http://www.tu-ilmenau.de

Logo TU Ilmenau


INHALTE

Student projects

Filter

Mobile Communications

Topic

SON Self-Coordination Framework using Restricted Boltzmann Machines based Recommender Systems

Kind of work:
Student research project, Advanced Research Project (MSCSP)
Supervisor:
M.Sc. Tanmoy Bag
Status:
available
Description:
Restricted Boltzmann Machine (RBM) based Recommender System (RecSys) is to be studied and implemented to learn its applicability in the domain of Self-Organizing Network functions (SFs). The objective is to model the dynamics between the concurrently executing SFs in order to recommend appropriate network configurations according to the changing state of the environment.

Tasks
- Literature study on the conflict between ICIC and CCO SFs.
- Literature study on Deep Learning based Restricted Boltzmann Machines and Recommender Systems.
- Implement RBM based RecSys with the available infrastructure and compare results with state-of-the-art approaches.

Requirements
- Programming experience in Python & C++
- Basic knowledge of Machine Learning algorithms and frameworks

References
1. T. Bag, S. Garg, D. F. Preciado Rojas, A. Mitschele-Thiel, “Machine Learning based Recommender Systems to achieve Self-Coordination between SON Functions”, in IEEE Transactions on Network and Service Management, Accepted Sept. 01, 2020.
2. H. Ben Yedder, U. Zakia, A. Ahmed and L. Trajković, "Modeling prediction in recommender systems using restricted boltzmann machine," 2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC), Banff, AB, 2017, pp. 2063-2068, doi: 10.1109/SMC.2017.8122923.

Topic

Synthetic Data Generation for enhancing Network Data Analytics in Self-Organizing Networks

Kind of work:
Student research project, Advanced Research Project (MSCSP)
Supervisor:
M.Sc. Tanmoy Bag
Status:
available
Description:
Self-Organizing Networks (SON) driven by Artificial Intelligence demand a large amount of data to learn the state of the environment and determine the most suitable configuration parameters accordingly. In real networks, it is a challenge to collect a significant amount of rich and consistent network data that is sufficient for training complex ML models, representing the dynamics between several concurrently executing SON functions (SFs). The idea is to apply Deep Learning techniques to generate relevant synthetic network data that can complement a limited real existing dataset.

Tasks
- Understand the jointly implemented SFs and the corresponding involved variables in the dataset.
- Explore synthetic data generation approaches and implement a suitable one for the given dataset.
- Measure the quality of the generated synthetic network data and the accuracy enhancements it brings to the jointly modeled SON algorithms.

Requirements
- Programming experience in Python
- Basic knowledge of Machine Learning algorithms and frameworks

References
1. T. Zhang, K. Zhu and D. Niyato, "A Generative Adversarial Learning-Based Approach for Cell Outage Detection in Self-Organizing Cellular Networks," in IEEE Wireless Communications Letters, vol. 9, no. 2, pp. 171-174, Feb. 2020, doi: 10.1109/LWC.2019.2947041.
2. B. Hughes, S. Bothe, H. Farooq and A. Imran, "Generative Adversarial Learning for Machine Learning empowered Self Organizing 5G Networks," 2019 International Conference on Computing, Networking and Communications (ICNC), Honolulu, HI, USA, 2019, pp. 282-286, doi: 10.1109/ICCNC.2019.8685527.

Topic

Research on QoS strategies for MCPTT applications over LTE

Status:
available
Description:
The development of Mission Criticial Push To Talk (MCPTT) application over LTE requires the LTE infrastructure to meet the requirements of public safety mission criticial voice communication. This needs QoS strategies to be targeted specifically for MCPTT applications on both RAN (Radio Access Network) and the core network for higher reliability and lower latency.

Tasks:
- Understanding the LTE infrastructure with added modules of IMS, PCRF and VoLTE required for MCPTT functionality
- Research on state of the art QoS strategies on RAN as well as the Core Network
- Propose a QoS scheme with a focus on meeting the latency and reliability requirements of MCPTT application

Topic

Development of delay tolerant routing protocols for android based devices

Kind of work:
Student research project, Project Work, Bachelor Thesis, Advanced Research Project (MSCSP)
Supervisor:
M.Sc. Mehdi Harounabadi
Status:
available
Description:
Delay tolerant networks are intermittently connected networks where nodes apply the store-carry-forward mechanism to communicate. This mechanism relies on the mobility of wireless nodes and the forwarding of messages. However, messages must tolerate a long latency due to disconnections in a network. Regarding to the message forwarding in delay tolerant networks, different approaches may be applied such as epidemic, social tie based and history based message forwarding. The goal of this project is to develop an application for the Android platform to emulate the delay tolerant routing protocols.
Tasks:
- The store-carry and forwarding mechanism must be implemented for the android based devices.
- Different forwarding strategies must be compared. The student must test the implemented solution under different scenarios and analyze the results.

