printlogo
http://www.ethz.ch/index_EN
Welcome
 
print
  

Current

STMT Frontend

Low-power flexible electronic devices based on gate voltage amplification with ferroelectric materials [D-ITET, 1]

Electronic devices on flexible substrates are interesting for a variety of different applications eg. flexible displays, sensors and wearable technology. Especially for mobile devices there is a requirement for low-power consumption of their electronics. One approach to lower the operation voltage for field-effect transistor based switches and memory devices is the introduction of a ferroelectric layer into the gate stack. Previous studies have shown that the semiconductor/dielectric/ferroelectric system can be engineered in such a way that the ferroelectric material reveals a negative capacitance which lowers the overall capacitance value. This leads to a surface potential amplification at the semiconductor interface and hence the device can operate at a lower gate voltage. In this thesis you will fabricate and characterize transistor and capacitor structures in the state-of-the-art cleanroom facilities (FIRST) of ETH. You will vary the semiconductor doping concentration as well as dielectric and ferroelectric thicknesses and the goal will be to obtain knowledge how these variations improve the device performance and result in a lower operation voltage. You will have the opportunity to gain experience in micro-fabrication technology and electrical characterization methods.

Distribution: 20% literature, 40% microfabrication, 40% measurements and evaluation
Requirements: Interest in microfabrication technology and thin-film transistors
Main supervisor: Alwin Daus, H 97, E-Mail: dausa@ife.ee.ethz.ch, Telephone: +41 44 632 68 70
Second supervisor: Giovanni Salvatore, H90, E-Mail: giovanni.salvatore@ife.ee.ethz.ch, Telephone: +41 44 632 23 77
Project Title:
Professor: Prof. Tröster

Ultra-thin dielectrics in flexible thin-film transistors [D-ITET, 1]

One of today's most popular emerging technologies is flexible electronics. It offers novel functionalities for e.g. flexible displays or wearables. One of the key factors in that field is the need of low-power electronic devices which can run on low battery usage to ensure its application in mobile devices. In this work you will improve the basic element of flexible electronics: flexible thin-film transistors. The approach for lower operating voltage will be a reduction of the gate dielectric thickness. The goal of this work is to fabricate the gate insulator with techniques which ensure a good control in the sub-10-nanometer regime and a good film uniformity. You will fabricate these transistors in state-of-the-art cleanroom facilities (FIRST) of ETH and scale down the gate dielectric to thicknesses which reach the quantum-mechanic tunnel regime. You will characterize the limits of this down-scaling process and test the devices under the application of high-electric fields to monitor the tunnel behavior and possible device modifications encountered by that. You will have the opportunity to gain experience in semiconductor process technology as well as advanced electrical characterization methods and take a look into the device physics for the further understanding of the device modification by quantum-mechanic tunnel processes. Finally, the different thicknesses will be compared and conclusions about the required operating voltages as well as energy consumption will be made.

Distribution: 10% Literature, 40% Cleanroom fabrication, 50% measurements
Requirements: Knowledge of semiconductor devices such as transistors, Interest in cleanroom fabrication and device physics
Main supervisor: Alwin Daus, H 97, E-Mail: dausa@ife.ee.ethz.ch, Telephone: +41 44 632 68 70
Second supervisor: Stefan Knobelspies, H97, E-Mail: kstefan@ife.ee.ethz.ch, Telephone: 22741
Project Title:
Professor: Prof. Tröster

Advanced channel materials for flexible thin-film transistors [D-ITET, ST/MT, 1]

One research goal in the area of bendable thin-film transistors (TFTs) is the enhancement of the switching speed. For the improvement of the transistor characteristics, the mobility of the channel region is of prior interest. This project deals with the exploration of different materials or material stacks for the semiconductor layer in TFTs. After a preliminary literature review about possible channel materials and material combinations, you will fabricate the TFTs in a state-of-the-art cleanroom facility. Then, you will evaluate the transistor characteristics, i.e. mobility and switching speed, but also the performance under bending will be investigated.

Distribution: Literature 15%, Clean room fabrication 55%, experimental work (measurements) 15%, Report 15%
Requirements: Interest in clean room fabrication, transistors and thin-film electronics
Main supervisor: Stefan Knobelspies, H97, E-Mail: kstefan@ife.ee.ethz.ch, Telephone: 22741
Second supervisor: Alwin Daus, H 97, E-Mail: dausa@ife.ee.ethz.ch, Telephone: +41 44 632 68 70
Project Title:
Professor: Prof. Tröster

Deep Learning for Sperm Classification using a Smartphone [D-ITET/D-INFK, MT, 1]

Schwangerschaftstests werden heute routinemässig zu Hause durchgeführt. Nicht so beim Mann: Wir entwickeln deshalb ein Smartphone-basiertes Handheld-Mikroskop, das sogenannte «Spermphone», das es dem Mann erlaubt, die Konzentration und Geschwindigkeit (Motilität) seiner Spermien bequem zu Hause zu messen.

Moderne Smartphones mit Megapixel-Bildsensoren und leistungsfähigen Mikroprozessoren erreichen die nötige optische Auflösung, um einzelne Spermien (Mikrometer: 10-6m) zu erkennen. Weiterhin wird die Echtzeitanalyse der erfassen Bildsequenzen (Videos) der Spermien ermöglicht.

Ziele der Arbeit:
Evaluierung, Entwicklung und Benchmarking von Machine- & Deep-Learning Algorithmen für die Klassifizierung von Spermien, welche mit Hilfe eines modifizierten Smartphones aufgenommen werden. Insbesondere soll der Algorithmus folgende Features aufweisen:
- Prozessieren von verschiedenen Bild- und Objektgrössen
- Selbstlernend (z.B. on-line learning)
- Robustheit gegenüber:
- Artefakten wie Verunreinigungen und Luftblasen in Sample
- Aggregaten von sich überlappenden Spermien
- Verschiedene Morphologien bei Spermien (z.B. zwei Köpfe)
Der entwickelte Algorithmus soll in sowohl in einer Cloud wie auch direkt auf den Smartphone ausgeführt werden können - wobei Letzteres vermutlich eine Vereinfachung des Algorithmus voraussetzt.

Problemstellung:
Basierend auf realen Bildern von Spermien sollen zuerst Klassifikations-Algorithmen, welche vor allem auf Machine- und Deep-Learning-Ansätzen beruhen, evaluiert werden. Die vielversprechendsten Algorithmen (1-2) sollen dann auf einem Computer (Cluster) implementiert, getestet und validiert werden.

Im Weiteren soll der Algorithmus auf ein Smartphone portiert und getestet werden (Android und/oder iPhone). Die reduzierten Ressourcen auf dem Smartphone erfordern unter Umständen eine Vereinfachung des Algorithmus.

Technologien und fachliche Schwerpunkte:
- Bildverarbeitung, Machine- / Deep-Learning
- Caffee o.ä. (z.B. Torch), C++, OpenCV, MATLAB
- Mobile Plattformen

Distribution: 50% Algorithmik mit Fokus Machine- und Deep Learning, 50% Umsetzung & Benchmarking
Requirements: Grundkenntnisse in Machine- & Deep-Learning
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Life Science Data Mining with Knime [D-ITET/D-INFK, ST, 1..2]

At SCS, we use novel machine vision & mathematical morphology algorithms to analyze complex multi-modal Life Science images. The extraction of relevant information from such data sets is typically very challenging and time-consuming.

The scope of this thesis is to evaluation and development of different machine learning concepts to data mine real life science data sets.

Distribution: 30% Theory, 70% Software-Engineering
Requirements: Solid knowledge of at least two algorithms of the following categories: Artificial Neural Networks (for classification), Clustering, Decision Trees, Support Vector Machines, PCA; Knowledge of Knime is not required
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

F# for Image Processing [D-ITET/D-INFK, 1]

F# is an open source functional programming language for the .NET platform. We think the functional paradigm fits the problem domain of image processing well. We therefore aim to leverage F# as a platform which allows us to quickly prototype image processing algorithms and create production-ready implementations. The goal of this work is to design a workflow which deals with all challenges involved in an image processing project on the basis of an existing project.

Distribution: 40% Theorie, 60% Software engineering
Requirements: Image Processing von Vorteil
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Towards User-independent 2D/3D Object Classification of Complex Life Science Images [D-ITET/D-INFK, MT, 1..2]

SCS develops novel machine vision & mathematical morphology algorithms to analyze complex multi-modal Life Science images.

The scope of this thesis is the development of 2D/3D image segmentation & classification algorithms for (semi-)automated analysis of complex cellular structures in plants & liver images from Electron Micrographs (EM).

Project plan:
1) Develop & benchmark different 2D/3D algorithms to segment, classify and quantify specific organells in sectional EMs:
- Liver cells: vesicles, mitochondria, red blood cells, ER
- Plant cells: thylakoid membrane, chloroplasts, yeast sections
2) Develop a graphical user interface that bests supports:
- The training phase (object classification)
- Post-corrections of non- & erroneously classified objects

Distribution: 30% Theory, 70% Software Engineering
Requirements: Solid knowledge of Matlab and C++, Good knowledge of Image Processing
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

An event-driven architecture with CQRS and Event Sourcing [D-ITET/D-INFK, MT, 1]

Im Rahmen einer Masterarbeit sollen die Vor- und Nachteile in der Flexibilita?t und Skalierbarkeit einer CQRS und EventSourcing basierten Architektur gegenu?ber einer klassischen, datengetriebenen 3-tier Architektur aufgezeigt werden. Es sollen Guidelines erstellt werden, welche bei der Wahl zwischen einer datengetriebenen Architektur und einer Event-basierten Architektur helfen.

Distribution: 60% Softwareentwicklung, 20% Research, 20% Benchmarking
Requirements: Softwarearchitektur
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Sicherheitskonzept für den Bahnbetrieb [D-ITET/D-INFK, MT, 2]

Sicherheit im Bahnbetrieb spielt eine zentrale Rolle. In der Vergangenheit wurden die Sicherheitskonzepte kontinuierlich erweitert und damit sind auch die Komplexität (und Kosten) enorm gestiegen. In dieser Arbeit soll ein minimales Set von Sicherheitsregeln definiert werden. Diese sollen in einer Simulation auf das Schweizerische Bahnnetz angewandt werden. Es soll simuliert werden, welche Systemleistung mit dem neuen Regelset möglich ist. Diese Systemleistung soll mit dem heutigen System verglichen werden.

Distribution: 40% Daten- und Domänenverständnis, Kreativität, 20% Algorithmik, 40% Software-Entwicklung
Requirements: Kreativität, Programmiersprache
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Latenzbetrachtungen und Optimierungen auf aktuellen Rechnersystemen [D-ITET, MT, 1]

Aktuelle Prozessoren (beispielsweise Intel Xeon) wie sie in PC's und Servern zum Einsatz kommen besitzen mehrere Cores, arbeiten mit sehr hohen Taktraten und haben schnellen Speicher in Form von Caches. Sie erreichen damit immer höhere Rechenleistungen.

Im Gegensatz zur Rechenleistung bleibt die I/O-Latenz beispielsweise über die Netzwerkkarte in den letzten Jahren unverändert oder wurden durch die Umstellung von parallelen auf serielle Bussysteme noch schlechter. Weiter erhöht das Betriebssystem mit den Treibern und Unterbrechungen der Cores die Latenz zusätzlich.

Somit sind aktuelle Prozessoren und Betriebssystemen nur bedingt für latenz-kritische oder Echtzeitaufgaben geeignet.

Ziel dieser Masterarbeit ist es, die Netzwerklatenz eines aktuellen Intel-CPU basieren System mit Linux Betriebssystem zu messen und diese Latenz durch geeignete Massnahmen zu reduzieren.

Mögliche Massnahmen sind:
* Umgehung des Netzwerkstacks
* Reservation eines Cores für latenz-kritische Aufgaben
* Ausschalten der Interrupts für einzelne Cores
* ...

Distribution: 30% Theorie, 70% Umsetzung
Requirements: C/C++, Linux, Treiber
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Beat and Mood Analysis and Visualization of Music [D-ITET/D-INFK, MT, 1..2]

The visualization of music has a long history, but only since recently we do have the processing power to perform advanced music analysis techniques in real time. The goal of this thesis is to visualize live music by means of various stage light equipment.

To achieve a new level of synergy between music and light, the following analyses should be performed on live audio streams and translated to lighting control signals:
- Robust beat detection, possibly incorporating a Kalman filter
- Speech/singing recognition in order to visualize vowels with specific colours or effects
- Mood recognition: Recognize harmony and classify dark, happy, energetic, (...) music.
- More ideas up to the students

A possible technical framework would be the jack audio toolkit as an audio server and python for rapid prototyping of signal processing.

Distribution: 30% Theorie, 40% Umsetzung, 30% Lab Setup
Requirements: Signalverarbeitung, Spracherkennung, Python
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Modular Data Acquisition System with Nanosecond-Accuracy Time Synchronization [D-ITET, MT, 1]

SCS has developed a modular data acquisition system that is especially targeted towards measuring voltages and currents in the power distribution grid. One of the main requirements is a high time synchronization accuracy in a system of many distributed measurement devices. Until now, GPS has been used for time synchronization purposes. But as the installation of a GPS antenna is not always feasible, we'd like to extend the capabilities of our system to synchronize time based on NTP (lower accuracy) and PTP (higher accuracy) depending on the situation.

The work for this thesis includes the design of a suitable HW platform (might include circuit design) that allows for GPS, NTP and PTP synchronization based on an embedded system including a Zynq SoC. Then the synchronization concept should be sketched and implemented using standard Linux drivers with possible custom patches. Some VHDL needs to be written in order to integrate the SCS DAQ Link. Acheived synchronization accuracy will have to be measured in our lab. Finally, a software service needs to be programmed that serves the acquired and timestamped measurement data using some standard protocol to allow plug'n'play with a standard PC.

Distribution: 20% Theorie, 50% Umsetzung, 30% Test und Messung
Requirements: HW Design, Linux Kernel Knowhow, C/C++
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Optimierung eines komplexen Heizungssystems eines Mehrfamilienhauses [D-ITET, 1..2]

Moderne Gebäudeheizungssysteme bestehen zum Teil aus mehreren Komponenten wie Wärmepumpen, Speichern, Dachkollektoren, PV-Panels, Eisspeicher, etc.. Der optimale Betrieb solcher Systeme ist ein komplexes Optimierungsproblem. Im Rahmen einer Studienarbeit soll der theoretisch optimale Betrieb eines solchen Systems unter Einbezug von Umgebungs- und Systemparametern berechnet werden. Basierend auf gesammelten Messdaten soll der effektive Betrieb des Systems an der optimalen Lösung für ein Jahr gemessen werden.

Teil der Arbeit kann der Entwurf eines möglichst optimalen Regelsystems für solche komplexen Anlagen sein.

Distribution: 60% Theorie, 40% Umsetzung
Requirements: MATLAB, Regelungstechnik, mathematische Optimierung
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Know-How Relation Graph [D-ITET/D-INFK, 1..2]

Neue Graphen-Datenbanken ermöglichen das Abbilden komplexer Beziehungen und machen diese durchsuchbar. Es soll ein Modell ausgearbeitet werden, welches Projekte, Technologien und Mitarbeiter in Beziehung zueinander setzt und das Beziehungsnetz analysierbar und durchsuchbar macht. Ein einfaches und intuitives Webinterface ermöglicht die Navigation im Netz.

In der SCS tauchen häufig folgende Fragestellungen auf:
- "Wer hat an diesem Projekt mitgearbeitet?"
- "Welche Technologien wurden in diesem Projekt verwendet?"
- "Welche Technologien wurden in der SCS bereits verwendet?"
- "Wer kennt sich mit einer bestimmten Technologie aus?"
- "In welchem Projekt wurde eine bestimmte Technologie bereits eingesetzt?"
- ...

Ein Knowhow Relation Graph soll es ermöglichen, diese Fragen zu beantworten. Ein einfaches Webinterface ermöglicht die Beantwortung der obigen Fragen und erlaubt es dem User, sich durchs Beziehungsnetz durchzuhangeln, ähnlich wie bei IMDB.com. Weiter soll das Webinterface ermöglichen neue Beziehungen zu erfassen.

Allenfalls können auch temporale und finanzielle Aspekte miteinbezogen werden ("Wann wurde das Projekt durchgeführt?", "Wie gross war der Projektaufwand?").

Verschiedene Versionen der selben Technologie (z.B. Frameworks) sowie die Kategorisierung von Technologien stellen eine weitere Herausforderung an die Modellierung.

Distribution: 10% RE, 30% Modellierung, 60% SW-Entwicklung
Requirements: Software-Entwicklung
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Low-cost hand recognition [D-ITET/D-INFK, 1..2]

Hand identification has been extensively studied in the past and several approaches showed promising results.

The goal of this thesis is to develop a hand recognition algorithm that can deal with images from a low-cost camera and is suitable to run on low-power microprocessors.

Potential tasks:
- Literature research on state-of-the-art concepts/algorithms
- Evaluation of the most promising approaches for the application at hand
- Design and implementation of a prototype algorithm
- Acquiring a test set and bench-marking the algorithm
- Optimization for low-power microprocessors
- Integration into prototype hardware

Distribution: 30% Research, 40% Algorithm prototyping, 30% Optimization
Requirements: Basic image processing know-how
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Fehlerlokalisierung im Niederspannungsnetz [D-ITET/D-INFK, MT, 1]

Die wachsende Transparenz im Niederspannungsnetz ermöglicht es unter anderem transiente Vorgänge hochauflösend aufzuzeichnen. Somit wird es denkbar, mit geeigneten Algorithmen die Fehlerstelle im Störungsfall zu lokalisieren, wie es heute nur auf höheren Spannungsebenen gemacht wird. Die Kosten für die Suche nach der Fehlerstelle und die Interventionszeiten der Netzbetreiber könnten somit gesenkt werden und damit wird ein Mehrwert geschaffen.

Im ersten Schritt soll der Student im Rahmen seiner Arbeit die bestehenden Fehlerortungsalgorithmen auf deren Anwendbarkeit im Niederspannungsnetz untersuchen und eine Evaluation vornehmen. Im zweiten Schritt soll ein geeigneter Algorithmus implementiert und mittels simulierter Daten getestet werden. Zudem sollen Versuche im ewz Labor durchgeführt werden, um echte Messdaten zu gewinnen.

Distribution: 25% Theorie, 25% Evaluation, 50% Umsetzung und Simulation
Requirements: Kenntnisse über elektrische Verteilnetze, Signalverarbeitung
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Brain Computer Interface BCI - Stand-alone and Low Cost [D-ITET/D-INFK, 1..2]

Clinical ElectroEnzephaloGraphie (EEG) systems have long been expensive and bulky. With the increasing interest in wearable Human Machine Interfaces (HMI), the purely mind based interaction with machines have become reality in the last years (e.g. Cybathlon). But the hardware for the Brain Computer Interfaces (BCI) is still bulky and costly.

In the proposed thesis, the possibilities and performance of a low cost, stand-alone BCI device should be investigated: as input device, the 3D-printable EEG headsets Ultracortex from OpenBCI is suggested. As a signal processing unit, the PULPino platform from the ETH IIS could be considered along with other possible choices of open source processing platforms.

Besides putting together a BCI system, signal processing algorithms should be suggested, which take advantage of the low power platform chosen. Ideally the BCI system is used to steer a computer or other remotely controllable device.

Distribution: 30% Theorie, 70% Umsetzung
Requirements: Hardwarenahe Signalverarbeitung
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Bestimmung optimaler Sicherheitsparameter mit Machine Learning [D-ITET/D-INFK, MT, 1..2]

Die ZKE der SBB bestimmt anhand verschiedener Parameter, ob ein Alarm ausgelöst wird. Nicht jeder Alarm führt jedoch zum Stop eines Zuges. Im Idealfall wird ein Alarm nur ausgelöst, wenn ein Zug tatsächlich gestoppt werden soll. Aufgrund der Daten der Vergangenheit sollen daher mithilfe von Machine Learning und Big Data-Verfahren neue Mechanismen entwickelt werden, um die Parameter zu optimieren. Die Anzahl der Fehlalarme soll dabei reduziert werden ohne tatsächliche Risiken zu übersehen. Die Daten der Vergangenheit, welche als Trainings-Set dienen, sind vorhanden.

Distribution: 50% Algorithmik, 50% Umsetzung
Requirements: Grundkenntnisse in Machine-Learning
Main supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Second supervisor: Fabian Schenkel, -, E-Mail: fabian.schenkel@scs.ch, Telephone: +41 (43) 456 16 80
Project Title:
Professor: Prof. Gunzinger

Tattoo-like sensors for gesture recognition using liquid metal [D-ITET/D-INFK, ST/MT, 1..2]

Sensors on thin stretchable membranes (<50 um) that are worn on skin are called epidermal sensors. They are much like temporary tattoos, unobtrusive and removable but with embedded electronics (passive components and antennas). In this thesis, we will explore how such sensors can be used for gesture recognition. The fabrication method of a tilt sensor based on spraying of elastomers and liquid metals will be given. The goal is to explore the use of such sensor for recognition of body postures and gestures. The tasks include: fabrication of the sensor, building a system for the orientation readout, applying machine learning techniques to classify gestures.

Distribution: 30% fabrication, 30% readout system design, 30% applied machine learning, 10% documentation
Requirements:
Main supervisor: Matija Varga, H 96, E-Mail: matija.varga@ife.ee.ethz.ch, Telephone: +41 44 632 79 61
Second supervisor: Alberto Calatroni, H63, E-Mail: alberto.calatroni@ife.ee.ethz.ch, Telephone: +41 44 632 33 91
Project Title:
Professor: Prof. Tröster

2D materials for flexible and high speed electronics [D-ITET, ST/MT, 1]

Thin film transistors (TFTs) are important electronic devices because of their low price and the possibility of large area fabrication. Two-dimensional materials have interesting physical properties, and are attractive for use in next-generation
TFTs because, compared to one-dimensional materials, it is relatively easy to fabricate complex structures from them. The most widely studied two-dimensional material is graphene both because of its rich physics and its high electron mobility; however, it does not have a bandgap, a property that is essential for many applications, including transistors. Although, single layers of MoS2 have a large intrinsic bandgap of 1.8 eV and high mobility; moreover its integration in a transistor like structure has been already demonstrated showing a room-temperature mobility of at least 200 cm2/Vs, a current on/off ratio of 1x108 and ultralow standby power dissipation.
The main goal of the project is to fabricate MsO2 Thin Film Transistors (TFTs) on flexible plastic substrate. This could path the way to a new generation of TFTs for plastic electronics. The MoS2 will be deposited through exfoliation, this means thin layers of the material are transferred to the plastic substrate. The main part of this work will be performes in the FIRST cleanroom. This gives you the opportunity to get familiar with the most important semiconductor processing technologies. The finished TFTs will be electricaly characterized using state of the art characterization equipment.

Distribution: 10% literature; 70% fabrication; 20% characterization
Requirements: Interest in micro fabrication technologies
Main supervisor: Giuseppe Cantarella, H 90, E-Mail: giuseppe.cantarella@ife.ee.ethz.ch, Telephone: +41 44 632 06 94
Second supervisor: Alwin Daus, H 97, E-Mail: dausa@ife.ee.ethz.ch, Telephone: +41 44 632 68 70
Project Title:
Professor: Prof. Tröster

Flexible electronics and OLEDs for biomedical & optoelectronic applications [D-ITET, ST/MT, 1]

The emerging field of flexible electronics places new physical requirements on electronic components. For a broad variety of applications, such as smart textile or biomedicine, the tendency is to have complex and large area devices, capable to be flexible and biocompatible.
The main goal of the project is to fabricate Thin Film Transistors (TFTs)-based electronics (analog and digital circuits) together with Organic Light Emitting Diodes (OLEDs) on flexible plastic substrate. The idea is to create an entirely flexible system for future biomedical & optoelectronic applications.
The main part of this work will be performes in the FIRST cleanroom. This gives you the opportunity to get familiar with the most important semiconductor processing technologies. The finished TFTs will be electricaly characterized using state of the art characterization equipment.

Distribution: 10% literature; 70% fabrication; 20% characterization
Requirements: Interest in Micro fabrication technologies
Main supervisor: Giuseppe Cantarella, H 90, E-Mail: giuseppe.cantarella@ife.ee.ethz.ch, Telephone: +41 44 632 06 94
Second supervisor: Stefan Knobelspies, H97, E-Mail: kstefan@ife.ee.ethz.ch, Telephone: 22741
Project Title:
Professor: Prof. Tröster

Edible transistors [D-ITET, ST/MT, 1]

Progress in materials science and processing have recently enabled electronic devices with novel and unique capabilities. Research has been driven by question such as: can we realize implants which delivery diagnostics or therapeutic functions and then safely resorb in the body avoiding any additional surgery intervention? Can we design security systems which register and communicate data only for a prescribed timeframe and then completely disappear? Can we realize fieldable environmental sensors which harmless decompose to eliminate the need for their retrieval?
In 2012 the Rogers' group shows fully dissolvable, high performance and scalable silicon-based components which dissolve in water at 37°C. The palette of materials has broadened offering the perspective of biodegradable high performance electronic devices. The goal of the project is to investigate the dissolution process in IGZO and realize fully dissolvable thin film transistors.

Distribution: literature 10%, design 15%, cleanroom fabrication 50%, testing 25%
Requirements: electronics, chemistry, physics and passion for technology
Main supervisor: Giovanni Salvatore, H90, E-Mail: giovanni.salvatore@ife.ee.ethz.ch, Telephone: +41 44 632 23 77
Second supervisor: Giuseppe Cantarella, H 90, E-Mail: giuseppe.cantarella@ife.ee.ethz.ch, Telephone: +41 44 632 06 94
Project Title:
Professor: Prof. Tröster

Biodegradable sensor for envirnmental sensing and implants [D-ITET, ST/MT, 1]

Electronic integrated circuits are conceived to last as long as possible without any degradation in performance over time. Biology and living organisms, instead, evolves and changes with time. The concept of electronics devices which can be biodegradable and physically reconfigure has always fascinated the imagination of engineers. In the very recent years ?transient? electronic devices, i.e. devices which can dissolve in biological environments, have gain new momentum thanks to the progress in material science and processing. Research has been driven by question such as: can we realize implants which delivery diagnostics or therapeutic functions and then safely resorb in the body avoiding any additional surgery intervention? Can we design security systems which register and communicate data only for a prescribed timeframe and then completely disappear? Can we realize fieldable environmental sensors which harmless decompose to eliminate the need for their retrieval?
The goal of this project is to study, materials, designs and fabrication schemes to realize fully biodegradable array of temperature sensors which can be mounted on plant leaves or implanted.

Distribution: literature 10%, design 15%, cleanroom fabrication 50%, testing 25%
Requirements: electronics, chemistry, physics and passion for technology
Main supervisor: Giovanni Salvatore, H90, E-Mail: giovanni.salvatore@ife.ee.ethz.ch, Telephone: +41 44 632 23 77
Second supervisor: Giuseppe Cantarella, H 90, E-Mail: giuseppe.cantarella@ife.ee.ethz.ch, Telephone: +41 44 632 06 94
Project Title:
Professor: Prof. Tröster

Wireless soft implants [D-ITET, ST/MT, 1]

Optogenetics is a biological technique which involves the use of light to control cells in living tissue that have been genetically modified to express light-sensitive ion channels.Such technique offers unpaired specificity, accurate time and spatial resolution and find application in pain relief, rehabilitation/restoration, neural excitation. To enable sophisticated manipulation of cells with limited disruption of animal behavior, light-delivery systems beyond fiber optic tethering and large, head-mounted wireless receivers are desirable.
In this project we want to realize flexible wireless implants which are able to provide enough energy to activate micro leds with ms precision. The device will be implanted subcutaneously in mice and the mechanics will be tailored to accommodate the stress arising during animal locomotion. First prototypes and preliminary implantation experiments are ongoing in collaboration with the Institute of Toxicology at UZH.

Distribution: literature 10%, design 15%, cleanroom fabrication 50%, testing 25%
Requirements: electronics
Main supervisor: Giovanni Salvatore, H90, E-Mail: giovanni.salvatore@ife.ee.ethz.ch, Telephone: +41 44 632 23 77
Second supervisor: Lars Büthe, H61.1, E-Mail: lars.buethe@ife.ee.ethz.ch, Telephone: +41 44 632 02 93
Project Title:
Professor: Prof. Tröster

Stretchable ring-like device for health monitoring [D-ITET, ST/MT, 1]

Skin consists of a stretchable network of sensors that relay information about tactile and thermal stimuli to the brain, allowing us to maneuver within our environment safely and effectively. However, skin is not only an interface to the external world but it also represents the most intimate window on our body. So, epidermal optoelectronic devices can be also used to access vital signals in a more effective and unobtrusive manner. While a lot has been done in terms of mechanics and material processing, future efforts will focus on device engineering to achieve devices with more sophisticated capabilities.
The question we want to address in this project is simple: can we build a lab on skin for long term monitoring and provide clinical-quality to medical doctors?
The project involves challenges in electronics (interface the epidermal sensors to conventional high performance chips), mechanics (robust and intimate interface with skin), micro-technology (fabrication of sensors).

Distribution: literature 10%, design 15%, cleanroom fabrication 50%, testing 25%
Requirements: electronics, physics
Main supervisor: Giovanni Salvatore, H90, E-Mail: giovanni.salvatore@ife.ee.ethz.ch, Telephone: +41 44 632 23 77
Second supervisor: Lars Büthe, H61.1, E-Mail: lars.buethe@ife.ee.ethz.ch, Telephone: +41 44 632 02 93
Project Title:
Professor: Prof. Tröster

A flexible textile integrated MRI frontend [D-ITET, ST/MT, 1]

Magnetic resonance imaging relies on a strong static magnetic field as well as time dependent magnetic and radio frequency fields to image the Water distribution in the human body. Despite using more than 10 kW of radio frequency power to excite the Protons, the echo from the Protons, containing the image information, is in the micro Watt range. This requires the receiving antenna to be close to the body surface as well as a low noise amplifier near this antenna. Commercially available Antennas and amplifiers are rigid - in order to improve patient comfort we are developing flexible antennas and wearable electronics. Based on RF simulations it would be your task to evaluate designs for a flexible and textile integrated low noise amplifier as well as the corresponding detune and matching network which attaches to a flexible coil. After this evaluation you will build the most promising setup and validate its performance.

Distribution: 20% Literature study, 40% Simulation, 40% Hardware design
Requirements: Interest in radio frequency design
Main supervisor: Christian Vogt, H97, E-Mail: christian.vogt@ife.ee.ethz.ch, Telephone: +41 44 632 49 47
Second supervisor: Andreas Mehmann, H 96, E-Mail: andreas.mehmann@ife.ee.ethz.ch, Telephone: +41 44 632 52 80
Project Title:
Professor: Prof. Tröster

Litz wire with liquid metal [D-ITET, ST, 1]

Liquid core fibres that can be used in composite materials for damping are produced by EMPA. Our idea is to fill these fibres with liquid metal (eGaIn) to get conductive fibres. While a single fibre is thin and not very stretchable, a multifilament thread spun around an elastic fibre can achieve higher strains, similar to spandex (elastane) and polyester in our everyday clothing. Your task will be to design and build these fibres in collaboration with EMPA, discuss various ways of connecting them to electronic circuits and characterize the conductance, especially at higher frequencies.

Distribution: Literature: 20%, Fabrication: 50%, Characterization: 30%
Requirements:
Main supervisor: Andreas Mehmann, H 96, E-Mail: andreas.mehmann@ife.ee.ethz.ch, Telephone: +41 44 632 52 80
Second supervisor: Matija Varga, H 96, E-Mail: matija.varga@ife.ee.ethz.ch, Telephone: +41 44 632 79 61
Project Title:
Professor: Prof. Tröster

Adaptive Impedance Matching based on MOSFET [D-ITET, ST/MT, 1]

Bending and stretching of flexible antennas changes their impedance, which has to be counteracted to maximize efficiency. This so called impedance matching is usually done with reactive elements that can be varied. An alternative to varactors - voltage controlled capacitors ? are capacitor banks. A capacitor bank consists of an array of parallel capacitors that can be added or excluded from the circuit with switches. In this thesis you will design such a capacitor bank using MOSFETs. The challenge will be to achieve small capacitances in the range of pico-Farrad. After simulating the behaviour in Spice and finding a suitable design, you will implement it into a PCB. For a master's thesis, the scope of the thesis can be extended to do impedance detection.

Distribution: Literature: 30%, Simulation: 50%, Hardware implementation: 20%
Requirements: Knowledge of simulation software appreciated but not required.
Main supervisor: Andreas Mehmann, H 96, E-Mail: andreas.mehmann@ife.ee.ethz.ch, Telephone: +41 44 632 52 80
Second supervisor: Matija Varga, H 96, E-Mail: matija.varga@ife.ee.ethz.ch, Telephone: +41 44 632 79 61
Project Title:
Professor: Prof. Tröster

Quality Perception in Violin Playing [D-ITET/D-INFK, 1..2]

Acoustic and haptic feedback of a musical instrument play a central role for the quality perception of a musical instrument by a player.
In this semester project a low-cost violin or an e-violin is actuated with loudspeakers in order to change the radiation characteristics and the body vibrations of the instrument in real-time with a closed-loop system.
In a final evaluation experiment the influence on the quality perception of the instrument is assessed with the goal to improve the perceived quality of the electronically augmented violin.

Distribution: hardware (60%) / software (20%) / evaluation and documentation (20%)
Requirements: interest in music, hardware and programming beneficial
Main supervisor: Matthias Flückiger, H95, E-Mail: mflueckiger@ife.ee.ethz.ch, Telephone: 22996
Second supervisor: Tobias Grosshauser, H63, E-Mail: grosshauser@ife.ee.ethz.ch, Telephone: 0786961272
Project Title:
Professor: Prof. Tröster

Vibration Measurements at the Contact Points of Musical Instruments [D-ITET/D-INFK, ST, 1]

We are studying vibrations at the contact points of musical instruments to artificially generate and control vibrotactile feedback of musical instruments. Vibrotactile feedback has an influence on the perceived quality of an instrument and it is suggested that it plays an important role for timing, for pronounciation and for the precise control of musical expression.
This project focuses on studying and measuring vibrations at the contact points of musical instruments (e.g. piano, trumpet, violin, guitar).
Reference measurements will be made with a laser vibrometer, which will serve as a reference to work towards a mobile and portable measurement solution e.g. with a piezo-based vibration sensor.

Distribution: hardware (40%) / measurements (40%) / evaluation and documentation (20%)
Requirements: interest in musical instruments, measurements and signal processing beneficial.
Main supervisor: Matthias Flückiger, H95, E-Mail: mflueckiger@ife.ee.ethz.ch, Telephone: 22996
Second supervisor: Tobias Grosshauser, H63, E-Mail: grosshauser@ife.ee.ethz.ch, Telephone: 0786961272
Project Title:
Professor: Prof. Tröster

Finger Pressing Position Measurement in Piano Playing [D-ITET/D-INFK, 1..2]

In this semester project different approaches to measure finger pressing position in piano playing are developed and compared.
Capacitive sensing, ultrasound distance sensing and motion capturing are three promising candidates to measure finger pressing force in piano playing. All of them have different advantages and disadvantages. In this project the different approaches will be implemented on a single key piano action model and evaluated against each other.

Distribution: hardware (40%) / software (40%) / evaluation and documentation (20%)
Requirements: interest in music, hardware and software beneficial
Main supervisor: Matthias Flückiger, H95, E-Mail: mflueckiger@ife.ee.ethz.ch, Telephone: 22996
Second supervisor: Tobias Grosshauser, H63, E-Mail: grosshauser@ife.ee.ethz.ch, Telephone: 0786961272
Project Title:
Professor: Prof. Tröster

Estimation of an LTI System based on Sound and Vibration Measurements of Pianos [D-ITET/D-INFK, ST, 1]

In a recent study we were investigating vibrotactile feedback in piano playing. The vibration of piano keys were recorded with laser interferometry in several configurations. A professional pianist was playing on eleven different piano models from different manufacturers.
This semester project focuses on a part of the experiments to estimate an LTI system based on system identification algorithms. The radiated sound of the piano will be regarded as the input to the system and the vibrations on the piano keys as the output of the system.
Besides interesting insights about the properties of piano key vibrations, this project will yield to an algorithm to artificially generate vibrotactile feedback.

Distribution: software and algorithms (60%) / evaluation and documentation (40%)
Requirements: interest in software, algorithms and signal processing beneficial
Main supervisor: Matthias Flückiger, H95, E-Mail: mflueckiger@ife.ee.ethz.ch, Telephone: 22996
Second supervisor: Tobias Grosshauser, H63, E-Mail: grosshauser@ife.ee.ethz.ch, Telephone: 0786961272
Project Title:
Professor: Prof. Tröster

Development of a 2-mask fabrication process for flexible thin-film transistors [D-ITET, ST/MT, 1]

Flexible top gate thin-film transistors offer several opportunities to simplify their fabrication flow towards reduced process complexity and lower cost. In particular, if choosing an appropriate substrate material, additional layers like passivation and light-shading layer can be avoided, and with a proper choice of materials and patterning techniques, the number of lithography processes can be reduced to a 2-mask process.
However, flexible thin-film transistors in a top-gate gate configuration normally suffer from poor interface quality between semiconductor and dielectric which results in inferior electrical performance compared to a bottom-gate configuration.
This project deals with the exploration of a novel fabrication flow, which involves just two lithography photomasks. Furthermore, this process should enhance the interface quality. After a preliminary planning phase and literature review about possible material combinations and processing technologies, you will fabricate the TFTs in a state-of-the-art cleanroom facility. The fabricated devices will be electrically characterized and the performance will be compared to reference TFTs, which are fabricated with established process flows.


Distribution: Literature 15%, Clean room fabrication 55%, experimental work (measurements) 15%, Report 15%
Requirements: Interest in semiconductor process technology, clean room work, flexible electronics
Main supervisor: Alwin Daus, H 97, E-Mail: dausa@ife.ee.ethz.ch, Telephone: +41 44 632 68 70
Second supervisor: Stefan Knobelspies, H97, E-Mail: kstefan@ife.ee.ethz.ch, Telephone: 22741
Project Title:
Professor: Prof. Tröster

How much do you tremble? [D-ITET/D-INFK, 1..2]

In the context of monitoring patients (e.g., multiple sklerosis) at home, one question is to assess the severity of physical symptoms remotely - without in-office visits. In this thesis, you will use an IMU mounted at the finger and detect and rate trembling(none, liitle, medium, strong) and different movements of the hand with the goal to identify qualitative differences in the way of moving. To do so, you will use an existing tool chain for data recording and and compare different methods with respect to performance.

Distribution: 20% literature review, 40% design and implementation, 40% analysis
Requirements: motivation, python, wearable systems
Main supervisor: Vanessa Klaas, 97, E-Mail: vklaas@ife.ee.ethz.ch, Telephone: 20272
Second supervisor: Alberto Calatroni, H63, E-Mail: alberto.calatroni@ife.ee.ethz.ch, Telephone: +41 44 632 33 91
Project Title:
Professor: Prof. Tröster

Analysis of real-life data from remotely monitored patients [D-ITET/D-INFK, 1..2]

In order to improve the care of ambulatory cancer patients and patients suffering from cancer related fatigue (a chronical state of exhaustedness caused by the illness and its treatment), we examine the application of wearables to monitor the patients. Out of the collected data from smartphones and wristband, we want to detect changes in the behaviour and correlate this to changes in the symptoms like distress, pain and fatigue.
In studies with cancer patients, we collect data concerning physical and social activity using smartphones and wristbands. Since the patients are severly ill and have already a heavy burden, we cannot ask them to annotate their data. Therefor, we have to deal with lots of unlabeled data and little of labeled data.
In this project, you apply unsupervised methods to detect trends over time in single subjects.

Distribution: 20% literature review, 40% design and implementation, 40% analysis
Requirements: motivation, python, wearable systems
Main supervisor: Vanessa Klaas, 97, E-Mail: vklaas@ife.ee.ethz.ch, Telephone: 20272
Second supervisor: Alberto Calatroni, H63, E-Mail: alberto.calatroni@ife.ee.ethz.ch, Telephone: +41 44 632 33 91
Project Title:
Professor: Prof. Tröster

Performance Parameter Tracking for Musicians [D-ITET/D-INFK, 1..2]

Performance tracking is widely used in many sport activities and more and more in everyday scenarios.
Performance in music is not touched yet - although performance in this field is not only measuring single parameters, it's measuring complex coherences, changes in long term observation or even in performances on stage.
In these areas, performance can mean for example how synchronously the hands swirl over a guitar, or how bright or pure a tone is produced with a brass instrument.
Your task will be to create a software tool where users can analyze and track their music-making activities over time. The user tracking can be done in manifold ways, starting from simple manual input up to automatic activity recognition. The latter will be based on feature extraction, analysis and classification of audio only, combinations of audio and sensor data and sensor data only. A set of sensors for musical-instrument-related applications are readily available in our WearLab, starting from an augmented violin up to sensor setups for piano, brass instruments and drums.

Distribution: SW 35%, HW 35%, Documentation and Evaluation 30%
Requirements: Matlab, interest in building HW and Music beneficial
Main supervisor: Tobias Grosshauser, H63, E-Mail: grosshauser@ife.ee.ethz.ch, Telephone: 0786961272
Second supervisor: Alberto Calatroni, H63, E-Mail: alberto.calatroni@ife.ee.ethz.ch, Telephone: +41 44 632 33 91
Project Title:
Professor: Prof. Tröster

HighTech Musical Instruments - New Ways of Making Music [D-ITET/D-INFK, 1..2]

We look for student(s) interested in programming and sensing, further in machine learning, app programming, music or musical instruments.
In the last years, we developed different wireless sensors for performance measurements in the area of music making with musical instruments. With these sensors and software tools we measure 3D-motion, posture and forces applied while music making. On the other hand, several apps for data visualization and recording are developed for performance analysis of musical instrument playing. With all these technologies, there are 3 main application fields:
- new methods to learn musical instruments
- building completely new musical instruments and new interfaces for music making and sound production
- augmenting existing musical instruments to enhance the musical expression and performing possibilities.
If necessary, project details will be discussed with music students or professional musicians in advance to define a concrete final application.

Distribution: Hardware (40% or less), software (30% or more), evaluation and (30%)
Requirements: Interest in hard- or software, further programming and signal processing beneficial.
Main supervisor: Tobias Grosshauser, H63, E-Mail: grosshauser@ife.ee.ethz.ch, Telephone: 0786961272
Second supervisor: Alberto Calatroni, H63, E-Mail: alberto.calatroni@ife.ee.ethz.ch, Telephone: +41 44 632 33 91
Project Title:
Professor: Prof. Tröster

Music and Digital Score [D-ITET/D-INFK, 1..2]

In the last years, many technologies for musical instruments and music making have been developed, including score following. This consists in algorithms which are able to detect where a sound is to be placed within a music piece or a score and to track the position within the piece/score. Score following has been object of various research endeavors all over the world, also at the IfE within ETH.
Based on the existing framework of several working algorithms in this field, the main improvements and interests are in features/methods which are robust at the same time against tempo variations and noise in the recording and the lack of an integrated system which can blend and annotate existing scores (e.g. PDF files or scans of paper-based scores) with score-following technologies. In this ST or MT thesis we will tackle some of the mentioned points but also musicians should be able to quickly understand and make sense of what is displayed on traditional scores. This will require e.g. further development of a graphical user interface which allows to display traditional scores, overlaid with information about the point in the score, in a way that is satisfactory for musicians. Evaluations will be done with musicians.

Distribution: 10% sensors, 60% algorithms and 30% evaluation
Requirements: signal processing or data mining, beneficial c/c++ or python
Main supervisor: Tobias Grosshauser, H63, E-Mail: grosshauser@ife.ee.ethz.ch, Telephone: 0786961272
Second supervisor: Alberto Calatroni, H63, E-Mail: alberto.calatroni@ife.ee.ethz.ch, Telephone: +41 44 632 33 91
Project Title:
Professor: Prof. Tröster
 

Wichtiger Hinweis:
Diese Website wird in älteren Versionen von Netscape ohne graphische Elemente dargestellt. Die Funktionalität der Website ist aber trotzdem gewährleistet. Wenn Sie diese Website regelmässig benutzen, empfehlen wir Ihnen, auf Ihrem Computer einen aktuellen Browser zu installieren. Weitere Informationen finden Sie auf
folgender Seite.

Important Note:
The content in this site is accessible to any browser or Internet device, however, some graphics will display correctly only in the newer versions of Netscape. To get the most out of our site we suggest you upgrade to a newer browser.
More information

© 2017 ETH Zurich | Imprint | 21 October 2014
top