AI for Transportation Planning
The Texas Department of Transportation has digital information about its projects spanning decades, but it is difficult for planners to access it all for use when planning budgets, timelines and potential issues for future projects. A University of Texas at Arlington researcher is creating a way to use artificial intelligence to help those planners easily use all available data instead.
Mohsen Shahandashti, a civil engineering professor at UTA, received a three-year, $226,172 grant from the Texas Department of Transportation to create a deep-learning algorithm that incorporates natural language processing and data mining to sift through millions of pieces of data, such as project characteristics, weather conditions, topography and geo-environmental conditions, that will allow TxDOT planners to apply the most possible data to their projects.
Current software is not intelligent enough, and there is no single tool that can collect and analyze all of the available data so that planners can efficiently analyze all possible factors. Experienced planners often draw upon their own knowledge, but humans cannot have complete knowledge of current and past variables. Shahandashti’s hope is that his tool will allow planners to look not just at a single aspect of a project, but all aspects of a project to keep on time and on budget by analyzing data quickly and easily for any project. Civil Engineering professors Stephen Mattingly, Kyeong Ryu, and Nazmus Sakib will help Shahandashti to achieve these goals.
“I am very proud of this project. It will be a leap in civil infrastructure decision-making and taking advantage of data in the fullest. It’s fundamentally changing how transportation planners look at data because they can use it to avoid problems before they arise. My group does a lot of work with data analytics and we develop tools to help government agencies make decisions,” Shahandashti said.