Luxtrace

Type: Semester thesis
Student: Martin Burri and Christian Morf
Advisor: Dr. Julian Randall, Oliver Amft

In the domain of wearable computing determination of the users position and activity is an important issue. Since GPS isn't reliable indoors new methods have to be found. To this day solar cells are only used as energy harvesting sources for such systems. LuxTrace I focusses on how solar cells can provide information of the users environment and support other location tracking systems. Whilst other systems demand installation of a specific infrastructure, LuxTrace benefits from knowledge of the environment, no additional installation is needed. It's based on solar modules which are cheap, light, low volume and can be integrated into clothes. This work investigates their characteristics in a simplified scenario. A theoretical model of the environment and the cells is built and validated by measurements. In a second step a trained model based on measured data is developed for distance and speed determination. Distance estimations in a simple case indicate that a empirical model for a particular scenario achieves an accuracy of 18 cm with a confidence of 83 %.