Department of Statistics, The Ohio State University
Statistics and Biostatistics Colloquium Series

Wireless sensor networks (WSN) are a promising technology for structural health monitoring. Detecting a change in the state of a structure is often achieved by monitoring features of measured response, e.g. the natural vibration properties of the structure. These properties are estimated in this paper by an AR model fitted to the acceleration response of a bridge to ambient vibrations. Physically motivated restrictions on the multivariate AR model are introduced to reduce the variability of the estimated parameter for a small sample size, and to reduce the data traffic over the WSN thus extending its lifetime. Simulations of linear vibrating systems subjected to ambient excitations suggest the modal estimates based on the restricted AR model perform as well as those based on the full AR model. The trade-off between data traffic in the network and estimation accuracy is also studied.

Meet the speaker in Room 212 Cockins Hall at 4:30 p.m. Refreshments will be served.