Projects of Externally Funded Students

This page lists the projects that are funding the associated students within IRTG and illustrates how these projects and IRTG benefit from each other.

Alejandro Llaves (IFGI Muenster / MUSIL / ENVISION)

ENVISION overview: The ENVISION project provides an ENVIronmental Services Infrastructure with ONtologies that aims to support non ICT-skilled users in the process of semantic discovery and adaptive chaining and composition of environmental services. Innovations in ENVISION are: on-the-Web enabling and packaging of technologies for their use by non ICT-skilled users, support for migrating environmental models to be provided as models as a service (MaaS), and the use of data streaming information for harvesting information for dynamic building of ontologies and adapting service execution.

Association with IRTG research initiative B: Environmental monitoring deals with changes related to geospatial phenomena. In ENVISION, we are interested in extracting event-related information from time-series of observations. The focus of this research is on the integration of such information across different communities. The lack of standardized methods and models to process and represent environmental information causes interoperability problems when data sets are exchanged among several organizations. Diversity of data syntax or structure leads to syntactic interoperability. In the field of the Sensor Web, the Semantic Web Enablement (SWE) working group defines data models, encodings, and Web service specifications to overcome issues raised by syntactic heterogeneities. However, semantic heterogeneities present bigger challenges in this field, mostly caused by lack of application-specific knowledge, domain dependent perspectives, or multilingual settings.

Alejandro’s research: Environmental monitoring is an essential process in areas potentially affected by natural disasters. It is aimed to ensure public safety, to set up continuous information services and to provide input for spatial decision support systems. Here, the main challenge is the distributed processing of vast amounts of heterogeneous sensor data in real-time. Event processing tools allow to create an event abstraction layer on top of sensor data. Users can define event patterns to filter the information they are interested in and avoid irrelevant data. Extreme events are usually related to other occurrences, e.g. landslides are related (among others) to precipitation and earthquakes. To be able to determine whether an occurrence could potentially lead to an extreme event, domain knowledge is necessary. Ontologies are helpful for this task, since they are able to capture a representation of knowledge as a set of concepts and relations, within a specific domain. First, it is necessary to define the phenomenon occurrence we want to detect, and use an Event Pattern Language (EPL) to formalize it as an event pattern. Our domain ontology contains rules describing the potential consequences of such occurrence. When the event processing tool detects a relevant event, an instance is created in the domain ontology and the corresponding rules are triggered. This research work aims at combining event processing tools with semantic technologies to improve the integration of environmental data in decision support.

 

Auriol Degbelo (IFGI Muenster /MUSIL / DAAD & ENVISION)

The German Academic Exchange Service (DAAD) is a funding organization which supports the international exchange of students and scholars. I was granted a two-years funding by the DAAD to continue working on my PhD thesis. The International Research Training Group on Semantic Integration of Geospatial Information offers a good setting where the doctoral project can be conducted. Four points are worth mentioning:

  • international: the program in english is a great advantage for people not speaking german; the exchange period at the university of Buffalo enriches the international experience of the doctoral researcher;
  • interdisciplinary: there is an ample room for different perspectives on the research problems investigated because of the variety of backgrounds of the researchers involved in the program (computer science, geography, philosophy, cognitive science, geoinformatics);
  • regular monitoring of the progress: this enables the student to receive continuous feedback while advancing towards the completion of his doctoral thesis;
  • the idea of a cohort: working with other PhD students at the same moment on closely related topics provides a good platform for discussions and collaborations.

Auriol’s research: The goal of my PhD thesis is to provide a formal specification of concepts related to the spatial and temporal resolution of sensor observations. The PhD thesis fits therefore well with the research initiative B “Semantic mediation across communities and sensors” because formal specifications of concepts reduce the danger of miscommunication during the exchange of data across communities. A formal specification of resolution in particular is of paramount importance because resolution affects the patterns that can be detected during an analysis process. Drawing conclusion using a dataset at a resolution which is not appropriate for the phenomenon investigated results in errors referred to as cross-level fallacy.

 

Christoph Stasch (IFGI Muenster/STML/UncertWeb

UncertWeb overview: UncertWeb is an EC funded research project running from Feb 2010 – Jan 2013 with the goal to integrate tools for uncertainty propagation in the Model Web. The Model Web concept, formulated within the Global Earth Observation System of Systems activity, envisages the integration of complex resources, such as data  and models, to construct complex models, composed of chains of model and  data components exposed as web services. This offers exciting  opportunities for model development in a more loosely coupled, component  oriented manner, encouraging sharing, re-use and easy access of environmental resources in the Web. However, when combining services of limited, or unknown, quality, it is necessary to account for uncertainty if the outputs of the model web are  to be used for rational decision making. UncertWeb develops mechanisms, standards, tools and case studies to enable uncertainty management in an interoperable model web context.

Association with IRTG initiative C: The UncertWeb project is closely related to the initiative C “Semantics in Analysis and Reasoning”, as it focuses on uncertainties within the Model Web and thereby also helps building basic tools for the Model Web. The Model Web, as part of the Global Earth Observation System of Systems, can be seen as one mean to extract useful information from large spatio-temporal environmental datasets. In particular, the UncertWeb project also develops tools for spatio-temporal aggregation and for visualization of high-dimensional environmental data with uncertainties and hence contributes to an improved analysis and reasoning of spatio-temporal data.

