Posture determination using a body sensor network

Due to the large number of degrees of freedom of the human body, posture monitoring of human during activity regimes presents many research challenges. Several research groups world wide have engaged with the development of low-power wireless body sensor networks that are capable of providing real-time posture tracking for a variety of applications, such as dance and sports. The work reported here is concerned with the development of a wireless body sensor network that, as opposed to posture tracking, can: a) provide the identification and classification of eight human postures (standing, kneeling, sitting, crawling, walking, laying down on front and back, and laying on one side) in real-time and b) is able to relate this information wirelessly to a remote monitoring point. Posture information is an essential part of monitoring operatives in safety critical missions. The work sits within a larger project aiming to increase general safety of operatives in bomb disposal missions. The goal of the posture body sensor network developed here is to identify the eight named postures using data from nine accelerometers placed at various sites on the human body. A prototype implementation which fulfills the goal has been produced and evaluated and is reported.