Decentralized Infrastructure Health Monitoring Using Embedded Computing in Wireless Sensor Networks

Due to the significant risk posed to public safety by infrastructure ageing and deterioration, infrastructure health monitoring has been drawing increasing research interest in recent years. In the field of infrastructure health monitoring, wireless sensor networks have become particularly popular, because of their cost efficiency, flexibility, and reduced installation time as compared to conventional wired systems. Using embedded computing, it is possible to process the collected data directly on the wireless sensor nodes, thus reducing the wireless communication and the power consumption. In this paper, a methodology to fully decentralize the condition assessment process is presented. The paper showcases the development of embedded algorithms and numerical models to be embedded into the wireless sensor nodes, i.e. the monitored structure is divided into substructures where each sensor node is responsible for monitoring one substructure. The proposed methodology is validated through simulations on a numerical model of a four-story shear frame structure. The objective of the simulations is to test the performance of the algorithms and the quality of the numerical models embedded into the wireless sensor nodes with respect to decentralized condition assessment, taking into consideration the effects of external factors, such as ambient noise, that usually interfere with measured data.