DATA PROCESSING AND MANAGEMENT ASPECTS OF WIRELESS SENSOR NETWORKS FOR STRUCTURAL HEALTH MONITORING

Wireless sensor networks consist of a host (data logging and configuration unit), representing the data sink, and many nodes distributed over a structure, representing the data sources. Each node is equipped with sensors, a digital signal processing unit and a radio transceiver. In wireless sensor networks, data processing and management issues arise already at the network nodes. The software running on each node schedules the sensor readings, acquires the raw data, performs a part of the data processing and manages the data communication within the network. For medium and long term monitoring applications, the power management of the network influences the data handling. Since communication is the most energy consuming task, a significant data reduction has to be achieved on the nodes to transfer as little data as possible through the network. This data reduction has to be performed with very limited computational and data storage resources on the nodes. The paper presents and discusses the software architecture for data processing and data management of a prototype wireless sensor network that is designed for long term monitoring of civil structures. The discussion is focused on monitoring of dynamic processes requiring high data acquisition rates. In these cases, data processing and management are key issues for an effective application of wireless sensor networks.