Research area 1: Context Sensing and Analysis.
To top
Prof. Kirste: Models and Description of Physical Environments. The task of an ensemble is to behave "meaningfully", or "optimally" with respect to the needs of the affected user. Thus, it is necessary to develop a precise definition describing the meaning of "optimal", or "meaningful" respectively, in the context of ensemble's behaviour. Also the need for a normative theory of the world arises, where the necessary variables are identified and on the basis of the relevant situation and the optimal behaviour an assemble is defined. The stress of our research is on the user support when interacting with the real world. Thus aspects such as place, orientation, lightness, loudness, media perceptibility, have a central meaning and are the basis for formulating concepts such as "the best perceptible display". This model allows the possibility of defining an ideal ensemble behaviour for a given situation. From this arises the challenge to find the global ideal based on the possible local informations and interact.
Ph.D. student (01.11.2006-31.10.2010): Christoph Burghardt
Ph.D. student (15.05.2010-14.04.2012): Kristina Yordanova
To top
Prof. Tavangarian: Reliable, context-based device-sensor-networks in horizontal and vertical wireless communication architectures. The goal of this subproject is to provide a flexible device communication, which shall allow a reliable, context-based data transfer based on any wireless or wired communication technology, with respect to quality of service requirements of the mobile devices and the sensors of the environment. Thus, self-configurable horizontal and vertical communication links are established to enable a resource-friendly information exchange between sensors, mobile devices and central infrastructure such as system- and application servers. Context is taken into account for application execution as well as for the selection of available communication channels.
Ph.D. student (15.10.2006-14.01.2010): Enrico Dressler
Ph.D. student (01.08.2010-31.12.2012): Alexander Gladisch
To top
Research area 2: Multimodel Interaction and Visualization.
To top
Prof. Schumann: Visual Interfaces in smart ensembles. This project aims at developing and implementing an approach for the generation of visual interfaces in heterogeneous, adhoc environments. For this purpose specifically two aspects have to be investigated: a dynamic selection of information to be presented and moreover, an adaptive choice of appropriate visual methods to adequately represent the data at multiple devices. A particular challenge is considering dynamically changing context parameters to allow for an adaptive information presentation (e.g. a new topic, changing tasks or users can lead to new information that has to be selected and appropriately presented).
Ph.D. student (01.10.2006-31.12.2009): Conrad Thiede
Ph.D. student (01.10.2009-30.09.2011): Axel Radloff
To top
Prof. Forbrig: Usability tests in mobile environments. Software usability tests of mobile applications are a special challenge since testing under laboratory conditions is not sufficent. However, even laboratory tests are more difficult than normal ones. The aim is to develop a methodology which is able to record and analyse the activities of participating users without influencing their behavior. Therefore, it is neccessary to analyse the actual states of model instances that are based on meta models.
Ph.D. student (01.11.2006-31.01.2010): Stefan Propp
PH.D. student (01.06.2010-31.05.2011): Rene Zilz
Prof. Staadt: Interaction in Large High-Resolution Display Environments. Nowadays large high-resolution displays are used in a wide application area like product engineering, geospatial imaging or scientific visualization. The advantage of scientific visualization on high-resolution displays is the presentation of complex data in a higher level of detail as well as in the context of surrounding information. In the field of smart ensemble and multimodal appliances the usage of and the interaction with these display environments have not been sufficiently investigated yet. The topic for this PhD project is the development of methods and techniques for integrating large high-resolution display environments in smart appliance ensembles. The thesis focuses on the investigation of multi-modal 3D interaction techniques. The display environment can be used as an (inter-) active or passive device, depending on the user's current position, task, and workload.
PH.D. (01.01.2010-31.12.2011): Anke Lehmann
To top
Research area 3: Intention Recognition and Strategy Generation.
To top
Prof. Uhrmacher: Synthesis of services for user assistance in ad-hoc, multimedia environments. The objective is, to synthesize emergent services for assisting users while devices are spontaneously cooperating with each other in an ensemble. Planning techniques play an important role for a goal-oriented and coordinated composition of services according to the user’s preferences and the current context. In this regard, options, goals, and preferences have to be weighted up dynamically. Therefore, a flexible approach should be chosen taken the dynamic character and heterogeneity of the ensemble into account. A mobile broker should be developed, which identifies potentially new services for assistance in a bottom up manner using an ontology, an environmental model, preferences, planner, and an inference system. Thus, the composition of services is in the focus of this research.
Ph.D. student (1.12.2006-28.02.2010): Florian Marquardt
PH.D. student (15.12.2009-14.12.2011): Alexander Steiniger
To top
Prof. Kirste: Spontaneous cooperation of multimedia devices. By developing suitable cooperation techniques, multimedia appliances shall be enabled to adopt spontaneously and in a situation specific manner according to users’ goals. The latter belongs to the development of synthesized assistance services by Prof. Uhrmacher. Ad-hoc ensembles should be able to reproduce the behavior of dedicated intelligent environments. Furthermore, it should be ensured that every appliance is able to estimate only its own contribution to the collective goal without having any a priori knowledge about possible goals.
Thus, the main focus is on development of distributed cooperation strategies for dynamic application domains particularly with regard to ensembles with causal functional dependencies between the ensemble’s participants. In this process, strategies were developed which are able to solve these problems without a learning phase. These studies are complemented by Prof. Salomon’s research.
