AI & Decision Support Systems for Crash Preventability PAR Processing
Project Goal
The goal of this project is to evaluate the economic, technical, and operational feasibility of applying artificial intelligence (AI) and/or other decision support systems (DSS) for Crash Preventability Police Accident Report (PAR) Processing.
Background
In accordance with a recent Executive Order from the White House, all U.S. Government research departments must include AI and similar technologies as part of their research portfolios. This project provides a unique opportunity to assess the feasibility of AI and DSS tools in support of the Agency's Crash Preventability program.
Summary
Under the Agency's the Crash Preventability Demonstration Program, motor carriers or drivers may submit Requests for Data Review (RDR) through FMCSA’s DataQs system for eight specific types of crashes that occurred on or after June 1, 2017. To date, the program has received over 12,000 RDRs. Only 60 percent of these are eligible crash types with approximately 94 percent being not preventable. To have a crash reviewed in the program, the submitter must provide a PAR. Each State’s PAR collects different information and uses a different coding system, narrative requirements, diagrams, and point of impact information. Thus, it takes a human reviewer significant time to review each PAR against code sheets. Additionally, each human reviewer must compare the PAR to the crash report in the Agency’s Motor Carrier Management Information System (MCMIS) to ensure there is a match. An AI and/or DSS could potentially streamline this process and significantly reduce staff processing hours.
Contractor
VTTI
Final Report
The brief and report have been published in the National Transportation Library:
- Technology Brief: https://rosap.ntl.bts.gov/view/dot/78878
- Full Report: https://rosap.ntl.bts.gov/view/dot/78877