Requirement:
Programming for Android devices
Data base
Basics of communication networks

Topic

Modeling a Mobility-Controlled Wireless Node in a Network Simulator

Kind of work:
Advanced Research Project (MSCSP)
Supervisor:
M.Sc. Mehdi Harounabadi
Status:
available
Description:
A mobility-controlled node can leverage a new degree of freedom in wireless networks which is the trajectory planning for the node. The objective of this project is to implement a new mobility model for wireless nodes which can be controlled by an application in a network simulator.
Tasks:
- Implementing a new mobility model in a networks simulator (ONE, NS, OMNet++)
- Study on the impact of controlling mobility in wireless networks

Topic

Real time scheduling algorithm for 5G cellular communication on software defined radios

Kind of work:
Student research project, Bachelor Thesis, Master thesis, Advanced Research Project (MSCSP)
Supervisor:
Zubair Shaik
Status:
available
Description:
Applications with requirements of high reliability and low latency need to be addressed specifically when compared to the other applications. The focus of this project is to design a scheduling algorithm for future cellular communication that improves the performance of real time applications, with respect to latency and reliability.

Tasks:
1. Research on real time scheduling algorithms for LTE/5G
2. Implementation of the algorithm on open source solutions of Openairinterface5G or srsLTE.
3. To test the validation on real time hardware using SDRs

Requirements:
1. Knowledge of cellular communication systems
2. very good programming skills on C/C++
3. The ability to solve practical issues while working on the hardware setup (USRPs, Linux systems etc.)

Topic

Research of the available scheduling algorithms on a setup of NS3 and NI's Labview LTE Application framework using SDRs

Kind of work:
Student research project, Advanced Seminar, Bachelor Thesis, Advanced Research Project (MSCSP)
Supervisor:
Zubair Shaik
Status:
available
Description:
The idea is to first bringup the setup of National Instrument's Labview LTE Application framework on the SDRs and port it with the available solution of NS3.
The next step involves the research of available scheduling algorithms on NS3/labview and analyze them with respect to low latency and high reliability.
Advanced projects/thesis would require to design a novel scheduling algorithm for real time applications like VoIP, MCPTT.

Requirements:
1. Knowledge of cellular communication systems
2. very good programming skills on C/C++
3. The ability to solve practical issues while working on the hardware setup of USRPs.

Topic

Mission Critical Push To Talk (MCPTT) on LTE Network using SDRs

Kind of work:
Student research project, Bachelor Thesis, Advanced Research Project (MSCSP)
Supervisor:
Zubair Shaik
Status:
available
Description:
Design and implementation of Mission Criticial Push To Talk (MCPTT) on LTE/5G cellular networks.

Tasks:
1. Understanding the architecture of MCPTT based on the Rel. 13 standard
2. Research on QoS, MBMS or direct communication strategies to enhance MCPTT
3. Realising the MCPTT solution on LTE/5G infrastructure using software defined radios (SDRs)

Topic

Mode selection of D2I and D2D communication on LTE/5G

Kind of work:
Student research project, Bachelor Thesis, Advanced Research Project (MSCSP)
Supervisor:
Zubair Shaik
Status:
available
Description:
Based on the network performance, design an algorithm which selects the mode of an UE (D2D, D2I and relay) connected to the infrastructure network.

Tasks:
1. Research on cellular communication and D2D protocols
2. Design the mode selection algorithm with the following modes of D2I, D2D and relay.
3. Implementation of the algorithm on Omnet++ simulator

Topic

D2D Communication on Software Defined Radios (SDRs) for 5G Cellular Networks

Kind of work:
Student research project, Advanced Research Project (MSCSP)
Supervisor:
Zubair Shaik
Status:
available
Description:
The focus is on the implementation of device-to-device communication on software defined radios(SDRs) by using open source solutions of 5G communication.

