zur Startseite

Dr. rer. nat. Frank Dürr: Projekte

Current Projects

Com'N'Sense

Funded by: German Research Foundation (DFG)

The Com'N'Sense Project investigates how mobile phones and other mobile devices can be utilized to gather sensor data by crowds of mobile users (crowd sensing, mobile phone sensing, public/participatory sensing). The basic goal is to deliver a certain quality of sensor data while minimizing the energy consumption of mobile devices spent for sensing, communication, and processing.

With partners from photogrammetry, we investigate how to automatically generate indoors floor plans and 3D indoor models from crowd-sensed sensor data.

[Project Homepage]

PriLoc — Privacy-aware Location Management

Funded by: German Research Foundation (DFG)

The PriLoc Project investigates concepts for the secure management of private position information on non-trusted server infrastructures such as public clouds. In particular, we designed a novel position sharing approach with interesting properties such as no single point of failure w.r.t. privacy, graceful degradation of privacy with the number of compromised servers, and support of different, individual precision levels for location-based applications with different quality demands and trust levels.

[Project Homepage]

Integrated Controller Design and Communication Services for Digital Control Networks 

Funded by: German Research Foundation (DFG)

Together with partners from control theory, we investigate how to tightly integrate the design of networked control systems and the underlying communication services to guarantee a certain quality of control while optimizing the usage of required network resources.  

[Project Homepage]

Time-sensitive Software-defined Networks for Manufacturing

Funded by: German Research Foundation (DFG) within the Graduate School of Excellence advanced Manufacturing Engineering (GSaME)

In this project with investigate concepts for implementing IEEE 802 networks with deterministic (hard) time bounds to support real-time tasks in manufacturing environments. To this end, we integrate software-defined networking (SDN) and time-sensitive networking (TSN) methods.

[Project Homepage]

Simulations on Mobile Devices

Funded by: German Research Foundation (DFG) within the Cluster of Excellence Simulation Technology (SimTech) 

Novel mobile applications like mobile cyber-physical systems, augmented reality applications, or holographic computing utilize simulation results. To support these applications, the goal of this project is to enable mobile devices to perform complex simulations by distributing simulation tasks between the resource-constraint mobile device and powerful cloud computing infrastructures.

Consistency in Software-defined Networking 

Funded by: University of Stuttgart

Software-defined networks are based on a logically centralized but physically distributed control plane. In this project, we investigate how to enable consistent network updates and consistent control plane distribution in SDN.

Finished Projects

ARAMiS

Funded by: Federal Ministry of Education and Research (BMBF)

ARAMiS is a joint research project with partners from academia and industry investigating concepts and methods for utilizing multi-core technologies to increase the safety, efficiency, and comfort within the demains automotive, avionic, and rail.

In ARAMiS project, IPVS is focusing on two research topics: On the one hand, we investigate virtualization technologies for the safe and efficient sharing of resources (in particular, GPUs) between different automotive applications of different criticality. On the other hand, we investigated mobile code offloading methods to optimizing the performance and efficiency of mobile applications (in particular, automotive applications) by automatically offloading resource-intensive tasks to a mobile cloud infrastructure.  

[Project Homepage]


Project A2 "Context-aware Communication" [Collaborative Research Center 627]

Funded by: German Research Foundation (DFG)

In this project, we designed communication mechanisms that utilize context information of participants to selectively distribute information. In particular, we investigated geocast protocols using only location information to defined the set of recipients and contextcast protocols taking into consideration further context information for the efficient forwarding of messages in an overlay network of location-aware and context-aware routers.

[Project Homepage]

Project B3 "Hybrid Model Management" [Collaborative Research Center 627]

Funded by: German Research Foundation (DFG)

In this project, we investigated methods for managing context-information on mobile devices. In particular, we designed a first approaches for gathering context information using crowds of mobile devices equipped with sensors, taking into account the limited energy resources of such mobile systems. This research is carried on as part of the Com'N'Sense project described above.

[Project Homepage]

Project B5 "Dynamic, Time-Referenced Model Data" [Collaborative Research Center 627]

Funded by: German Research Foundation (DFG)

In this project, we concentrated on two research questions: First, we looked at the problem of how to track mobile objects efficiently and accurately in real time. To this end, we designed different position update protocols that allow for trading off efficiency – defined by the communication cost of sending mobile object positions to a mobile object database and the size of data stored by the database – and the accuracy of the trajectory.

Secondly, we designed a context broker for searching for heterogeneous context data stored on different servers. To this end, we designed formalisms and index structures for heterogeneous information sources.

[Project Homepage]

Project E3 "Distributed Situation Recognition Based on Evaluated Context Information" [Collaborative Research Center 627]

Funded by: German Research Foundation (DFG)

The overall goal of this joint research project with partners from artificial intelligence was to derive meaningful high-level context information (situations) from low-level sensor data.

In this context, we designed an overlay network for the in-network processing of streams of sensor data transmitted by spatially distributed networked sensors. We designed operator placement algorithms for distributing processing operators on overlay nodes such that the network load is minimized and a certain maximum end-to-end delay (quality of service) is guaranteed.

[Project Homepage]

Project E5 of Graduate School of Excellence Advanced Manufacturing Engineeing (GSaME)

Funded by: German Research Foundation (DFG)

Project E5 investigated methods to improve the consistency in product monitoring in a smart real-time factory scenario based on RFID technology. As a first step, a consistency stack has been developed that conceptualizes the different consistency issues into separate layers. Moreover, algorithms have been designed to provide probabilistic guarantees about observations of monitored objects utilizing a distributed system of RFID readers deployed along production lines. To this end, we designed concepts to increase the reliability of readings based on redundant observations as well as concepts for the automatic calibration of RFID readers.

[Project Homepage]

Supporting Complex Queries in Peer-to-Peer Systems

Funded by: Industry (Alcatel-Lucent)

In this project, we designed concepts for looking up content and users in a peer-to-peer system. In more detail, suitable system architectures and distributed index structures for peer-to-peer system supporting complex context-based queries beyond simple id-based look-ups have been investigated. In particular, we aim at the efficient support of multi-attribute range queries, which are essential for location-based queries for instance.

[Project Homepage]