University of Twente Student Theses


Device-free sensing and deep learning : analyzing human behavior through CSI using neural networks

Klein Brinke, J. (2018) Device-free sensing and deep learning : analyzing human behavior through CSI using neural networks.

[img] PDF
Abstract:Monitoring human beings is becoming a more pressing matter in this society. The desire to monitor ourselves becomes apparent by how fast different technologies are being adapted into monitoring systems. Device-free sensing is one of these new monitoring systems. This is a method that allows the monitoring and localization of humans by measuring the characteristics of these wireless signals. A method to do so is by looking at the way these signals propagate through the channel state information (CSI). Deep learning is an upcoming field within machine learning used to classify human activities. It attempts to replicate the way humans perceive information. Of all the techniques within deep learning, convolutional networks have proven to be capable of dealing with signal processing, yet it has not been widely adopted in human activity recognition through CSI. This research presented in this thesis looks into how convolutional networks perform against current state-of-the-art systems by analyzing both static and dynamic activities. The experiments performed in this research were conducted with multiple participants over three days to investigate the scalability. The findings indicate that for dynamic activities, convolutional networks achieve higher accuracies than static postures.
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science MSc (60300)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page