Tasks:
1. Research on future mobile communication systems (5G systems)
2. Implementation of D2D on SDRs

Requirements:
1. Knowledge of cellular communication systems
2. very good programming skills on C/C++
3. The ability to solve practical issues while working on the hardware setup (USRPs, Linux systems etc.)

Topic

Wireless mesh network of mobile eNBs for LTE backhaul communication

Kind of work:
Student research project, Bachelor Thesis, Master thesis, Advanced Research Project (MSCSP)
Supervisor:
Zubair Shaik
Status:
available
Description:
As the mobile eNBs serve the cellular UEs, the idea is to form a wireless mesh network of the eNBs which are not connected directly to EPC. For the backhaul communication, the mesh network would be formed using an additional interface of LTE on the eNB. The project is to be implemented using Omnet++/NS3 simulator.

Tasks:
– Understanding LTE infrastructure and its radio access protocol
– Design and implementation of resource allocation for the mesh network (Centralized/distributed)
– LTE mesh network formation eventually connecting to EPC for the backhaul communication
– Use of an existing routing protocol in the simulator for the multi-hop communication on the mesh network
– Evaluation of network performance of the cellular and mesh network for VoLTE application

Topic

Novel Dynamic Path Planning Algorithm for localization of wireless nodes in multi-UAV scenario using implicit signaling in DTN

Kind of work:
Master thesis
Supervisor:
M.Sc. Mehdi Harounabadi
M. Sc. Alina Rubina
Status:
available
Description:
In unknown environments, dynamic trajectory planning for a UAV represents an effective solution for localizing wireless nodes. The UAV must localize wireless nodes accurately and efficiently without wasting resources.
However, employing a single UAV for localization in vast areas causes long latency and may not be feasible. To overcome these problems, multiple UAVs can be applied in the network.
In this case, cooperation between UAVs is required to avoid redundant visits of already localized nodes. Also, covering the same area by different UAVs will waste time and time is the most crucial parameter in disaster scenarios.
As UAVs may not be always connected to exchange information about their planned trajectory and localized nodes, one possible solution will be employing a signaling of control information among UAVs.
There are two types of signaling. In the first case an explicit signaling is used, where a UAV exchanges the control information whenever it encounters another UAV. However, in the vast networks this cannot be applicable as UAVs have short communication range.
The other feasible solution is implicit signaling through wireless nodes. This can also be called delay tolerant signaling. To implement this, a list of localized nodes is stored in the network among localized nodes.
Wireless nodes act as relay nodes to exchange signaling information among UAVs (indirect signaling of UAVs).
By indirect signaling of UAVs through nodes, the redundancy in time, covered area and number of localized nodes can be decreased significantly. It can improve the latency of localization and decrease the traveled distance of UAVs.
Tasks of the student:
1- Literature review about existing work in static and dynamic UAV trajectory planning for localization and delay tolerant networks routing
2- Developing a cooperative multi-UAV dynamic trajectory planning approach for localization of wireless nodes employing indirect signaling among UAVs
a. Modeling the problem
b. Proposing a solution for the modeled problem
3-Implementation and evaluation of the proposed algorithm comparing two scenarios: no signaling and indirect signaling

Topic

Implementation and Evaluation of Different Path Planning Algorithms for UAV in Wireless Networks

Kind of work:
Master thesis
Supervisor:
M. Sc. Alina Rubina
Status:
available
Description:
Quite recently, considerable attention has been paid to the problem of anchor placement, i.e. a design of a trajectory for a mobile anchor (UAV in our case). Previous studies on this issue have shown that a well-planned mobile anchor trajectory increases the speed and accuracy of the localization process [1].
In this paper, one very challenging scenario is considered where an unmanned aerial vehicle (UAV) is flying over an urban area, which suffers from a disaster. The purpose of a UAV is to localize ’survived’ devices, enabled with Wi-Fi module. The obtained information will help to accelerate the rescue process of people who are stuck inside a building.
This work will be mainly divided into two parts:
1) Simulation part
Implementation of chosen three trajectories (namely, Sierpiński curve, Lebesgue curve and Peano curve) in Python.
Simulation of those and comparison with the existing trajectories.
Consideration of two scenarios, with and without a building.
Based on the simulation results, choosing one or two for the experimental evaluation.
2) Experimental part
Evaluation of the chosen trajectory in a real-world experiment considering two scenarios.

