Smart Cities and Mobility Health Ecosystem
Public Services Oriented Architecture
Business Process Management Business Intelligence & Analytics
Cloud Computing  

Methods for Inferring Human Daily Life Activities from Sensing Data in Real-Time
Authors: João Paiva Cardoso
João Mendes Moreira
João Gama (DEIC)
The automatic identification of daily life human activities is of paramount importance for context-aware systems, health monitoring, etc. In this project we will research semi-supervised methods to classify daily life human activities using streaming sensing data obtained by sensors embedded in vests/mobile devices. Those sensors include: 3-axis accelerometers, temperature, light, sound, GPS, ECG, and virtual sensors such as time and calendar. We will use data collected representing daily life human activities bearing in mind their ultimate implementation in embedded systems (such as PDA or smartphone). We will focus on classifying activities, e.g., “running”, “sleeping”, “at the office”, etc.                
IBM CAS Portugal
IDMEC - Pólo FEUP, 6º andar
Rua Dr. Roberto Frias
4200-465 Porto

P: +351-22 508 1639 / 2133
F: +351-22 508 1538