Abstract Lane closures, when required during road repair and maintenance, can cause traffic congestion in adjacent open lanes. It is problematic when oversized vehicles are present, as they can create safety risks for workers and other drivers in work zones. The existing technologies in this regard are customized only for overheight vehicle detection and ignore the horizontal span of the vehicles. Therefore, those solutions cannot be extended directly to address the problem at hand. Additionally, the existing methods rely on expensive sensors such as lidars and radars for autonomous vehicle detection. Exorbitant costs restrict the largescale use of those devices. As a more economical solution, this study will leverage inexpensive RGB-D sensors for accurate learning-based vehicle size estimation. To address this issue, this project aims to develop an intelligent early warning system that uses low-cost 3D sensing cameras and artificial intelligence (AI)-based detection algorithms. The system will estimate the size of approaching vehicles and issue a real-time warning to any vehicle that is too large for the open lanes. This will help prevent potential accidents and encourage these vehicles to take alternate routes or slow down to ensure everyone's safety.
Description Lane closures are often necessary for the repair and maintenance of roads. This strategy can cause traffic congestion in the adjacent open lanes. It is especially problematic when oversized vehicles are present, as they can create safety risks for workers and other drivers in work zones. The existing technologies in this regard are customized only for overheight vehicle detection and ignore the horizontal span of the vehicles. Therefore, those solutions cannot be extended directly to address the problem at hand. Additionally, the existing methods rely on expensive sensors such as lidars and radars for autonomous vehicle detection. Exorbitant costs restrict the large-scale use of those devices. As a more economical solution, this study will leverage inexpensive RGB-D sensors for accurate learning-based vehicle size estimation. To address this issue, this project aims to develop an intelligent early warning system that uses low-cost 3D sensing cameras and artificial intelligence (AI)-based detection algorithms. The system will estimate the size of approaching vehicles and issue a real-time warning to any vehicle that is too large for the open lanes. This will help prevent potential accidents and encourage these vehicles to take alternate routes or slow down to ensure everyone's safety.
Objective "This project will address the following 2022-2026 USDOT strategic goal:
(a) Safety - Make our transportation system safer for all people. Advance a future without transportation-related serious injuries and fatalities.
(b) Equity – The AI tool will be trained with non-discriminatory data."
Impacts/Benefits Instant notification of oversized vehicle detection to alarm the approaching vehicles as well as overground workers. Reduced fatalities and accident rate in work zones and improved safety in consequence. Reduced damage to infrastructure and vehicles. Increased confidence in traffic management system in and around an active work zone