Optimising Pavement Crack Sealing with a Depth-Informed Reinforcement Learning Framework and Synthetic Crack Trajectory Modelling

Key Points

• The study aims to enhance pavement crack sealing by employing a depth-informed reinforcement learning framework and synthetic crack trajectory modeling.
• Utilized a reinforcement learning framework to optimize crack sealing techniques.
• Developed synthetic models to simulate crack trajectories for better decision-making.
• Assessed the depth-informed approach for its impact on sealing effectiveness.
• Improved sealing effectiveness was observed with the depth-informed approach compared to traditional methods.
• Demonstrated a significant reduction in maintenance costs associated with pavement repairs over time.
• Increased longevity of sealant applications was reported, enhancing overall pavement performance.