Data Acquisition and Processing Using Artificial Intelligence and Machine Learning

Artificial intelligence (AI) and machine learning (ML) offer state transportation agencies promising tools to help them advance their missions by informing decision making; improving information accuracy, completeness, and timeliness; and automating tedious tasks to free up valuable DOT resources. The objective of this research was to provide MoDOT with tools, information, and examples to implement and leverage these technologies. The project reviewed existing work efforts and explored opportunities where AI and ML could replace existing work activities or augment and improve them. This was completed by developing a universe of potential DOT applications, screening them based on key criteria, scoping five potential projects, and executing two pilot projects. Findings from the idea generation process and execution of the pilots can be applied to MoDOT’s future AI and ML endeavors.  

VIEW FINAL REPORT

Report number: cmr 24-009
Published: July 2024

Project number: TR202215

Authors: Ryan Loos, Mark Egge, and Anna Batista

Performing organization: High Street Consulting Group

Striping Program

The use of wider (greater than 4 in) edge line pavement markings is recognized as a proven safety countermeasure by the Federal Highway Administration (FHWA). However, providing wider edge lines can pose challenges due to budget and striping capacity constraints. One strategy that could be used to address the capacity constraints involves reducing the amount of striping performed on low volume roads. The objective of this research study is to synthesize existing research and DOT practices regarding the use of wider edge line markings on all roads and the use of pavement markings of any width on low volume roads to facilitate the evaluation of tradeoffs between different pavement marking strategies. The research methodology to meet these objectives includes a literature review (e.g., research studies and DOT guidance and standards) and department of transportation (DOT) interviews. Results from the literature review indicate crash reductions for wider pavement markings ranging from 7 percent to 30 percent (total crashes) and from 14 percent to 51 percent (fatal and injury crashes) and benefit-cost ratios ranging from 24:1 to 55:1. Regarding prevalence of wider pavement markings, most state DOTs use 6-in markings (or, in some cases, 5-in markings) to some extent. Previous research studies on the use of pavement markings on low volume or narrow roads have shown mixed results. The literature review identified 12 state DOTs with state-specific warrants for center line or edge line markings on low volume roads. Implementation challenges noted by DOTs during the interviews include communication, budget and capacity constraints, and equipment needs. Overall, the research findings indicate that increased use of wider pavement markings and reductions in the use of pavement markings on lower volume roads is a viable strategy that can be explored further for implementation.

VIEW FINAL REPORT

Report number: cmr 24-008
Published: July 2024

Project number: TR202402

Authors: Henry Brown, Praveen Edara, Zhu Qing, and Ingrid Potts

Performing organization: University of Missouri-Columbia

Developing a Hazard Detection and Alert System to Prevent Incidents

The proposed project aimed to develop a cost-effective, user-friendly, and reliable hazard detection and alert system to prevent crashes between construction equipment/vehicles and workers on foot in work zones. The envisioned system includes wearable proximity sensors for the workers and an in-vehicle portable detection system for the operators. The system was designed to notify operators and workers promptly when equipment/vehicles are backing up and workers are nearby. The system employed Bluetooth Low Energy (BLE) beacon technology that enables a bidirectional detection and alert system to prevent crashes between equipment/vehicles and workers. However, the developed system had limitations associated with environmental and natural factors. The project concludes that BLE beacon technology is unsuitable for developing a safety detection and alert system for work zones.

VIEW FINAL REPORT

Report number: cmr 24-007
Published: July 2024

Project number: TR202214

Authors: Sejun Song, Sungyop Kim, and John Kevern

Performing organization: University of Missouri-Kansas City