Topic

Applying the solution of vehicle routing problem for path planning of multiple message ferries

Kind of work:
Diploma Thesis, Project Work, Master thesis
Supervisor:
M.Sc. Mehdi Harounabadi
Status:
available
Description:
Message ferrying is a solution to exchange delay tolerant messages among disconnected wireless nodes that are located out of the transmission range of each other. A ferry node (e.g. UAV), travels among disconnected nodes and deliver their messages. The mobility of a ferry node can be controlled and its travel path can be given in advance. Vehicle routing problem (VRP) is a combinatorial optimization problem which defines set of optimal routes for a fleet of vehicles.
The goal of this project is to model the message ferry path planning problem as VRP and to propose a solution to minimize the message delivery delay in a message ferry network.

E-Learning / Remote Lab

Topic

E-Assessment auf Basis interaktiver Simulationswerkzeuge

Kind of work:
Master thesis
Supervisor:
Dr.-Ing. Heinz- Dietrich Wuttke
Christian Saul
Status:
available
Description:
Existierende E-Assessment-Systeme, sowohl Alleinstehende (z.B. Questionmark Perception, Respondus) als auch in Lernplattformen (z.B. moodle, ILIAS, Backboard) integrierte Werkzeuge setzen nur einfache Fragetypen (z.B. Multiple-Choice) für die Überprüfung des gelernten Wissens ein. Dies ermöglicht meist nur das Abprüfen von Faktenwissen. Eine selbstständige Bearbeitung komplexer Verständnis-, Anwendungs- und Analyseaufgaben kann damit nicht überprüft werden. Bezogen auf die Bloom’sche Taxonomie bedeutet dies, dass die unteren Stufen durch bestehende Assessment-Systeme und -Werkzeuge abgedeckt werden, wohingegen die oberen 3 (Analyse, Synthese, Evaluation), auch als Kompetenzen höherer Ordnung bezeichnet, ein Problem darstellen.

Ziel dieser Arbeit ist die konzeptuelle Integration von interaktiven und simulativen Werkzeugen in die Prozesse eines E-Assessment-Systems und deren prototypische Implementierung am Beispiel des am FG Integrierte Kommunikationssysteme entwickelten Online-Laboratoriums RemoteLab sowie des am Fraunhofer IDMT entwickelten E-Assessment-Systems askme!.

Topic

semantic net in the domain digital systems

Kind of work:
Student research project
Supervisor:
Dr.-Ing. Heinz- Dietrich Wuttke
Status:
available
Description:
Semantische Netze eignen sich zur Darstellung komplexer Zusammenhänge und eignen sich als Darstellungsmittel zur Erschließung eines Wissensgebietes.
Ziel der Arbeit ist die Aufbereitung der Begriffswelt o.g. Domäne sowie deren webbasierte Darstellung als Nutzerinterface für eLearning-Angebote. Hierzu sind einerseits geeignete Relationen zur Darstellung zu definieren und andererseits geeignete Werkzeuge für eine prototypische Implementierung zu finden und anzuwenden.

Topic

Reuseable learning objects

Kind of work:
Student research project
Supervisor:
Dr.-Ing. Heinz- Dietrich Wuttke
Status:
available
Description:
Multimediale Lehr- und Lerninhalte (oder kurz „Lernobjekte“) sind in den unterschiedlichsten Technologien bezüglich Erzeugung, Speichermedium und Distributionsmöglichkeiten in den letzten Jahren entwickelt worden. Es wurden unterschiedliche Autorenwerkzeuge eingesetzt, die die weitere Verwendbarkeit der so erzeugten Lernobjekte wesentlich beeinflussen. Sie werden entweder auf Datenträgern und Servern der Autoren oder komfortabler in Contentmanagement- (CMS) oder Lernmanagement- Systemen (LMS) verwaltet. Um derartige Materialien zu nutzen gibt es gegenwärtig nur die Möglichkeit, in Kontakt mit den Autoren zu treten und individuelle Kooperationen einzugehen. Sollen die Materialien dann zusätzlich noch in einen neuen, eigenen Kontext gestellt werden, so ist der hierfür notwendige Arbeitsaufwand meist höher als ein Neuentwurf.

Zeil des Hauptseminars ist es, exemplarisch wiederverwendbare Lernobjekte zu realisieren, die auf unterschiedlichen Plattformen laufen können.

