Artificial intelligence in bridge inspection

By Peter Murphy

Published May 10, 2019

Research assistant professor Xiao Liang shared his work on several different bridge topics at two separate events last month.

Print

Liang featured his paper Estimation of Rail Vertical Profile Using an H-Infinity Based Optimization with Learning at the 2019 American Society of Mechanical Engineers (ASME) Joint Rail Conference (JRC). Liang and co-author Minghui Zheng, a mechanical and aerospace engineering assistant professor, propose a learning-based detection algorithm for rail track geometry monitoring.

“This paper proposes an algorithm to estimate the rail vertical profile using the vertical acceleration of the vehicle resulting from train-track dynamic interaction,” Liang says, “the proposed algorithm possess several advantages, including easy design, little tuning effort and low computational cost.”

ASME’s JRC featured international researchers across engineering disciplines. Specialists with civil, mechanical, electrical and systems engineering backgrounds, and experts in rail safety, planning, design, financing, operations and management were also featured at the conference.

Liang’s other engagement was with Michael Baker International, an engineering and consulting firm with offices throughout the United States. The firm has a webinar series, and Liang was selected to present his research, Artificial Intelligence in Vibration-Based Structural Health Monitoring and Bridge Inspection.

“The inevitable aging of bridge structures, as one of the most critical components in transportation infrastructure systems, makes them vulnerable to future extreme events,” Liang says, “considering limited resources for various post-disaster activities, this work presents a three-level image-based approach for bridge inspection using deep learning to interpret images from unmanned aerial vehicles.”

Read Xiao's paper at this link