A low cost posture monitoring system based on an ambient light sensor network

This paper applies wireless sensor networks to monitoring sitting condition (presence and posture). Specifically, we develop a sitting condition monitoring system based on an ambient light sensor network. As the system is hard to be modeled, a feature learning experiment has been conducted to learn about the features to design a classifier. We conducted an evaluation experiment in five different environments. Our experiment results show that our system has an accuracy around 82%, and it is robust to five different environmental noise. Keywords-sensor networks; sitting condition; monitoring