0
Hopp til hovedinnhold

Disputerer på behandling av infodata i nødsituasjoner

Vimala Nunavath

Although lots of information is available, the information utilization by emergency responders is not optimal. The reason for this is that the available information is heterogeneously distributed, in different formats, with different semantics, stored in different emergency response organizations’ data sources, and technically not accessible to one another.

Vimala Nunavath

Ph.d.-kandidat / forsker

Ved en større ulykke, nødsituasjon eller katastrofe strømmer nødhjelpere fra mange forskjellige organisasjoner til. Informasjonsdeling er svært viktig, men kommunikasjonen vanskelig. Vimala Nunavath har forsket på bedre samhandling med bruk av IKT-verktøy.

Informasjonsinnsamlingen skjer ofte manuelt av hjelpemannskaper fra forskjellige etater og organisasjoner, og blir lagret i de respektive organisasjonenes systemer for informasjonsbehandling. De er ofte forskjellige, og kommuniserer ikke med alle de andres systemer. Alle deltakerne kan dermed ikke nyttiggjøre seg viktig informasjon.

Vimala Nunavath forsker på en modell som øker samforståelsen mellom alle hjelpegruppene. Det har tre komponenter: Ordforståelse, behandling av datamateriale og presentasjonen av datamaterialet.

Hun har fulgt doktorgradsprogrammet ved Fakultet for teknologi og realfag med spesialisering i IKT, og disputerer for ph.d.-graden mandag 30. oktober 2017 med avhandlingen Model-Driven Data Integration for Emergency Response.

Slik beskriver kandidaten selv essensen i avhandlingen:

Model-Driven Data Integration for Emergency Response

When any emergency happens, a complex network of emergency responders from various emergency response organizations such as police service, fire and rescue service, health care, municipality, military, and non-governmental organizations, work together in teams to carry out different tasks depending on the type of an emergency.

In order to perform different tasks, emergency responders need plenty of information to share with each other i.e., within or among teams in a timely manner. However, during any kind of emergency occurrence, enormous amount of information is generated and available at various places. This generated and available information is collected manually or semi-automatically by different emergency responders and then stored in their respective organizations’ data sources after the emergency occurrence.

Improve information accessibility

Although lots of information is available, the information utilization by emergency responders is not optimal. The reason for this is that the available information is heterogeneously distributed, in different formats, with different semantics, stored in different emergency response organizations’ data sources, and technically not accessible to one another. As a result, searching and finding the relevant information become time consuming. In such a situation, emergency responders face difficulties in obtaining a sufficient understanding of the emergency situation and poor decisions may be made. Therefore, in this dissertation, the research is focused on bringing the available heterogeneously dispersed information together to improve the information accessibility for emergency responders.

As the outcome of the research, a framework for solving the semantic heterogeneity problem to support various emergency responders for improving access to the needed and relevant information from different existing data sources has been developed. The framework consists of three main components i.e., the semantic model, the data source handling, and the information presentation. To test the applicability of the framework, an indoor fire emergency use-case scenario in a multiple storey building has been considered.

Building fire emergency response ontology

With the help of this case, a building fire emergency response ontology has been developed to explore the semantics of fire emergency response in a building domain. The developed building fire emergency response ontology enables data exchange and knowledge sharing among different emergency responders and across heterogeneous emergency response organizations’ data sources during the search and rescue operation. A semantically-enhanced mediator-based approach has been used to connect the existing data sources with the developed information model. To make the data accessible and available in a reliable way to ERs, web services have been utilized to present the relevant information on a graphical user interface. The evaluation of the prototype was done in a workshop session with participants from emergency rescue services.

Disputasfakta:

Kandidaten: Vimala Nunavath was born in India, in 1984. She received her M.S degree in ICT (Information and Communication Technology) with a system development specialization from the University of Agder, Norway in 2011 and B.Tech (bachelor in technology degree) in Computer Science and Engineering from Sagi Rama Krishna Raju Engineering College (affiliated to Andhra University), Bhimavaram, India in 2005. From feb.2013 to feb.2017, she was a PhD student in ICT at University of Agder. Her research interests are in the areas of domain modelling, data integration, knowledge management, semantic technologies, data analysis, artificial intelligence and machine learning. Since Sept 2017, she has started working as a senior researcher in Centre for Artificial Intelligence Research group at University of Agder, Campus Grimstad, Norway.

Disputas: Mandag 30. oktober 2017 kl 12:00

Dekan, professor Michael Rygaard Hansen, leder disputasen.

Prøveforelesning kl 10:00

Prøveforelesning og disputas finner sted i Rom C2 040, Campus Grimstad

Oppgitt emne for prøveforelesning: Ethical, legal and societal implications (ELSI) of emergency response systems

Tittel på avhandling: Model-Driven Data Integration for Emergency Response

Søk etter avhandlingen i AURA - Agder University Research Archive, som er et digitalt arkiv for vitenskapelige artikler, avhandlinger og masteroppgaver fra ansatte og studenter ved Universitetet i Agder. AURA blir jevnlig oppdatert. Avhandlingen vil være tilgjengelig til utlån ved Universitetsbiblioteket. Det vil bli også lagt ut noen eksemplarer av avhandlingen til utlån i lokalet hvor disputasen finner sted.

Opponenter:

Førsteopponent: Professor Peter Herrmann, NTNU

Annenopponent: Professor Frank Fiedrich, Bergische Universität Wuppertal

Bedømmelseskomitéen er ledet av førsteamanuensis Jan Pettersen Nytun, Institutt for IKT, UiA

 

Veiledere i doktorgradsarbeidet var professor Andreas Prinz, Institutt for IKT, UiA (hovedveileder) og professor Martina ComesInstitutt for IKT, UiA (bi-veileder)