Grid computing on map-reduce architecture

Mario Miosga successfully defending Bachelor thesis

Seville, June 2015. Mario Miosga defended successfully his Bachelor Thesis at the Universidad de Sevilla. Mario is a student from HTWG Konstanz moving with the ERASMUS agreement to Seville. At the Universidad de Sevilla he implemented a map-reduce system on a comparable small hardware infrastructure (Intel Galileo) and in comparison he used the high-performance computing cluster in Seville. The thesis has been co-supervised by Prof. Dr. Juan Antonio Ortega and Prof. Dr. Ralf Seepold. There is a long successful research cooperation between both laboratories. This thesis was running also in this frame of cooperation. We congratulate Mario for his excellent work and we thank you very much our colleagues in Seville for their support.

HTWG visits MEI in Moscow

24 years of international cooperation renewed (link to the German version)

Moscow April/May 2015. Scientific cooperation has a long tradition for researchers in Konstanz and Moscow. For nearly 24 years, HTWG and the Moscow Power Engineering Institute MEI (National Research University) are running a cooperation agreement for exchanging advanced students in Bachelor or Master degrees, PhD students or even on professors level but also executing applied research projects.  

Biovital data analysis

Big data processing based on raw sensor data input

April/May 2015. Prof. Dr. Soria Morillo (Universidad de Sevilla) staying at Ubiquitous Computing Laboratory contributed to the lecture ‘Ubiquitous Computing’ forming part of a specialization in the Applied Informatics degree on Bachelor level. He reported that a remarkable portion of noise will accompany data captured with the help of a sensor that is e.g. connected to the body and measuring the heart rate.

#GirlsDay@the UC-Lab 2015

The Girls' Day 2015 was prepared with the motto Applied Informatics. The same name like the computer science bachelor degree at HTWG Konstanz. The Girls Team at the Ubiquitous Computing Lab prepared two experiments for the young group of pupils: The self-made piano and the programmable pendant.

Follow the main routes of tourists

The objective of this project is the tracking of tourists with the help of positioning systems like GPS in the area of the Lake Constance, Germany. The tracking can be realized via common Smartphones extended by an App or with dedicated sensors like position loggers. Movement patterns of tourists in the Lake destination are unknown so far. The goal is to track tourists and to visualize movement patterns. These patterns are of high importance for the tourism industry regarding the design of tourism offers and products. The faculty of Computer Science developed a smartphone app that provides access to the GPS component available in nearly all smartphones on the market. This app provides a functionality similar to a GPS-Logger but offers an interface to the user in order to provide a simple setup for personalized use. The app will track the movement of a person and push impersonalized data to a backbone server that stores the movement during a certain period (e.g. one day). Besides this app an alternative is offered in case a person done likes installing an app on the smartphone or in case the person is not using a smartphone. In this case a GPS logger is used, providing a similar functionality than the smartphone app. In the case of the logger, the data can be store on the logger without immediate push to the backbone server, while the app would use an online connection to the server in order to report the movement profile.

Stress visualized with Correlations Plots

Heart Rate Variability indicating Stress

Granada, April 2015. Stress is recognized as a factor of predominant disease and in the future the costs for treatment will increase. The presented approach tries to detect stress in a very basic and easy to implement way, so that the cost for the device and effort to wear it remain low.  

Prof. Dr. Soria Morillo discussing with researchers

Scienfic advancements in biovital data capturing

Prof. Dr. Soria Morillo (Universidad de Sevilla) visited the Ubiquitous Computing Laboratory to present his recent developments in biovital data capturing. In one of his main research areas, which is captuiring the Electroencephalography (EEG), he presented the technique of measurement and processing data with a specific sensor. The realtime analysis of the traced data supports a prediction of the behaviour. "With the help of this system, we can forecast that a driver will press the break even before he started to do so.", says Prof. Soria Morillo. Several scenarios like a driver assitance system can benefit from this development but there are also applications beyond automotive systems. In a special session the status of research with respect to biovital signal capturing and processing in both groups have been discussed. Finally, possible cooperations have been identified which are beyond the running cooperation of both groups. 

