanisms for predicting of requirements and to understand operational challenges
partitioning (based on GPS or neighborhood information). during a disaster. Sahana has been deployed and/or tested in
Data replication mechanisms, distributed directories must be many countries including: Sri Lanka, Philippines, Indonesia,
implemented, and key management group are also supported China, United States of America, Australia, Lebanon, Spain,
by the middleware. Portugal, Germany, Poland, United Kingdom, South Africa,
Ecuador, Argentina, Sudan, etc. These efforts led to great
E. Sahana FOSS Disaster Management System improvements in Sahana capability and robustness after
The Sahana [9] project grew up in Sri Lanka just after the each deployment. The main lessons learned from these
December 2004 tsunami that hit more than 12 countries in experiments are summarized in [4] and [14].
Asia. It immediately gained widespread attention from
developers and humanitarian consultants worldwide. Sahana F. DDT project
aims to develop an integrated set of pluggable, web-based DDT Project [17] on rescue robots and related
disaster management applications that provide solutions to technologies is a special project for earthquake disaster
large-scale humanitarian problems in the relief phase of a mitigation in urban areas, launched by the Japanese MEXT9
disaster. The system is built on the stable FOSS (Free and in 2002. It was managed by a non-profit organization,
Open Source Software) technology stack, AMP (Apache, International Rescue System Institute (IRS)10, and more than
MySQL, PHP/Perl). The components diagram of Sahana is hundred researchers and students have contributed during
given in Fig. 7. five years. The aim of the project is practical development
of technologies related to robotics against earthquake
TABLE I. MAIN APPLICATION MODULES IN SAHANA. disasters, which include robot systems, intelligent sensors,
Module Goal
information equipment, and human interfaces, that support
emergency response activities (urban search and rescue,
Missing Person Helping to reduce trauma by effectively finding missing information gathering, and communications). Typical
Registry persons
technologies include teleoperated robots for victim search in
Organization Coordinating and balancing the distribution of relief hazardous disaster area, and robotic systems with distributed
Registry organizations in the affected areas and connecting relief sensors for gathering disaster information to support human
groups allowing them to operate as one
decision making. The DDT Project consisted of the
Request Registering and tracking all incoming requests for support following four Mission Units (MU), organized as research
Management and relief up to fulfilment and helping donors connect to groups. The objective of research of each MU is as follows.
System relief requirements
Camp Registry Tracking the location and numbers of victims in the • Aerial Robot System Mission Unit (ARS):
various camps and temporary shelters setup all around the Intelligent helicopters, balloons, image processing, and
affected area human interface;
Volunteer Coordinate the contact info, skills, assignments and • Information Infrastructure System Mission Unit (IIS):
Management availability of volunteers and responders
Distributed sensors, RFID tags, integration protocol,
Inventory Tracking the location, quantities, expiry of supplies stored database, and mapping;
Management for utilization in a disaster
• In-Rubbles Robot System Mission Unit (IRS): Serpentine
Situation Providing a GIS overview of the situation at hand for the
Awareness benefit of the decision makers robots, advanced rescue tools, advanced search cams,
advanced fiber scopes, sensors, and human interface;
The system is available and distributed under the GNU
Lesser General Public License (LGPL). The main • On-Rubbles Robot System Mission Unit (ORS): Tracked
applications built into Sahana and the problems they address vehicles, jumping robot, ultra-wide-band radar, semi-
so far are as summarized in Table I. autonomous movement, ad-hoc communications, self
The Sahana development team is giving more and more 9
priority to real deployments in order to obtain a complete set Ministry of Education,Culture, Sports, Science and Technology
10
http://www.rescuesystem.org/
CNTA ’09 Université A.MIRA BEJAIA

