MobilitApp project: user transportation activity recognition via mobile device sensors
Tutor / director / evaluatorAguilar Igartua, Mónica
Document typeMaster thesis
Rights accessRestricted access - author's decision
The thesis presents our developments of algorithms and software for detecting means of transportation of citizens using motion data from the sensors built in Android smart phones. First, we worked on adding sensors data logging functionalities to an Android application called MobilitApp. Second, we worked on logging sensors data via using MobilitApp?s new functionality. Third, using the data collected, we worked with machine learning concepts to be able to detect the specific transportation activities. Eventually, we assessed multiple solutions to integrate the machine learning logic implementations with Android operating system.
MobilitApp project (http://mobilitat.upc.edu//): We offer Master/Degree thesis. The student will join the MobilitApp project that we develop in collaboration with the ATM (Autoritat del Transport Metropolità) of Barcelona. MobilitApp is able to detect the type of transportation and take statistics. Current on-going tasks: send an emergency message upon the occurrence of an accident, suggest alternative types of transportation, among other features. Sensor data from the smart phones is gathered and periodically sent to our Raspberry Pi server, where it is analysed. The goal