Integrated HW/SW Systems

Topic

Dynamic self-organized algorithm for UAVs swarm

Kind of work:
Diploma Thesis, Advanced Seminar, Project Work, Bachelor Thesis, Master thesis
Supervisor:
M. Sc. Victor Casas Melo
Status:
available
Description:
UAVs haben viele Gründe um zusammen zu fliegen. Ein Grund dafür ist ein gemeinsamer Weg erledigen zu müssen. Während des gemeinsamen Fluges, UAV sollen Kohäsion und Separation bewahren, sowie Kollisionen mit externe UAVs zu vermeiden. In der Literatur beschäftigen verschieden Arbeiten sich mit UAVs, die von Anfang bis zum Ende eine Mission zusammen durchziehen. Für diese uns ist der Fall interessant, wo UAVs mit unterschiedlichen Ziele, aber mit teilweisen gemeinsamen Strecken temporäre Schwärme bilden.


Ziel dieser Arbeit ist einen selbstorganisierten Algorithmus zu entwickeln, der ermöglicht:

- dass UAVs erkennen, wann nähe UAVs eine gemeinsame Strecke haben.

- dass UAVs mit möglich wenig Aufwand und ohne Kollisionen zusammen, wenn eine gemeinsame Strecke erkennbar ist.

- dass UAVs mit möglich wenig Aufwand und ohne Kollisionen sich trennen, wenn der Strecke einer der UAVs nicht mehr passt.


Der Algorithmus soll in Simulation validiert werden, kann aber auch mit realen UAVs validiert werden.



Topic

Survey on benchmark problems for swarm algorithms for UAVs

Kind of work:
Student research project, Advanced Seminar, Project Work
Supervisor:
M. Sc. Victor Casas Melo
Status:
available
Description:
There are different works in the state of the art about swarm algorithms for UAVs. They ensure that UAVs keep cohesion, separation, and alignment while flying together. For the validations of such algorithms, there are as many scenarios, assumptions, and targets., as works out there. Therefore, the comparison of different algorithms is difficult.

The goal of this work is to summarize relevant scenarios, assumptions, and targets for the validation of swarming algorithms for UAVs. With his experience from the literature, a set of benchmark-problems should be proposed. These benchmarks problems seek the standardization of the validation of swarming algorithms.

Topic

Entwicklung einer selbstorganisierten Landungsstrategie für Quadkopternschwärmen

Kind of work:
Advanced Seminar, Bachelor Thesis, Master thesis
Supervisor:
M. Sc. Victor Casas Melo
Status:
available
Description:
Quadkoptern die zusammen fliegen, haben das gleiche Ziel. Die können aber nicht alle auf denselben Punkt landen. Deswegen brauchen die Quadkoptern eine Strategie um kollisionsfrei zu landen, wenn sie am Ziel sind.


Aufgaben:


- Literaturrecherche zum Thema: Landungsstrategien für Quadkoptern


- Entwicklung und Implementierung einer Landungsstrategie für einen Quadkopternschwarm


- Optional: Realer Test der Strategie mit drei Quadkoptern(Wetterabhängig)

Topic

Simulation of quadcoptern swarm in gazebo

Kind of work:
Advanced Seminar
Supervisor:
M. Sc. Victor Casas Melo
Status:
available
Description:
Simulation von Quadkoptern-Schwarmen ist ein sehr interessantes Thema. Je realitätsnäher die Simulation ist, desto erfolgreicher und sicherer die experimentellen Erprobungen sind. Wir haben ein Schwarm Algorithmus mit Morse simuliert und mit realen Koptern getestet. Die Aufgabe ist das Schwarm Algorithmus mit dem Simulator für Robotern Gazebo zu simulieren und festzustellen, welcher Vorteile Gazebo gegenüber Morse hat.

Aufgaben
- verschiedene Quadkoptern Modelle für Gazebo in der Literatur zu suchen
- ein Modell auswählen
- Schwarm Algorithmus in Gazebo simulieren
- Ergebnisse mit Morse und reale Experimente vergleichen
- In Python programmieren


J. García and J. M. Molina, "Simulation in Real Conditions of Navigation and Obstacle Avoidance with PX4/Gazebo Platform," 2019 IEEE International Conference on Pervasive Computing and Communications Workshops (PerCom Workshops), Kyoto, Japan, 2019, pp. 979-984.