disaster management systems are developed or currently


under development [18].
Based on the analysis briefly presented in this paper, we
can see that there is a growing recognition by either research
community, governments and private institutions that a
disaster management system and more specifically one that
uses mobile and wireless IT is useful to minimize the impact
of disasters [13][19]. As a result of this recognition, efforts in
developing disaster management systems and applications
which make use of IT are increasing.
IV. THE CERIST'S PROJECT ON DISASTER MANAGEMENT
Mindful that Algeria is, in a high probability, subject to
Fig. 8. Integration of gathered data from robots and systems
disaster scenarios, and wanting to use our background in IT
using Mitigation Information Protocol (MISP) into DaRuMa and particularly mobile and wireless IT, we offered to
database
explore opportunities to apply them to emergency
localization and mapping, human interface, and sensor management of disaster relief. We started our disaster
data processing management project in early 2008. We do not plan to support
all aspects of the so complex and multidisciplinary problem
Before an incidence, IIS systems continuously monitor of disasters management, but to provide tools that can help
the situation in houses. Right after the incidence (large-scale concerned persons (managers of the crisis and the teams
urban earthquake disaster occurs), residents’ information involved in the field) to capture the dynamic realities of a
which has been gathered by distributed sensors, Rescue catastrophe more clearly and help them make better
Communicators of IIS, is transferred to disaster response decisions more quickly to facilitate their operations. Our
organizations immediately after receiving Earthquake Early methodology for conducting research within our disaster
Warning (EEW)11 before the shake. The intelligent ARS management project is based on three main elements: (i)
systems automatically fly to gather overview information of observe and learn lessons from actual and past events,
the affected area. The robots of ORS are brought by first domain experts and existing literature, (ii) identify research
responders to the disaster site to collect victim information needs and priorities in crisis response, (iii) finally, develop
and to investigate structural damage and hazardous materials. and nurture a research agenda that ensures the transfer of
All the gathered data are integrated into a distributed solutions to end-users community. Currently, we are
database DaRuMa via XML-type standardized protocol studying and testing the deployment of the Sahana open
MISP (Mitigation Information Sharing Protocol), and can be source platform on a wireless ad hoc network for rescue
referred and searched by SQL commands and viewers such situations. On the other hand, we are dealing with two VoIP
as Google Earth as shown in Fig. 8. The information in the deployment issues, which are of paramount importance in
database can be attributed, added and processed afterwards the context of mesh networks and MANETs, namely:
via Internet. Such information would be helpful to improve decentralizing SIP (Session Initiation Protocol) and
efficiency of decision making [2].In order to test the conceiving new QoS aware routing protocols. Solving such
developed systems and technologies, several experiments, problems can help to offer sustainable multimedia services in
demonstration and exercise were performed. Firefighters in the case of a disaster.
active service organized a volunteer unit, and made intensive
testing and demonstrations to evaluate research results.
V. CONCLUSIONS
III. DISCUSSION
The survey presented in this paper is a step of the project
A variety of information technologies such as networks, first phase that aims at making a state of the art on the
mobile and distributed systems, databases, data analysis and advancements achieved in disaster management using IT.
mining, image processing, security, decision-support tools, The objective is twofold. First, it allows us to better
etc. are incorporated in the research activities with the understand the various facets of the project and channel our
objective of revolutionize the ability to gather, manage, efforts in organizing our work into objectives targeting (a)
analyze and disseminate information in crisis response. In rapid development of applications for which the technology
Table II, we summarise the principle characteristics of the has reached maturity and (b) more or less long term research
projects presented in this paper. However, we note that other on still open issues. Second, it shows that our initiative is
sound, and can convince the various actors involved in crisis
management, including government, to join the project.
11
EEW system provides advance announcement of the estimated Indeed, the ultimate goal is to implement a technological
seismic intensities and expected arrival time of principal motion platform, placed under government control, which offers key
using difference of speed of the primary wave and the secondary services for emergency management.
wave of earthquake.
CNTA ’09 Université A.MIRA BEJAIA

TABLE II. SUMMARISATION TABLE OF DESCRIBED PROJECTS


Project Type Period Country Minimal Equipment needed Deployment in real situations
DUMBONET Wireless 2006-2007 Thailand WiFi and satellite equipments Yes. Experiments with two
communication isolated sites and simulated
headquarter
WISECOM Wireless 2006-2008 Germany GSM, WiFi backhauling over satellite, No
communication WiMAX
RESCUE Software 2003 up to USA Web server, Ad-hoc equipments, wireless Yes. Tsunami Southeast Asia
Hardware now communications (2004) & Southern California
wildfires (2007)
Ad-hoc Middleware 2003-2007 Norway (Oslo Ad-hoc equipments No
InfoWare university)
Sahana Web 2004- still in Sri Lanka Web server Yes
application progress
DDT Robotics, 2002-2006 Japan Robot systems, intelligent sensors, Yes. Yamakoshi town, Niigata-
related information equipment, and human interfaces Chuetsu earthquake (2005)
technologies