Seminar: Innovation Lab SS15

Master Informatik
(Open for HTWG and University students)
First meeting: Thursday 19 March 2015, at 17:30 in F-128.

The seminar will address different up to date topics currently discussed in research. Based on literature and/or literature and demonstrators, Master students will receive material helping to understand the new domain and typical models or systems available this field. After studying the introductory material, the students should be able to understand and summarize the domain. In a second step, the student is asked to develop a new idea based on the knowledge obtained but contributing to a gap that has to be detected. The idea will be presented and discussed before the student is asked to prepare a detailed description of the idea and a suited model to realize it, i.e. contributing to close the gap. Depending on the complexity, a prototype solution can be part of the task. At the end, the work will be presented in two parts: (i) Scientific paper and poster, and (ii) Oral presentation (and demonstration – if possible).

The objective of this seminar is to improve the ability working in a new domain based on related work, detect missing parts and develop a new idea how to address this gap. The student may use this seminar to prepare further studies (like a Master Thesis) or a fundament for setting up a business.

Proposed topics (preliminary)

  1. Mobile QRS complex analysis and stress detection
  2. Hadoop meets Android (Real time big data analysis)

Intel boards for embedded developments at UC-Lab

Empowered by Intel Galileo

Intel selected the Ubiquitous Computing Lab to receive the new Intel® Galileo Gen 2 Development Board for their projects. The donation is part of Intel’s University Program supporting proposers in applying the new board throughout different activities. The board is designed to be hardware-, software-, and pin-compatible with a wide range of Arduino Uno R3 shields.

mBiDA is supporting German-Italian Dialog

Mobile Networks for Biometric Data Analysis (mBiDA)

Ancona, November 2014. Prevention and treatment of common and widespread (chronic) diseases is a challenge in any modern society and vitally important for health maintenance in aging societies. Capturing biometric data is a cornerstone for any analysis and treatment strategy. The event Mobile Networks for Biometric Data Analysis (mBiDA) contributed with seven different activities during two days to this raising topic.

Determining stress with the QRS complex

QRS complex detection, analysis and pre-processing to determine stress

Ancona, 31 October 2014. Daniel Scherz presented a algorithmic calculation for stress detection. The system is designed to enable real time monitoring of stress by using a mobile, cheap and light-weighted device. The results have shown that stress can be detected even in the case that people may have a naturally low or a height heart rate. The prototype can be connected to different systems (PCs or smartphones). The algorithm for stress detection has been implemented with a small footprint and it can be deployed in prototyping board, for example, an Arduino Uno has been used. The system has been tested while doing sport activities but the interference generated by the movement of the electrodes generates artefacts in the signal, which has to be filtered or corrected. 

Dynamic threshold calibration for RR interval

Dynamic threshold calibration for RR interval detection in a QRS complex

Ancona, 31 October 2014. These days the computer analysis of ECG (Electrocardiograms) signals is common. There are many a real-time QRS recognition algorithms. One of these algorithms is Pan-Tompkins Algorithm, where it is working to detect QRS complexes of ECG signals. The algorithm will analysis the data stream of the heartbeat based upon the digital analysis of amplitude, bandwidth, and slope. In addition to Mohammed Rajab proposes a new algorithm has to compare whether the current heartbeat is similar or different to the last heartbeat after detecting the ECG signals. So the algorithm will determine the stress and stress detection for the patient in real-time. In order to implement it, the algorithm is modeled in C language using different function as beat detector and beat classifier of stress. His results show that it can be used for correlation detection between stress and the heartbeat.

Intelligent navigation routes

Intelligent generation of navigation routes based on individual stress level

Ancona, 30 October 2014. The impact of stress of every human being has become a serious problem. Reported impacts on persons are a higher rate or health disorders like heart problems, obesity, asthma, diabetes, depressions and many others. An individual in a stressful situation has to deal with altered cognition as well as an affected decision making skill and problem solving. This could lead to a higher risk for accidents in dynamic environments such as automotive. Different papers faced the estimation as well as prediction of drivers’ stress level during driving. Another important question is not only the stress level of the driver himself, but also the influence on and of a group of other drivers in the near area.

News Archive