TU Berlin

Service-centric NetworkingInternet of Services Lab - ST2016


zur Navigation

Es gibt keine deutsche Übersetzung dieser Webseite.

Internet of Services Lab (ST2016)


This SNET project is about carrying out a whole project. A given topic will be analyzed, specified, implemented, tested and documented. These topics can vary within the following groups of related subjects:

  • Mobile Computing
  • Context-aware Computing
  • Linked Data
  • Data Science and Analytics
  • Digital Communities
  • Cloud Computing
  • Web Services

Depending on the topic, the project will be supervised by different SNET researchers, T-Labs employees or other guest lecturers. At the beginning of the semester, the participating students can apply for one of the specific tasks, which will be announced at the beginning of the semester in the first lecture. Then the supervisor will define groups and regular meetings with the students, explain the task and provide materials and tools. During the semester, each participant of the project has to present their topic, progress and final results in different talks.

The course is conducted in English and requires programming knowledge

Target Group

This compulsory course (Wahlpflichtfach) addresses Master students of

  • Computer Science (Informatik)
  • Computer Engineering (Technische Informatik)
  • Business Informatics (Wirtschaftsinformatik)
  • EIT Digital
  • Compulsory Elective in other degree programs possible if course is not full


After organizing in groups, the main part of the project is managed by you. Additionally you regularly meet with your supervisor (normally once a week or bi-weekly) and she/he will assess your advance. Furthermore, there are two mandatory milestone appointments, where you present your progress and results in front of the whole group.


The introductory lecture will take place at TEL20 Auditorium 1. Note: The lecture will start on sharp 10:00 (s.t.). Please be on time. The date for the Milestone workshop will be announced at the introductory lecture.

Schedule and Location
Day of week
Start/End date
Introductory Lecture
10:00 s.t. - 12:30
April 19, 2016
TEL20 Auditorium 1
Milestone Workshop
H 3004
Final Workshop
14:00 - 17:00
July 20th, 2016
H 3004


The method of examination is the portfolio exam ("Portfolioprüfung"). All in all 100 portfolio points can be achieved:

  • Practical implementation (25 portfolio points)
  • Process of the project (25 portfolio points)
  • Presentation (25 portfolio points)
  • Written report (25 portfolio points)

The final grade under the terms of § 47 (2) AllgStuPO is calculated according to the grading scheme no. 2 of faculty IV. Attendance of all appointments is mandatory.

Application and Distribution

Each topic requires different number of participants and is limited in general. Students with substantial knowledge in Java or other object-oriented programming languages can register for this course on ISIS until April 18th, 2016.

At the introductory lecture, the topics (will be announced on ISIS beforehand) will be presented and any registered student may choose a topic. Early registration gives you the opportunity to choose the topic first. However, in any case you are not entitled for any topic beforehand.

Project Topic Proposals

Here you can find a tentative list of topics for the summer term 2016. For a more detailed description see ISIS.

Automatic Profiling

Through sensors and webservices, smartphones are able to unobtrusively collect data and describes its user in a profile. This profile, or parts of it, could then be used for further services that for example give recommendations. In this project, you will extend an existing Android App implementation that does such profiling, logging music listened to, locations, weather, and movement. The goal is to show the user in appropriate, interactive, and fun ways, what profile data the phone has about her and lets her select sub-sets of the data for further use.


Inferring transportation modes from mobile network trajectories 

The goal of this project is to infer what mode of transportation (i.e. public transport, subway, s-bahn, bus, car, bike, walking) was chosen by a person throughout the day – based only on a series of mobile network towers a user has been connected to. You will be given Android code that collects the mobile network information. This project will require you to extend the android application, allowing users to label data. (i.e. a user traveled by subway from 9:30 to 10:15. With the labeled data you collect yourself, you will then try to predict the labels with machine learning algorithms and evaluate its accuracy on differing features and algorithms.


Graphical HATEOAS Representation for REST API Development 

An important aspect of Restful APIs is the concept of “Hypermedia as the engine of application state” (HATEOAS). In this project you will develop a graphical REST API prototyping system (like Restlet Studio, studio.restlet.com/apis/local) as a web application and include the concept of HATEOAS.


Building a platform for analysis of driver behavior 

The objective of this topic is to develop a user interface for querying and visualizing data generated by automobile sensors.  The final product should be a web-based dashboard for visualizing trends among the different data and for identifying how drivers react to various situations on the road.  Students will gain experience in building a front-end user interface that connects with back-end querying platform, using of state-of-the-art technologies for both querying and visualizing data.


Deep Learning for Image Classification

With the help of open source libraries like R and tensorFlow we will use deep learning and boosting algorithms to classify image data and compare the results.


Yelp datasets: Comparing friendships and similarity graph

Yelp has made public lots of its data. We will use it to create a similarity graph of the yelp users using LDA. This is then compared to the friendship graph using readily available methods from graph theory.


Privacy-aware Geolocations Using Variable Geofences and Proximity-based Geocoding 

The whereabout of a person should be revealed to a sender. The location of the person changes over time. The sender should get information about the person’s whereabout depending on his distance to the target. If the sender is in the same city as the person a more detailed location, e.g., the district or street name is revealed. If the sender moves closer to the person, e.g., the exact GPS location is disclosed. The project outcome should consist of a geocoding application, a mobile app, and a server component. Technologies, such as (Geo)XACML policies, Nominatim, and Android OS and/or iOS, as well as other components developed at SNET, should be used/included.


Continuous and on-demand opportunistic crowdsensing with Smartphones

Extend a mobile application to collect sensor data on smartphones in a continuous and on-demand manner, develop a backend that receives and stores the sensor data and create a web site for the visualization of sensor data and control of the crowdsensing process. 


The topics are still subject to change. Depending on the number of students it is possible that a project is not conducted at all.



Schnellnavigation zur Seite über Nummerneingabe