REFRENCES
[1] I. F. Akyildiz and X. Wang, “A Survey on Wireless Mesh Networks”, [12] T. Plagemann, E. Munthe-Kaas, K. S. Skjelsvik, M. Puzar, V.
IEEE Radio Communications, pp. 523-530. Sep. 2005. Goebel, U. Johansen, J. Gorman, and S. P. Marin. A Data Sharing
[2] H. Asama, Y. Hada, K. Kawabata, I. Noda, O. Takizawa, J. Meguro, Facility for Mobile Ad hoc Emergency and Rescue Applications. In
K. Ishikawa, T. Hashizume, T. Ohga, K. Takita, M. Hatayama, F. Proceedings of the First Intl Workshop on Specialized Ad Hoc
Matsuno, S. Todokoro: "Information Infrastructure for Rescue Networks and Systems (SAHNS’2007), Toronto, Canada, June 2007.
Systems", Workshop on Rescue Robotics - DDT Project on Urban [13] R. R. Rao, J. Eisenberg, and T. Schmitt. Improving Disaster
Search and Rescue, 2007 IEEE/RSJ Intl Conference on Intelligent Management: The Role of IT (Information Technology) in
Robots and Systems, pp. 45-56, San Diego, USA, Oct. 2007. Mitigation, Preparedness, Response, and Recovery. National
[3] M. Berioli, N. Courville, and M. Werner. Integrating Satellite and Academies Press, Washington, USA, 2007.
Terrestrial Technologies for Emergency Communications: The [14] Nah Soo Hoe. “Managing Disasters: Sahana, Sri Lanka.” In Breaking
WISECOM Project. In Proceedings of the Int. Conf. on Barriers:” The Potential of Free and Open Source Software for
Heterogeneous Networking for Quality, Reliability, Security and Sustainable Human Development; ed. Nah Soo Hoe, Bangkok: Asia-
Robustness, August 2007. Pacific Development Information Program, Regional Center in
[4] M. Careem, C. D. Silva, R. D. Silva, L. Raschid, and S. Bangkok, United Nations Development Programme; New Delhi:
Weerawarana. Sahana: Overview of a Disaster Management System. Elsevier, pp: 56–63. 2006.
In IEEE Intl Conference on Information and Automation, 2006. [15] N. C. Sanderson, K. S. Skjelsvik, O. V. Drugan, M. Puzar, V. Goebel,
[5] T. Clausen and P. Jacquet, “Optimized Link State Routing Protocol E. Munthe-Kaas, and T. Plagemann. Developing Mobile Middleware
(OLSR)”, available at http://www.ietf.org/rfc/rfc3626.txt, Oct 2003. - an Analysis of Rescue and Emergency Operations. Technical Report
358, Dept of Informatics, University of Oslo, Norway, June 2007.
[6] K. Kanchanasut, A. Tunpan, M. A. Awal, D. K. Das, T.
Wongsaardsakul, and Y. Tsuchimoto. DUMBONET: a Multimedia [16] J-M. Scheuren, O. le Polain de Waroux, R. Below, D. Guha-Sapir,
Communication System for Collaborative Emergency Response and S. Ponserre. Annual Disaster Statistical Review: The Numbers
Operations in Disaster Affected areas. International Journal of and Trends 2007. Catholic University of Louvain, Belgium, 2007.
Emergency Management, 4(4):670–681, 2007. Catholic University of Louvain, Belgium, 2007. Free access to the
CRED database is available at http://www.emdat.be.
[7] K. Kanchanasut, A. Tunpan, M. A. Awal, D. K. Das, T.
Wongsaardsakul, and Y. Tsuchimoto. “A Multimedia [17] S. Tadokoro, F. Matsuno, H. Asama, M.Onosato, K. Osuka, T. Doi,
Communication System for Collaborative Emergency Response H. Nakanishi, I. Yokokohji, M. Murata: "DDT Project: Background
Operation in Disaster-affected Areas”, Technical Report No. and Overview", Workshop on Rescue Robotics - DDT Project on
TR_2007-1, Internet Education and Research Laboratory Asian Urban Search and Rescue, IEEE/RSJ07 Intl Conference on Intelligent
Institute of Technology, Thailand, January 2007 Robots and Systems, pp. 1-22, San Diego, USA, October (2007).
[8] S. Mehrotra, C. Butts, D. Kalashnikov, N. Venkatasubramanian, R. [18] F. Souza and I. Kushchu, Mobile Disaster Management System
Rao, G. Chockalingam, R. Eguchi, B. Adams, and C. Huyck. Project Applications - Current Overview And Future Potential –Proc. Of the
RESCUE: Challenges in responding to the unexpected. SPIE Journal First European Mobile Government Conference MGOV2005,
of Electronic Imaging, pages 179–192, 2004. Brighton, UK, 10-12 july, pp. 455-466.
[9] Official site of the Sahana project: http://www.sahana.lk [19] Committee on Information and Communications Technology,
Information and communications technology-enabled disaster risk
[10] Official site of the DUMBO project: reduction in Asia and the Pacific, E/ESCAP/CICT/2, United Nations
http://www.interlab.ait.ac.th/dumbo/index.php Economic and Social Commission for Asia and the Pacific (ESCAP),
[11] T. Plagemann, V. Goebel, C. Griwodz, and P. Halvorsen. Towards First Session, 19 to 21 November 2008, Bangkok, Thailand.
Middleware Services for Ad hoc Network Applications. In [20] J. Lickfett; N. Ashish, S. Mehrotra, N. Venkatasubramanian and J.
Proceedings of the 9th IEEE Workshop on Future Trends of Green. “The RESCUE Disaster Portal for Disasters and Emergency
Distributed Computing Systems, pages 249–257, May 2003. Response”, Proceedings of the 5th International ISCRAM
Conference, Washington DC, USA, 2008.

Anda mungkin juga menyukai