Human daily behavior accumulation in real home environment via distributed sensors in a long term

This paper reports the case study about long-term sensor data accumulation of natural human behavior in a real home environment. Human behavior measurement in long term and in the real environment is difficult because of devices, software and system problems. We rented an apartment and introduced 4 sensor modules (pyroelectric sensor module, proximity sensor module, electric current sensor module and LRF module) into the rooms. The systems for data collection from modules, database system for long-term accumulation and sensor data visualization software for behavior annotation are also developed. We measured the sensor data during 4 years continuously. This data will assist to develop methods for typical behavior extraction, behavior recognition and anomaly behavior detection.