Center for Safety Equity in Transportation

rural • isolated • tribal • indigenous

Enhancing Vehicle Sensing for Traffic Safety Performance Improvements Using Roadside LiDAR Data in RITI Communities

  • Active

    CSET Project #: 2203

    Project Funding: UHM

  • Start Date: March 2022

    End Date: July 2023

    Budget: $$60,239.00

Project Summary

The Federal Motor Carrier Safety Administration (FMCSA) developed its 2015-2018 strategic plans to identify four strategic focus areas including “Safety 1st” culture and comprehensive data utilization and leveraging technology. Compared to crashes occurring in urban areas, traffic crashes in Rural, Isolated, Tribal, or Indigenous (RITI) communities are associated with a series of significant attributes, such as high speed, low seatbelt usage rate, poor weather and pavement conditions, inferior lighting conditions, considerable distractions, etc. Recent technological advancements in computer vision algorithms and data acquisition devices have greatly facilitated the research to enhance traffic sensing for traffic safety performance improvements. Significant research efforts have been devoted to developing and deploying new technologies to better detect, sense, and monitor traffic dynamics and rapidly identify crashes in RITI communities. This project aims to develop a set of deep learning-based traffic object recognition from roadside Light Detection and Ranging (LiDAR) data to automatically detect vehicle dynamic characteristics, such as vehicle speed, headway, and spacing, and then track vehicle trajectories in real-time to better identify and predict traffic crashes to mitigate crash injuries with minimum response time. Currently, a majority of the existing methodologies applied deep learning-based techniques, especially Convolutional Neural Networks (CNNs), for traffic detection and tracking on autonomous driving datasets. Very fewer studies were focused on deep learning-based traffic detection using roadside LiDAR data, partially due to the lack of publicly available roadside LiDAR datasets for network training and testing. In this study, we plan to develop a novel framework based on CNNs and LiDAR data for automated vehicle detection. It leverages the domain knowledge of CNNs trained on large-scale autonomous driving datasets for vehicle detection from roadside LiDAR data. The project results will better use roadside LiDAR data to monitor traffic operation flow conditions and identify crashes dynamically and help transportation agencies improve traffic safety performance developing timely countermeasures in the States of Alaska, Washington, Idaho, and Hawaii.