In order to develop such an integrated framework, three composite bridge decks on the UCSD (University of California, San Diego) campus, instrumented with heterogeneous sensors [4] including a video camera [5], are employed as a testbed for a pilot study. Live data from selected sensors and the video feed are now available on-line on a 24/7 basis over a web-site for worldwide access at bridge.ucsd.edu/compositedeck.html.

As an important element in the integrated framework, a traffic pattern recognition system is developed in this paper. Using features extracted by the Principal Components Analysis (PCA) technique [6], a neural network [7] is employed in this system to identify the classes of crossing vehicles based on changes in traffic-induced strain response. For this purpose, neural network training is accomplished in a supervised-learning mode [7], which requires a large number of labelled patterns. Considering that data labelling is labor-intensive and time-consuming, an active learning procedure [8,9] is proposed to provide a more efficient solution for practical implementation.

Using the real data collected from the UCSD bridge deck testbed, comparison studies were performed to test and validate the proposed approach. Final results show that a neural network involving PCA is a feasible and efficient solution for strain-based vehicle identification. In addition, the time, labor, and expense for data labelling can be significantly reduced by taking advantage of the developed active learning procedure.

1

E. Aktan, S. Chase, D. Inman, and D. Pines, "Monitoring and Managing the Health of Infrastructure Systems", Proceedings of the 2001 SPIE Conference on Health Monitoring of Highway Transportation Infrastructure, SPIE, March 6-8, 2001.

2

A. Elgamal, J.P. Conte, S.F. Masri, M. Fraser, T. Fountain, A. Gupta, M. Trivedi, and M. El Zarki, "Health Monitoring Framework for Bridges and Civil Infrastructure", Proc. of the 4th International Workshop on Structural Health Monitoring, Stanford University, Stanford, CA, September 15-17, 2003.

3

A. Elgamal, J.P. Conte, L. Yan, M. Fraser, S.F. Masri, M. El Zarki, T. Fountain, and M. Trivedi, "A Framework for Monitoring Bridges and Civil Infrastructure", (Invited Paper) Proc. of the 3rd China-Japan-US Symposium on Structural Health Monitoring and Control, Dalian, China, October 13-16, 2004.

4

L. Zhao, "Characterization of Deck-to-Girder Connections in FRP Composite Superstructures", Ph.D. Dissertation, University of California, San Diego, La Jolla, CA, USA, 1999.

5

M. Fraser, A. Elgamal, J.P. Conte, V. Karbhari, L. Zhao, and F. Seible, "Pilot Health Monitoring for Composite Bridge-deck Panels", Proceedings of the Third Middle East Symposium on Structural Composites for Infrastructure Applications, Aswan, Egypt, December 17-20, 2002.

6

I.T. Jolliffe, Principal Components Analysis, Spring-Verlag, New York, 1986.

7

C.M. Bishop, Neural Networks for Pattern Recognition, Oxford University Press, 1995.

8

S. Thrun, and K. Moller, "Active Exploration in Dynamic Environments", in Advances in Neural Information Processing System 4, Moody, J., Hanson, S., and Lippmann, R. (Eds.), Morgan Kaufman, 1992.

9

D.A. Cohn, Z. Ghahramani, and M.I. Jordan, "Active Learning with Statistical Models", Journal of Artificial Intelligence Research, 4, 129-145, 1996.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)