Asset management is necessary to keep transportation moving. Generally, it’s a process of maintaining, upgrading, and expanding physical assets. And we are on the verge of using technology to make it even more exciting. With limited staff and budgets, advanced technology appeals to managers in airports, sea ports, rail systems, transit agencies, and state departments of transportation alike.
University of Washington Professor of Civil and Environmental Engineering Yinhai Wang is the incoming chair of TRB’s
Standing Technical Committee on Artificial Intelligence and Advanced Computing Applications. He sees the next big development of artificial intelligence (AI) in transportation asset management as a way to address safety.
“Volume and traffic demand forecasting are helpful for operations, but there are more important applications of AI in transportation asset management, such as using it to predict which bridge is getting close to the end of its useful lifespan in time to prepare and budget. I believe there is a great potential to use AI for infrastructure risk and deterioration assessments,” he says.
TRB recently offered a webinar that prepared participants to do just that. Presenters offered a closer look at how AI can be applied to highway data to better explain and predict bridge system performance. The session also noted the potential of AI and deep learning to improve sensor signal data, and explained how these technologies can support design, operations, and management of highway systems.
“Data management skills are particularly needed in the transportation industry. Very often, traffic data are collected and stored by separate data management systems without sufficient quality control. To make existing datasets ready for AI-based applications, transportation professionals need to use data management tools and skills to ensure data quality, properly integrate available datasets, and label and mark them as needed,” reiterates Wang.
In talking about his work with the
PacTrans Workforce Development Institute he emphasizes the shortage of AI engineers in transportation and compares their role to that of a restaurant chef.
“You can have all the materials at hand, but the chef is the most critical part to turn them into delicious meals for a restaurant. AI engineers are critical to making valuable and innovative uses of transportation data. Given the lack of AI engineers, state departments of transportation may need to find a way to collaborate with universities and IT companies to get ready for the coming smart infrastructure age.”
Those considering a career options, including those in AI, may be inspired to find their calling by watching
Your Career in Transportation.
Understand and plan for the shifting and rapidly increasing role of technology
As new tools, technologies, materials, and approaches proliferate in the transportation system, the ways highway agencies maintain, preserve, and renew infrastructure is expected to shift. The TRB National Cooperative Highway Research Program's (NCHRP)
Strategic Issues Facing Transportation, Volume 7: Preservation, Maintenance, and Renewal of Highway Infrastructure places emphasis on preparing for plausible future scenarios. The authors develop a pathway to guide transportation agencies in advancing the implementation of emerging practices through a process involving awareness, advocacy, assessment, adoption, and action planning.
AI offers high potential for practical value. NCHRP is accepting proposals through March 25 to
aid state departments of transportation in transitioning to a more advanced state of practice with machine learning by demonstrating the feasibility and understanding skills, resources, costs, benefits, and limitations of the technology. To make the most of the technology, agencies also need to share frameworks, tools, and guidance when there are common use cases.
Combining two new technologies, a study published in
Transportation Research Record (TRR) presents a
drone-mountable, real-time AI framework for road asset classification. With a small target dataset, the authors leveraged transfer learning to fine-tune an existing dataset to their target. Overall, their results demonstrated 81.33% accuracy on the test set, showing that it is possible to use AI in this way.
Rail offers unique challenges that may make AI even more appealing. Modifying track structures is particularly disruptive and many factors can influence reference measurements. A TRB Rail Safety Innovations Deserving Exploratory Analysis (IDEA) report explores an approach for
stress state prediction of continuous welded rail using contactless acoustic sensing and machine learning methods. An impulse-based experimental device was designed and deployed to excite multiple vibrational modes in rail simultaneously. What’s more, this can be done without track structure modification, nor hazards to humans, and it eliminates the need for reference measurements and reduces influences.
TRB’s Transit Research Cooperative Program (TCRP) also notes the potential for AI on the rails.
Maintenance Planning for Rail Asset Management – Current Practices explores video assistance and suggests that AI could be implemented to assist in testing for detection of rolling contact fatigue and possible internal rail defects that develop.
Resources to help manage data
Data management is different from data analysis. Both require quite a bit of preliminary work and understanding.
In October 2020, TRB offered a webinar to help attendees dip their toes into the sea of
data governance and asset management. Presenter slides are available that cover a variety of tools, including one specifically for the maritime system.
TRB's Airport Cooperative Research Program (ACRP) has taken a look at Computerized Maintenance Management Systems (CMMS) for asset management across various airport systems. ACRP’s
Guidebook for Advanced Computerized Maintenance Management System Integration at Airports develops guidance on the steps necessary to implement a CMMS, factors for consideration in prioritizing which systems to include using a phased approach, and the steps for integrating CMMS data into performance management and business decision making.
An increasing number of systems depend on deep neural networks, from autonomous vehicles to social media platforms that influence political discourse. Scientists are also beginning to rely more on deep learning for AI as a knowledge discovery tool as research becomes ever more data driven. The recording of the National Academy of Sciences’ lecture on the
science of deep learning is available.
Scratching the surface with Edge AI
Wang is particularly excited to see how Edge AI will affect the future of transportation asset management and connected vehicle operations. Edge AI implements Machine Learning algorithms in edge devices to conduct quick data collection, fusion, analysis, and communication with roadway users. Benefits of Edge AI include quick response time, low bandwidth to communicate with transportation management centers, and easy deployments.
“We’ve developed an Edge AI technology at the University of Washington Smart Transportation Applications and Research Laboratory (STAR Lab). We are currently testing the technology in several locations. For example, sponsored by the Federal Highway Administration State Transportation Innovation Council program, we are working with the City of Bellevue and the City of Lynnwood in Washington State to use the technology for road surface condition and traffic sensing, data fusion, and data processing. A warning message will be sent to road users when warranted through connected vehicles or social media for traffic safety improvement. This technology has a lot of potential in rural areas where broadband communication infrastructure may not be available.”
Launch into the future of asset management with TRB
Join TRB this summer for the
National Conference on Transportation Asset Management. Open to all registrants, the conferences’ tracks are implementation, data governance and tools, managing risk, partners and peers, and sustaining asset management.
Use
MyTRB.org to become a friend of TRB’s
Standing Technical Committee on Artificial Intelligence and Advanced Computing Applications and connect with experts like Wang. Friends of committees receive updates on and can volunteer to participate in committee activities.
Get involved in this work with the Cooperative Research Programs. Look for
ongoing information on new projects, requests for proposals, or to nominate yourself or others to serve on a project panel. Submit problem statement research ideas and find new announcements in
TRB’s weekly newsletter or on the homepages for the
ACRP,
NCHRP, and the
TCRP.
TRB reports cited in this article:
TRB events:
IDEA reports:
Cooperative Research Program active projects:
TRB Standing Committees:
Articles published in TRR:
Additional TRB resources:
Additional National Academies of Sciences, Engineering, and Medicine resources:
Contact:
Beth Ewoldsen, Content Strategist
Transportation Research Board
202-334-2353;
bewoldsen@nas.edu
Published March 1, 2021
This Summary Last Modified On: 3/1/2021