Marc Mueller

My Projects

• safety and reliability research through stochastic modelling

  In the context of the research project Stella at the University of Applied Sciences Bonn-Rhein-Sieg, I developed a universal stochastic model to identify safety- and reliability critical states of energy storage devices. In account to a realistic quantification of these states, the model covers up the control and power electronics, the electric circuit, the communication bus, the sensors, the battery management system and the motor controller with its software. Critical states while loading the energy storage devices are detected in real-time and passed to a fault preventing system. The Objective of the project was to act before the energy storage devices react irreversibly to its environment. Furthermore the model will be integrated to a framework consisting of standards and recommendations to increase the safety. To mathematical verify the model, it was formalized and CTL model checking was applied to it.

  • ClientUniversity of Applied Science Bonn-Rhein-Sieg
  • Websitewww.stella-mobility.weebly.com
  

• mobility research with signal processing

  The objective of this project was to show that the data of tracked bluetooth devices are sufficiently dense to detect commuter patterns based on a Fourier analysis.

  • ClientFraunhofer IAIS
  • Websitewww.iais.fraunhofer.de

  

• image recognition system

  A prototype web-service to find similar fashion based on uploaded images. The uploaded image is transformed to hue color space and afterwards features are builded with signal processing methods. Thus I was able to search a preprocessed image database to propose the user in nearly realtime (>1s) the similar fashion results, based on the features.

  • ClientBucktownUG
  • Websitewww.fashiongram.de
  

• bluetooth tracking device

  The objective of this project was to use a reliable sensor system based on a single-board computer (BeagleBoard and RaspberryPi) that is able to collect Bluetooth data autonomously and send the data to a server via UMTS. Furthermore, the system has a maintenance interface that is secured with a public key infrastructure. This was especially important when it comes to sensor placements in public spaces since the data has to be protected from illegal external access. Furthermore, different reliability and availability mechanisms are implemented in order to receive continuous sensor data. For example, a disconnection of the UMTS stick has to be intercepted, and after a reconnection the sensor has to send the incurred data to the server. The whole process runs autonomously since the sensor’s accessibility is limited once setup up.

  • ClientFraunhofer IAIS
  • Websitewww.iais.fraunhofer.de
  

• Fashion.Gram

  Fashion.Gram is an self developed app of me and colleague to shop popular fashion from Instagram. It enables an easily way to shop the hottest fashion trending on Instagram. Therefor fashion streams are crawled and the data is processed to present it to the user. In late 2014 there were about 4.000 Android and 20.000 iOS downloads, with more than 1.500 inapp views per day (source: google analytics).

  • ClientBucktownUG
  • Websitewww.fashiongram.de
  

Scientific Research