Disco Network Optimization

Teleoperation is an upcoming application over modern networks such as 5G. Teleoperation data flows require very strict delay guarantees that can only be given if the network is tightly controlled. One method to calculate those guarantees is the DISCO Network Calculator which derives bounds on the flow’s end-to-end delay given a graph of network nodes and the routing of the flows crossing these nodes.

In this project group, we aim to optimize a modern deterministic industrial campus network where teleoperation flows (flows for remote controlling machines) as well as flows for video and audio meet. The main target is to embed these teleoperation flows into the network while keeping very tight delay guarantees needed for smooth teleoperation. A secondary objective is to embed as many of these flows as possible, i.e., scaling. Here it is important that the network optimizer runs near real-time.

As a participant in this project group, you will obtain broad knowledge on the role of network optimization, different routing policies as well as the different factors that affect the delay of teleoperation network flows. In addition, we may have the possibility to run your network optimizer in a real-world 5G network for teleoperation.

Authentifizierung über Verhaltensbiometrie in XR

Behavioral Biometrics, particularly utilized in XR (Virtual Reality, Augmented Reality, Mixed Reality), is currently a hot research topic in human-centered security. In this project group, we are going to explore novel approaches to authenticate users in XR. From a technical standpoint, we are going to create one or several prototypes through 3D programming (e.g., Unity as an engine and C#) and evaluate it empirically in a user study by making use of Machine- or Deep-Learning techniques (e.g., Python, Keras, Pytorch). The goal of this project group is to extend the horizon of user authentication in XR, making passwords in XR a little bit more obsolete, and to publish this knowledge in an adjunct academic publication.

Web-based Organization Mining

An important goal of many universities is to increase the number of startups that transform innovative research results of the university into sustainable businesses. To reach this goal, it is necessary to connect the researchers that have generated promising results with the business advisors and innovation coaches of the university that help scientists to successfully launch a new business.

The goal of this project group is to design and implement a web-based tool that supports the advisors and coaches of a university in identifying research groups or researchers that are working on innovative topics. As primary input the tool shall process the web pages of a university to automatically extract relevant information. Some examples for this are:

• The organization and structure of the university (names of the faculties, research groups and researchers, etc.)
• The research projects of the different research groups (project name and topic, project duration, funding scheme and budget, etc.)
• The publications of the different researchers (authors, title, type of publication, etc.)

The algorithms implemented as part of the tool shall be generic to support the data extraction from the web sites of different universities. To visualize the data, the project group shall develop a web-based application that enables business advisors and innovation coaches to browse and search the extracted information.

From a technical perspective, the project will encompass the development and integration of a web-crawler, a search index, a data mining framework with the associated templates to extract the desired information and a web application to access the data. For the web-crawler and search index, we are currently planning on using Scrapy and Elasticsearch. The technologies used to perform the actual data mining can be freely defined by the students.

From a theoretical perspective, the project group covers fundamental concepts related to web search and data mining in theory and practice. This includes web crawling and search as well as data extraction and information integration. In addition, the participants will prepare individual seminar talks and papers on selected research topics related to web search and data mining.

Coordinated Adaptation in Cloud Computing

Cloud computing offers compelling advantages for deploying and running software. Advantages include the instantaneous access to seemingly infinite computing resources without the need for installing and maintaining costly IT equipment. However, the cloud is a complex and highly dynamic environment. For example, the active user base of cloud services is continuously changing, and so is the intensity with which cloud users use the services. Moreover, the requirements of users are always evolving, requiring new services, the upgrade of old services, or the removal of deprecated ones. To adapt to such kind of changes, cloud services must dynamically scale in or out by replicating (or discarding) software components cloud servers.

With the increasing scale of modern cloud systems, more and more companies use Kubernetes to dynamically adapt their cloud systems at runtime. Kubernetes is an open-source solution introduced by Google. Kubernetes facilitates managing, deploying, scheduling, and scaling applications on large-scale cloud computing clusters.

Kubernetes works in a centralized way, which means that adaptation decisions are taken by a central adaptation component. Such a centralized adaptation component has many shortcomings: it exhibits a single point of failure, it leads to performance issues in the case of large-scale systems, and it assumes total visibility and control over system components.

The overall goal of this PG is to enhance Kubernetes to support decentralized adaptation. In particular, this requires coordinating the decentralized adaptation components. The PG explores two complementary directions to facilitate coordination: (1) a model-based approach using the distributed constraint optimization problem (DCOP) framework, (2) a learning-based approach using multi-agent reinforcement learning.

SmartPostBox

Die „Smart PostBox“ ermöglicht die sichere Paketzustellung in Abwesenheit durch moderne SmartHome- und VoIP-Technik.

• Ausführliche Informationen zu den einzelnen Projektgruppen gibt es unter Moodle.
• Nicht-UDE-Studierende ohne UDE-Unikennung müssen sich für Moodle erst registrieren. Den Einschreibeschlüssel fragen Sie bitte bei den Ansprechpartnerinnen und Ansprechpartnern nach.