Abstract: Modern mobile platforms allow for cost-effective mobile sensing. This development has boosted research in the area of social signal processing (SSP) which aims to enable computers to understand human social interaction. In this thesis we describe a research approach closely related to SSP that uses mobile smartphones for detecting social co-activity. Social co-activity is shared activity between one or more humans that are mutually aware of each other. Our system uses most of the available sensors within modern smartphones. It detects the existence of social co-activity between two persons and it detects transitions from one co-activity to another. We developed a prototypical implementation of the system that uses iPhone 4 commodity smartphones as a mobile sensing platform. Additionally, we validated our approach with an evaluation study.
Thesis PDF coming soon.