Sensor data based user activity modeling with deep neural networks

OData support
Supervisor:
Dr. Gyires-Tóth Bálint Pál
Department of Telecommunications and Media Informatics

Nowadays portable devices - like smartphones, smartwatches - have become an important part of our daily lives. These devices contain increasing amount of personal information and we are not only using them to communicate, but during work and other daily activities too. The currently available smart devices can record a great deal of sensor data, which can be used estimate the user's current activities. The recognised behavioral patterns can be used to improve human-computer interaction by automatically performing the neccesary steps in the current context of the user, instead of manual execution.

In my thesis I reviewed the scientific literature connected to my topic of research. I developed a Google Android based data-recording application which I used to record sensor data when executing different activities from 9 person. I studied the theory behind Deep Neural Networks. I used the recorded data to build models based on Deep Neural Networks to estimate the user's current activity. According to tests, the system can recognise the user's activity with mean accuracy of 94%.

Downloads

Please sign in to download the files of this thesis.