Search System for Behavior Time Segments from Accumulated Sensor Data in Room Environment

This paper proposes a search system for a database that stores heterogeneous sensor data in home environment. In the system, sequential sensor data are treated as time segments. The time segments are generated with Flexible representation among programs for time segment generation, database table and tokens in query. Various search and aggregation of time segments about daily behavior (ex. average of sleeping time for every week and frequency of watching TV when sitting) are realized with the simple query language for generated time segments. We demonstrated the search system with a large amount of sensor data accumulated in our Sensing Room, which is a room including many heterogeneous sensors.