Christoph’s Research: In the UncertWeb project, I’m involved in the technical working packages dealing with data models and representation and integration of several models in one common workflow. In particular, I’m focusing on spatio-temporal aggregation of observations and model outputs/inputs. Though aggregates in space and time are already available in the Web, for example weather portals, a flexible and on-demand spatio-temporal aggregation of observations in the Sensor and Model Web is currently not available. My research addresses this issue by (i) defining an technical architecture for web-based spatio-temporal aggregation, (ii) addressing the issue of meaningful spatio-temporal aggregation, and (iii) by propagating uncertainties during spatio-temporal aggregation. Hence, my research perfectly fits in the initiative C of IRTG, as it connects statistical models for aggregation with semantics about the observed process towards a meaningful aggregation.

Jae Hee Lee (Cognitive Systems Bremen & SFB/TR 8 Spatial Cognition, Project R3-[Q-Shape])

SFB/TR 8 Spatial Cognition overview: The interdisciplinary Transregional Collaborative Research Center Spatial Cognition: Reasoning, Action, Interaction has been established by the Deutsche Forschungsgemeinschaft (DFG) on 01 January 2003 at the Universities Bremen and Freiburg.
Spatial Cognition is concerned with the acquisition, organization, utilization and revision of knowledge about spatial environments, be it real or abstract, human or machine. Research issues range from the investigation of human spatial cognition to mobile robot navigation. The goal of the SFB/TR 8 is to investigate the cognitive foundations for human-centered spatial assistance systems.

The R3-[Q-Shape] project is part of the transregional collaborative research center SFB/TR 8 Spatial Cognition. Its goal is the development of qualitative high-level representation and reasoning methods for cognitive agents operating in a spatial environment and communicating about space. The project started in January 2003 and is currently in its third and last funding phase which will last until end of 2014.

Association with IRTG initiative C: Processing spatio-temporal information is one of the challenges in this research initiative. Once the sensed quantitative data is abstracted to qualitative data, high-level reasoning with this data is possible. Within the R3-[Q-Shape] project I work on devising efficient algorithms for reasoning with such qualitative spatial data.

Jae’s Research: Commonsense reasoning is a major challenge in Artificial Intelligence. As qualitative descriptions comprise a great deal of our everyday language, commonsense reasoning is concerned with developing and processing such qualitative representations. This applies especially to commonsense reasoning about space, where spatial objects are in particular characterized by their spatial relations to each other. Consequently, to endow spatial applications (e.g., CAD, GIS) with the commonsense reasoning ability, a thorough understanding of qualitative spatial relations is required. Qualitative Spatial Reasoning (QSR), a subfield of AI, approaches this very challenge. QSR aims at providing calculi which allow a machine to represent and reason with spatial relations using a finite set of symbols. In my thesis, I want to investigate qualitative spatial reasoning with directional information in regard to its reasoning properties. As reasoning with directional information, in particular with relative directions is intrinsically hard, I am investigating heuristic and approximative methods.

 

Jia Wang (IFGI Muenster/SIL/Sketchmapia)

SketchMapia overview : SketchMapia project (SketchMapia: A Framework for Collaborative Mapping) aims to develop a framework that contains the complete workflow of collection, recognition, interpretation, integration and visualization of sketch maps. In the context of Volunteered Geographic Information (VGI), SketchMapia employs sketch maps to contribute geographic information. This project develops a qualitative computational model to represent sketch maps in a computer-understandable way. SketchMapia integrates information from various sketch maps and metric maps into one data repository which can be queried by users via a query-by-sketch interface. Finally, spatial information from sketch maps is integrated with quantitative data to be represented on metric maps.

Association with IRTG initiative A: SketchMapia  project focuses on human strategies to obtain and organize spatial knowledge of local environment and aims to develop a framework for collaborative mapping using free-hand sketch maps. The aim of SketchMapia is perfectly aligned with the focus of initiative A : development of a new form of interaction with geospatial information by integration of qualitative and quantatitive representation formats used either by human or robots. The cooperations with other IRTG attendants as the former IRTG post-doctoral researcher Jan Oliver Wallgrün and Malumbo Chipofya have broadened the horizon in knowledge representation and reasoning for free-hand sketch maps; and Klaus Broelemann’s research on automatic understanding of sketch maps facilitates the data processing of sketch maps by valid sketching data recognition and extraction using computer vision methods.

Jia’s Research: As externalized representations of mental maps, sketch maps appear to have certain characteristics that already are well proved with ample evidence, i.e.,schematization structures and distortions on direction and distance (spatial), and size and shape (non-spatial). Take the inaccuracy into account during sketch map analysis, we attempt to explore relevant sketching aspects to be aligned with corresponding real-world configurations represented on planar metric maps. Qualitative sketching aspects indicating orientation, distance, topology, serial or cyclic order and etc are extracted and analyzed for their reliability and relevance for alignment. From obtaining such a list of sketching aspects, sketch maps can be used to identify cognized elements of the environment and further aligned and integrated with existing metric maps.

Activities and scientific outputs with IRTG funding:

  • Las Navas 20th Anniversary Meeting on Cognitive and Linguistic Aspects of Geographic Spaces (A full paper presentation) – IRTG scholarship
  • GIScience 2010 (A short paper and a paper at the Doctoral Colloquium) – IRTG scholarship
  • AGILE 2011 (A full paper presentation with its empirical experiment)- partially funded by IRTG
  • COSIT 2011 (Workshop organizer and main workshop proceeding editor). This workshop proposal was inspired by a group discussion of IRTG common senario during IRTG retreat at Borkum in Oct 2011. The workshop was cooperated with former IRTG funded post-doc Jan Oliver Wallgrün
  • IEEE Conference on Cognitive Informatics and Cognitive Computing 2012 (A full paper presentation and cooperated with Human Factors in GIScience Lab, GeoVISTA Center at Penn State University) – Mobility to the U.S fully funded by IRTG