Hi-Res activity recognition system based on EEG and WoT

Abstract

Nowadays, the recognition of physical activity (PA) is a well-known problem with many solutions. Sev eral kind of algorithms, using MEMS sensors, al low determine the most likely activity. Indeed, these applications work well when physical activity is performed for long periods of time and steadily. However, indoors, these systems are not entirely suitable and have several problems. In this paper, thanks to the introduction of new context infor mation, such as EEG, and through communication between WoT based elements interface at home, it would be possible to perform a more accurate and low-level recognition. By using uPnP proto col and additional services, information from other smart housing elements with user device itself can be shared, enriching traditional systems based on ac-celerometry.