Abstract
To realize intelligent environments that understand and support our daily activities, it is important to observe positions and identities of people in the environment and many studies have been proposed using sensor networks. Floor sensors can reliably detect current positions of people and prevent invasion of privacy, but it is difficult to identify people. To solve the problem, we propose to integrate acceleration sensors that are attached to the human body. Since the signals from floor sensors and acceleration sensors are correlated when they observe the same person, these signals are not independent. The correlation between the signals is evaluated based on a statistical test to find correct association of positions to IDs. People tracking examples are shown to confirm the effectiveness of the proposed method. Significant improvement in correct association rate is achieved compared to the results using only floor sensors.
