Date of Award

Fall 2017

Project Type

Thesis

Program or Major

Civil Engineering

Degree Name

Master of Science

First Advisor

Erin Santini Bell

Second Advisor

Tat S Fu

Third Advisor

Ricardo A Medina

Abstract

Machine Learning and Computer Vision are the two technologies that have innovative applications in diverse fields, including engineering, medicines, agriculture, astronomy, sports, education etc. The idea of enabling machines to make human like decisions is not a recent one. It dates to the early 1900s when analogies were drawn out between neurons in a human brain and capability of a machine to function like humans. However, major advances in the specifics of this theory were not until 1950s when the first experiments were conducted to determine if machines can support artificial intelligence. As computation powers increased, in the form of parallel computing and GPU computing, the time required for training the algorithms decreased significantly. Machine Learning is now used in almost every day to day activities. This research demonstrates the use of machine learning and computer vision for smart infrastructure management. This research’s contribution includes two case studies – a) Occupancy detection using vibration sensors and machine learning and b) Traffic detection, tracking, classification and counting on Memorial Bridge in Portsmouth, NH using computer vision and machine learning. Each case study, includes controlled experiments with a verification data set. Both the studies yielded results that validated the approach of using machine learning and computer vision. Both case studies present a scenario where in machine learning is applied to a civil engineering challenge to create a more objective basis for decision-making. This work also includes a summary of the current state-of-the -practice of machine learning in Civil Engineering and the suggested steps to advance its application in civil engineering based on this research in order to use the technology more effectively.

Share

COinS