An Agent-Based Architecture for Sensor Data Collection and Reasoning in Smart Home Environments for Independent Living

There has been a tremendous growth in new sensor technology and wireless, handheld computing devices that give rise to new opportunities for smart home applications. With advances in the technology, reduced costs of operation, and the ubiquity of WiFi and cellular networks, the reach and value of smart home technology is growing giving rise to applications of smart home systems to support independent living of the elderly. Such systems must collect and analyze data in the home to recognize unusual activities and alert care givers and/or family members in emergency situations. To address this challenge we developed a flexible and scalable multi-agent system for sensor data collection, integration, processing, and alert management. A prototype system for activity recognition has been built and preliminary tests demonstrate that activities can be successfully recognized based on data captured in the home. As a next step, the sensor network will be deployed in an inpatient residence facility to collect real-world data for evaluating the system’s performance and to develop new applications.