2026 Undergraduate Research Showcase

AI-Enabled Pavement Management Using LTPP Database

Document Type

Student Presentation

Presentation Date

4-24-2026

Faculty Sponsor

Dr. Yang Lu

Abstract

Across the United States, the average driver lost nearly one work week in 2024 due to travel delays as congestion continues to increase along major highways. At the same time, state and local transportation agencies are responsible for maintaining aging roadway infrastructure under budget constraints and performance expectations. As pavements deteriorate, agencies must decide when and how to intervene, often balancing short-term repair needs with long-term performance. This research addresses the challenge of determining which roads require treatment and when maintenance should occur to support resilient, cost-effective infrastructure management. Specifically, this study investigates whether past pavement performance data can be used to develop a Markov Decision Process (MDP) model that identifies optimal, long-term pavement maintenance strategies. An MDP, a mathematical decision-making model used to evaluate choices under variability, will predict how pavement conditions change from year to year and determine effective maintenance actions for each condition level. I will use data from the Long-Term Pavement Performance (LTPP) database that tracks performance data on national highways, and categorize pavement values into condition states such as good, fair, and poor. I will then use Microsoft Excel to estimate how pavements transition between these states and evaluate maintenance strategies based on expected long-term performance and cost. Although data collection is still in progress, this research aims to demonstrate how performance-based decision tools can support transportation agencies in shifting from short-term repairs to proactive planning, ultimately improving infrastructure resilience, optimizing maintenance, reducing lifecycle costs, and decreasing congestion nationwide.

This document is currently not available here.

Share

COinS