Massive Sensor Data Management in Elderly People Monitoring Systems

In an Elderly People Monitoring System (EPMS), each monitored person is equipped with one or more sensors, monitoring his/her health conditions and other status. All the sensor sampling data are sent to the data center for centralized storing, processing, analyzing, and emergency handling. An EPMS can monitor 50,000 or more people at the same time.

In EPMS however, we also face a lot of challenges. First, an EPMS may contain various kinds of sensors whose sampling data have heterogeneous data structures. Second, queries about sensor data cannot be finished by keyword search only. In a lot of cases, spatial-temporal computations are involved in query processing. Third, the data to be managed in EPMS are massive and dynamically changing stream data.

To deal with the above problems, we propose the "SeaCloudDM" - a Sea-Cloud Data Management Framework for Sensor Data. The framework is composed of four layers: the sensor deployment layer, the sea-computing layer, the cloud data management layer, and the data analysis and application layer. SeaCloudDM can manage the historical and present sampling data from various kinds of sensors involved in elder people monitoring systems, and also, it can conduct fast query processing and efficient online analysis based on the sensor data. A lot of triggers can be defined in SeaCloudDM so that alarms can be sent whenever emergency occurs, which can be very helpful for elderly people monitoring.

Even though the SeaCloudDM is currently used for EPMS, it is designed as a fundamental supporting module, aiming to provide general support for sensor-based systems. Therefore, we cordially solicit potential collaborations in this area for further research.

-> Download presentation