PX4 Development guide, Gazebo simulation: https://dev.px4.io/v1.9.0/en/simulation/gazebo.html

F. M. Noori, D. Portugal, R. P. Rocha and M. S. Couceiro, "On 3D simulators for multi-robot systems in ROS: MORSE or Gazebo?," 2017 IEEE International Symposium on Safety, Security and Rescue Robotics (SSRR), Shanghai, 2017, pp. 19-24.

Topic

Verbesserung von Simulationsmodellen für Quadkopternschwarm

Kind of work:
Advanced Seminar, Bachelor Thesis, Master thesis
Supervisor:
M. Sc. Victor Casas Melo
Status:
available
Description:
Realitätsnahe Simulationsmodelle sind ein wertvolles Werkzeug um Algorithmen zu entwickeln, die später mit realen Quadkoptern getestet werden. Wir haben einen Simulator, mit dem man einen Quadkopternswarm simulieren kann. Wir verfügen auch über eine Flöte von Quadkoptern, mit der man Schwarmalgorithmen testen kann. Ergebnisse von realen Quadkoptern unterscheiden sich von Simulationsergebnisse. Das liegt an Parametern, die Windwiderstand und PID-Reglern des Koptern beeinflussen. Solche Parametern lassen sich nicht eins zu eins von realen Quadkoptern in den Simulator übertragen.

Ziel der Arbeit ist Werte für die oben genannte Parametern zu finden, mit den der Simulator realitätsnah Ergebnisse liefert. Dafür könnte Mann eine Brute-Force-Methode (erschöpfende Suche) anwenden, indem alle mögliche Kombinationen von Parametern testet, bis die Simulation die Realität ähnelt. Da diese Methode Zeit intensive ist, wird erwünscht, Methoden wie Evolutionäre Algorithmen oder Gradientenverfahren bei der Suche auszuprobieren.


Aufgaben:


- Unserer Kopternsimulator kennenzulernen (Python)


- Realen- und Simulationsergebnisse von einer Schwarstrategie zu vergleichen


- Implementierung eines Suchverfahren, um die Simulatorergebnisse realitätsnäher zu machen

Topic

Research about modern test possibilites on the basis of BSCAN

Kind of work:
Bachelor Thesis, Master thesis
Supervisor:
Dr.-Ing. Heinz- Dietrich Wuttke
Status:
available
Description:
This work is a cooperation between TU Ilmenau and a company in Jena. The exact task assignment will be discussed with the advisors and the company if the student is interested.
All topics will be around the following keywords:
- JTAG
- BSCAN
- structural testing of PCBs

Requirements:
- knowledge of VHDL and/or C/C++ is usefull

Further details can be discussed with the advisor.

Mobile Communications

Topic

Recommender System with Deep Learning based AutoEncoder for coordinating SON Functions online

Kind of work:
Advanced Research Project (MSCSP)
Supervisor:
M.Sc. Sharva Garg
Status:
in progress
Description:
Edge network data analytics can support the complex modelling of the complementing and conflicting interactions between the diverse SON functions in a data-driven manner. In this project, Autoencoder based Recommender System is to be implemented as a cloud solution that can recommend SON configurable parameters online, according to the changing state of the mobile network.

Tasks
- Literature study on Autoencoder based Recommender Systems
- Understand 5G service based architecture with a focus on Network Data Analytics Function (NWDAF).
- Implement Autoencoder based RecSys as a service for SONs.

References
1. T. Bag, S. Garg, D. F. Preciado Rojas, A. Mitschele-Thiel, “Machine Learning based Recommender Systems to achieve Self-Coordination between SON Functions”, in IEEE Transactions on Network and Service Management, Accepted Sept. 01, 2020.
2. Yao Wu, Christopher DuBois, Alice X. Zheng, and Martin Ester. 2016. Collaborative Denoising Auto-Encoders for Top-N Recommender Systems. In Proceedings of the Ninth ACM International Conference on Web Search and Data Mining (WSDM ’16). Association for Computing Machinery, New York, NY, USA, 153–162.
3. Florian Strub, Jérémie Mary. Collaborative Filtering with Stacked Denoising AutoEncoders and Sparse Inputs. NIPS Workshop on Machine Learning for eCommerce, Dec 2015, Montreal, Canada.

