Low intrusive ehealth monitoring: Human posture and activity level detection with an intelligent furniture network

Assisting elderly people living at home is a topical issue for Information and Communication Technology (ICT) developers. The motivation is in tracking the resident behavior and detecting abnormal living patterns. We take an approach for such an eHealth monitoring by an intelligent furniture network. Human behavior in the form of postures and activity levels is monitored using a set of intelligent furniture with very low cost low-intrusive capacitive proximity sensors. The sensor system relies on wireless sensor network technologies and is extended with data management and monitoring user interfaces via the internet. Our experimental tests show that compact algorithms based on nearest neighborhood classifiers and filter banks with Infinite Impulse Response (IIR) filters or Haar wavelets can identify the state of the furniture user in the form of postures and activity levels. Changes in posture and activity patterns can reveal behavioral anomalies, like restlessness and wandering, indicating possible health related unrevealed complications.