Automatic Composition of Sensor Data for Behavior-Driven Application in Smart Room

This paper proposes a software for sensor data composition in home environment. The software supports behavior detection for applications based on occupant´s behaviors in home environment. The software generates program components for sensor data processing with a strong name resolver for program components in network middleware and knowledge description about sensors and behaviors. Behavior descriptions are prepared for behavior detection. The behavior is defined as time-series combination of behavior elements (e.g. sitting and lying). The program components capture behavior element through sensor data processing. The program including the behavior description detects the behavior with the detected behavior element and the description. Since the program components are generated automatically according to the sensor in the environment, the behavior detection with the software adapts to the sensor configurations. The demonstration of behavior detection in sensing room and an application of information display shows automatic data processing for behavior detection in home environment.