Topic

Offline-online decison making for multi-ferry networks

Kind of work:
Diploma Thesis, Master thesis
Supervisor:
M.Sc. Mehdi Harounabadi
Status:
in progress
Description:
Message ferry networks are disconnected wireless networks where message ferry nodes are employed to travel among disconnected nodes and deliver their messages. The travel path of ferries can be planned offline or ferries can decide their mobility online (on-the-fly). The goal of this project is to propose a message ferry network with multiple message ferries where the path for some ferries are planned offline and other ferries make decision on-the-fly. Each ferry may visit all nodes or a subset of nodes. The assignment of ferries to nodes and type of decision making for a ferry(offline or online) must be done in such a way which the average delay for message delivery in a network is minimized.
The optimization problem should be solved (approximated applying heuristic methods) and the proposed solution must be compared to the networks with pure offline path planning for ferries and networks with pure on-the-fly decision making.
Requirements:
- Understanding of heuristic methods
- Python

Topic

Implementation of localization software for UAV in real-time

Supervisor:
M. Sc. Alina Rubina
Status:
in progress
Description:
In this project we are considering mobile anchor based localization. We are using an UAV as a mobile anchor and mobile phones represent wireless sensors. UAV explors the area and measures received signal strength of mobile phones which need to be localized. Localization process includes mapping those received signal strengths to distances. Also, knowing the distances we can estimate the position of the node.
The main goal of this project is to develop a localization software for UAV. This software needs to process data in real-time.

Topic

Performance evaluation of different multi-ferry path planning approaches

Kind of work:
Master thesis
Supervisor:
M.Sc. Mehdi Harounabadi
Status:
in progress
Description:
Message ferrying is a solution to exchange delay tolerant messages among distant wireless nodes that are located out of the transmission range of each other. The ferry node (UAV), flies to visit ground nodes and in each contact exchange information with them. Ferry is a wireless node with controllable mobility. It picks up messages from a ground node, flies to the destination and offloads messages. Decision about which ground node to visit (ferry path planning) impacts the performance of message forwarding in network. Having a single ferry in networks that has been established in large areas cannot support communication needs of different traffic types in the network. Multi-ferry approach employs several ferries in the network to meet the performance requirements of the network. Several approaches has been proposed in literature for multi ferry path planning. The objective of this project is:
1. Literature study on existing multi ferry path planning approaches
2. Implementation of some of the selected multi ferry approaches on the existing simulator (python based simulator)
3. Investigation on results to make conclusion

E-Learning / Remote Lab

Topic

Implementierung automatisierter Wissenstest zur Unterstützung der Digitalisierung im Hochschulbereich

Kind of work:
Project Work
Supervisor:
Dr.-Ing. Heinz- Dietrich Wuttke
Dr.-Ing. Prof. h. c. Karsten Henke
Students:
Steffen Krug
Status:
in progress
Description:
Im Zuge der Digitalisierung der Lehre werden verstärkt individuell zugeschnittene Formen der Inhaltsvermittlung angestrebt, die durch digitale Medien Unterstützt werden. Dabei kommt dem Testen des erreichten Kenntnistandes eine wesentliche Rolle zu. An der TU Ilmenau existieren gegenwärtig unterschiedliche Systeme, die hierfür Verwendung finden.
Aufgabe der Projektarbeit ist es, zu recherchieren, ob und wie zwischen diesen Systemen vorhandene Inhalte ausgetauscht werden können. Dabei ist einerseits auf die zur Verfügung gestellten Fragetypen einschließlich Bewertungsmethoden einzugehen und andererseits zu vergleichen, welche Daten- und Medienformate unterstützt werden. Im Ergebnis der Recherche soll an ausgewählten Beispielen die Transformation von Inhalten vorgenommen werden. Dazu ist ein entsprechendes Programm zu implementieren, welches die automatisierte Übertragung von vorhandenen Inhalten zwischen den Systemen ermöglicht.

Integrated HW/SW Systems

Topic

Set up and extension of programming lab equipment

Kind of work:
Project Work
Supervisor:
Dr.-Ing. Prof. h. c. Karsten Henke
Status:
in progress
Description:
Only available in English

Mobile Communications

Topic

Survey: Aerial Base Station Placement in Future Cellular Networks

Kind of work:
Advanced Seminar
Supervisor:
M.Sc. Oleksandr Andryeyev
Status:
finished
Abstract: