Semester and Master theses
Wearable Sports assistant - FS12

---

Making Rowing Boats Faster: Performance Metrics [D-ITET/D-INFK, 1..2]

Rowing is a very complex sport which strongly relies both on physiological strength and technical skills.
No matter if you have already watched a rowing event at the Olympic Games, the annual Uni-Poly-Match here in Zurich or you even have own experiences in rowing boats - the main question for rowers and spectators is: What exactly makes a boat fast?
Our project aims at equipping state-of-the-art rowing boats with miniaturized motion sensors to investigate possibilities to improve rowing technique and speed. We are partnering with Swiss and German pro rowers and coaches.
The needed hardware (miniaturized acceleration sensors) is already available and tested in a pre-study.
Your tasks are:
- Find the locations for the sensors on the boat that lead to best results
- Record data of rowers ranging from beginners to professional rowers
- Analyze recorded data and find quantitative measures to describe different movement patterns and common mistakes in rowing technique.
Finally, the goal is to reveal relationships between movement measures and rowing technique/experience.
Your work will find an answer to the question: Which measures qualify for supporting a coach in improving the rower's technique?

Project Website: http://www.wearable.ethz.ch/research/groups/sports/rowing


Distribution: 40% Experiments, 50% Theory and Signal Processing, 10% Optimization
Requirements: Programming, Matlab, Rowing basics would be beneficial but not necessary
Main supervisor: Franz Gravenhorst, H97, E-Mail: gravenhorst@ife.ee.ethz.ch, Telephone: +41 44 632 76 41
Second supervisor: Bernd Tessendorf, H96, E-Mail: tessendorf@ife.ee.ethz.ch, Telephone: 25191
Project Title: Project Links
Professor: Prof. Tröster

---

Making Rowing Boats Faster: Framework and Algorithm [D-ITET/D-INFK, 1..2]

Rowing is a very complex sport which strongly relies both on physiological strength and technical skills.
No matter if you have already watched a rowing event at the Olympic Games, the annual Uni-Poly-Match here in Zurich or you even have own experiences in rowing boats - the main question for rowers and spectators is: What exactly makes a boat fast?
Our project aims at equipping state-of-the-art rowing boats with miniaturized motion sensors to investigate possibilities to improve rowing technique and speed. We are partnering with Swiss and German pro rowers and coaches.
The needed hardware (miniaturized acceleration sensors) is already available and tested in a pre-study. You can either record own data or use available data sets. Your job is to develop a segmentation algorithm which detects the different phases of a rowing stroke based on the recorded sensor data. Finally, you implement a framework with a GUI to visualize and evaluate your algorithm by comparing its results to reference segmentations which are performed manually.
With the help of the framework you can vary your segmentation approach (such as simple signal thresholds or wavelet filters) and optimize the applied parameters.

Project Website: http://www.wearable.ethz.ch/research/groups/sports/rowing



Distribution: 30% Signal Processing, 50% Framework Implementation, 10% Segmentation Algorithm, 10% Optimization
Requirements: Programming, Matlab, Rowing basics would be beneficial but not necessary
Main supervisor: Franz Gravenhorst, H97, E-Mail: gravenhorst@ife.ee.ethz.ch, Telephone: +41 44 632 76 41
Second supervisor: Bernd Tessendorf, H96, E-Mail: tessendorf@ife.ee.ethz.ch, Telephone: 25191
Project Title: Project Links
Professor: Prof. Tröster

---

Modelling running movement to predict sensor readings [D-ITET/D-INFK, 1..2]

Wearable motion sensors such as miniaturized Inertial Measurement Units (IMUs) allow for unobtrusive monitoring of athletes in the field. The gathered sensor data can be used to identify typical mistakes in running technique or to assess skill level. With a movement model we aim at predicting sensor readings of running gait. Simulating e.g. typical errors in running technique allows to develop algorithms that are robust across genders, age, speed etc. and to identify important sensor positions.
Your task will be to integrate a framework predicting sensor readings from Motion Capture data into our framework and to develop a model of running gait using animation software (e.g. 3ds max, Blender). A case study (e.g. a typical running mistake) will be investigated to validate your model.
Code, sensors, and data from previous running experiments will be provided to support your work.


Distribution: Algorithms: 60% - Math: 20% - Experiments: 20%
Requirements: Matlab, signal processing, linear algebra (Python is beneficial)
Main supervisor: Christina Strohrmann, H 64, E-Mail: strohrmann@ife.ee.ethz.ch, Telephone: +41 44 632 05 44
Second supervisor: Franz Gravenhorst, H97, E-Mail: gravenhorst@ife.ee.ethz.ch, Telephone: +41 44 632 76 41
Project Title:
Professor: Prof. Tröster

---

RunConomy: Development of an Android App to predict a runner?s economy [D-ITET/D-INFK, 1]

Expended mechanical energy in running can be assessed with a simple spring/mass model. The state-of-the-art approach implies the use of a kinematic arm that is fixed to a treadmill to estimate the expended mechanical energy during running. To improve running economy, mechanical energy should be reduced while maintaining running velocity.
Your task will be to investigate the use of body-worn sensors to estimate the expended mechanical energy during running. Sensor modality, sensor positioning and accuracy will be investigated. Your final solution will be implemented on an Android phone. A user study will reveal if your App helps runners improve their running economy.


Distribution: Algorithms: 60% - Experiments: 40%
Requirements: Matlab, Android programming, interest in sports
Main supervisor: Christina Strohrmann, H 64, E-Mail: strohrmann@ife.ee.ethz.ch, Telephone: +41 44 632 05 44
Second supervisor: Franz Gravenhorst, H97, E-Mail: gravenhorst@ife.ee.ethz.ch, Telephone: +41 44 632 76 41
Project Title:
Professor: Prof. Tröster

---

Orientierungsberechnung von körpergetragenen Sensoren [D-ITET/D-INFK, 1]

Körpergetragene Sensoren ermöglichen durch Messen der Bewegung, z.B. im Sport, eine Bewegungsanalyse, welche nicht an äußere Gegebenheiten (z.B. ein mit Infrarotkameras ausgestattetes Labor) gebunden ist.
Typischerweise werden hierzu IMUs (Inertial Measurement Units - Inertialsensoren) verwendet. Diese beinhalten einen Beschleunigungssensor, einen Drehratensensor und einen digitalen Kompass. Beschleunigungssensor und Magnetfeldsensor können die Orientierung im Raum sehr gut bestimmen, solange die Bewegung statisch ist. Bei dynamischeren Bewegungen wird daher die Schätzung durch die Messung des Drehratensensors ergänzt.
Deine Aufgabe besteht darin, einen solchen Algorithmus in einem Sportszenario (z.B. Laufen) zu testen und diesen anschließend online auf dem Sensor zu implementieren.


Distribution: Algorithmen: 80% - Experimente: 20%
Requirements: Matlab (C++ und Mikrocontroller-Programmierung von Vorteil)
Main supervisor: Christina Strohrmann, H 64, E-Mail: strohrmann@ife.ee.ethz.ch, Telephone: +41 44 632 05 44
Second supervisor: Franz Gravenhorst, H97, E-Mail: gravenhorst@ife.ee.ethz.ch, Telephone: +41 44 632 76 41
Project Title:
Professor: Prof. Tröster