Ph.D. student (01.02.2007-31.01.2010): Christiane Plociennik (geb. Reiße)
Ph.D. student (01.05.2010-30.04.2012): René Leistikow
To top
Prof. Salomon: Global cooperation through self-organization, learning and artificial evolution. Smart-appliances ensembles consist of intelligent devices that interact with each other and that are supposed to support their users in autonomous, non-invasive way. Since both the number and the composition of the participating devices may spontaneously change at a time without any notice, traditional approaches, such as rule-based systems and evolutionary algorithms, are not appropriate mechanisms for their self-organization. Therefore, a new evolutionary framework is needed that accounts for the inherent system dynamics by distributing all data structures and all operations across all participating devices.
Ph.D. student (01.10.2006-28.02.2009): Ulf Ochsenfahrt
Ph.D. student (01.03.2009-28.02.2011): Enrico Heinrich
To top
Research area 4: Data Management, Resources, and Infrastructure Management.
To top
Prof. Cap: Content-Decoupled Communication in Heterogeneous, Mobile Network Topologies. Ad-hoc smart environments are created around the user by communication and collaboration of all devices that surround the user in a given environment. There are applications on these devices that need to share information and therefore act as message producers and applications that need information and act as message consumers. The main challenges in such environments are their heterogeneous nature and their mobility. These environments are made up of devices utilizing different, paritally incompatible or administrative separated communication technologies and they change rapidly. Therefore this project investigates new means of decoupled communication that joins messages publishers and messages subscribers even though they are separated by different communiction technologies, situated in a rapidly changing and unreliable communication topology and the messages that are published need to be processed before they can be understood by their subscribers. Therefore other devices are automatically included into the communication to bridge incompatible communicatino technologies or translate and/or process messages.
Ph.D. student (01.10.2006-30.09.2010): Henry Ristau
Ph.D. student(01.02.2010-31.01.2012): David Gaßmann
To top
Prof. Heuer: Query Managment in Ad-Hoc Environments. Obviously P2P solutions are able to manage unstructured data sets or binary data in general. To adress the aspect of data management at MuSAMA two facts have to be considered: Complex data objects and therefore complex queries have to be administrated regarding the fact of mobility. As complex queries can be decomposed subqueries have to be analysed on different devices at the same time. On the other hand we also have to look at different query types i.e. queries on structured data sets or IR like queries. So we have to focus on query processing solutions and data exchange mechanism which are able to adress the dynamic of ad-hoc environments. Finally real time test and simulations are essential to evaluate efficiency and effectivity of our solutions.
PH.D.: (01.10.2008-30.09.2011) André Peters
vacant: 01.10.2010
To top
Prof. Cap: Confidentiality Models in Smart Appliance Ensembles. Environments which store and exchange interactions and context for smart services need to be open for the ubiquitous access of service providers and sensors components which participate in an ad-hoc fashion. At the same time such environments contain very sensible information, raising various security and privacy related issues. Consequently, a trust concept for such context storing spaces has to be developed. For this purpose, the storage space has to be divided into different layers (e.g. raw sensor data, derived context, device interaction), each implemented with a corresponding pseudonym scheme. Individual layers are only accessible for specific purposes within the overall environment architecture, additionally guarded by sandbox approaches. Various previous work in the field of secure ad-hoc device interactions can be used as a basis for this problem, especially approaches to profile based security management. Security requirements of applications and the level of trust components have towards the ensemble strongly depend on the current context of a user, e.g. activity and location. Additionally, an exact modeling of security problems often fails in practice. Stochastic approaches and methods of fuzzy reasoning, already proved to work in practical applications, are more likely to succeed.
Ph.D. student (1.2.2007-31.1.2010): Christian Bünnig
Ph.D. student (1.3.2010-28.2.2012): Till Wollenberg
To top
Prof. Tavangarian: Context-aware Device Communication. Context-aware applications need to be enabled to react on changes in real world. A learning scenario is for instance changed if two students meet up with each other in the library. The event “students meet up” is a situation caused process with relation to the real world. This can be monitored by evaluating sensor data (e.g. GPS data) and needs to be communicated to the students. In opposition to a sender- or receiver-initiated communication the sender and receiver are determined by the event itself, so it is a event-initiated communication. Known event-driven services are primary used for static network structures. The definition and creation of suitable abstractions for the registering of events is essential for situation-aware event monitoring - independent of the concrete sensor device and data processing. A solution approach for these aspects is offered by a system with an abstract model of the communication environment, which handles the context-aware capture and processing of events for specified scenarios. This system considers the situation of users in context-aware applications within real environments.
In the present task definition an approach towards the communication of nodes based on abstract models of the environment and scenarios is followed. Particularly aspects like heterogeneity of the network infrastructure, monitoring and semantics of events as well as QoS requirements are considered. The system represents a logic communication network consisting of heterogeneous connection structures between the appliances with different communication channels. Within the network the communication architectures for efficient and reliable event monitoring and processing are designed.
Ph.D. student (1.2.2007-31.1.2010): Raphael Zender
Ph.D. student (15.1.2010-14.1.2012): Philipp Lehsten
To top
