<rss version="2.0" xmlns:atom="https://www.w3.org/2005/Atom">
  <channel>
    <title>Research in Progress (RIP)</title>
    <link>https://rip.trb.org/</link>
    <atom:link href="https://rip.trb.org/Record/RSS?s=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" rel="self" type="application/rss+xml" />
    <description></description>
    <language>en-us</language>
    <copyright>Copyright © 2026. National Academy of Sciences. All rights reserved.</copyright>
    <docs>http://blogs.law.harvard.edu/tech/rss</docs>
    <managingEditor>tris-trb@nas.edu (Bill McLeod)</managingEditor>
    <webMaster>tris-trb@nas.edu (Bill McLeod)</webMaster>
    <image>
      <title>Research in Progress (RIP)</title>
      <url>https://rip.trb.org/Images/PageHeader-wTitle-RIP.jpg</url>
      <link>https://rip.trb.org/</link>
    </image>
    <item>
      <title>Enhancing Utah's Rest Area Safety Through AI-Based Near-Miss Detection and Risk Mapping</title>
      <link>https://rip.trb.org/View/2672765</link>
      <description><![CDATA[State-maintained rest areas are facing increasing safety risks as infrastructure ages and traveler demand grows, raising concerns about both user well-being and operational efficiency. Utilizing open-source data and video from surveillance cameras in rest areas managed by the Utah Department of Transportation (UDOT), this project proposes to incorporate a safety-focused analysis layer into rest area management in Utah. The research team will develop computer vision pipelines for multi-object tracking and scene calibration to extract vehicle and pedestrian precise trajectories and speeds, then calculate surrogate safety measures to identify near-miss interactions in parking areas and walkways. These metrics will be integrated into spatiotemporal hotspot maps and a site-level Safety Performance Index (SPI) that ranks locations and time periods with the highest safety risks. The SPI will guide a risk-based shortlist of targeted countermeasures, including signage enhancements, speed-calming features, bollards, and lighting upgrades, each accompanied by projected risk reduction and cost estimates to support informed decision-making.]]></description>
      <pubDate>Sun, 22 Feb 2026 10:38:34 GMT</pubDate>
      <guid>https://rip.trb.org/View/2672765</guid>
    </item>
    <item>
      <title>AI-Driven Telematics Solutions for Detecting Near-Crash Events and Safety Hotspots in Texas Transportation Networks</title>
      <link>https://rip.trb.org/View/2606398</link>
      <description><![CDATA[The research team will leverage telematics data to proactively identify near-crash events and hotspots, strengthening transportation safety management across Texas. The work will advance analytical methodologies in four areas: (1) trajectory-based analysis of vehicle movement patterns; (2) event-based analysis of critical driving behaviours such as hard braking and abrupt manoeuvres; (3) multi-criteria analysis integrating mobility, safety, and environmental performance; and (4) data fusion techniques that combine telematics with other sources, including traffic sensors and historical crash records. Building on this foundation, the research team will apply spatial-temporal analyses and machine learning predictive models to detect current and forecast future near-crash hotspots. An interactive artificial intelligence (AI)-powered decision-support system will be developed to provide transportation agencies with actionable insights for targeted safety interventions. Rural and urban case studies will demonstrate the platform’s applicability and validate its effectiveness through comparisons with historical crash data. An implementation roadmap will guide integration into agency safety management practices. The project will deliver robust analytical tools and evidence-based recommendations that can be seamlessly integrated with existing platforms—such as geographic information system (GIS) systems, roadway networks, and performance dashboards— ensuring compatibility and significantly enhancing proactive traffic safety measures statewide.]]></description>
      <pubDate>Thu, 02 Oct 2025 09:44:48 GMT</pubDate>
      <guid>https://rip.trb.org/View/2606398</guid>
    </item>
    <item>
      <title>Evaluation of Signs at Highway-Rail Crossings</title>
      <link>https://rip.trb.org/View/2601423</link>
      <description><![CDATA[This project evaluates the safety impacts of replacing STOP signs with YIELD signs at highway–rail grade crossings. The study examines whether this transition has resulted in measurable reductions in crashes and near misses, while also analyzing key crossing characteristics—including traffic volume, train frequency, visibility, and roadway type—that may influence safety outcomes. Findings will be used to provide data-driven recommendations on whether to retain, modify, or reverse the signage change at specific locations, supporting evidence-based decisions for improving transportation safety.]]></description>
      <pubDate>Wed, 17 Sep 2025 16:18:51 GMT</pubDate>
      <guid>https://rip.trb.org/View/2601423</guid>
    </item>
    <item>
      <title>Evaluation Of The Relationship Between Near Misses And Crash Outcomes At Intersections In Oregon</title>
      <link>https://rip.trb.org/View/2594039</link>
      <description><![CDATA[Over 44,000 crashes occurred on Oregon roads in 2022, 41% of which were at intersections, according to data from Oregon Department of Transportation's (ODOT’s) Crash Analysis and Reporting Unit data. Intersections are also discrete locations where design improvements can reduce crash rates while minimizing costs, when supported by relevant data. Crash analysis is a useful tool for characterizing the safety of a system, then refining the system’s design. However, assessing the safety of novel transportation designs or existing facilities using real crash data may not be feasible because of the long observational periods and low numbers of crashes, typically at least three years of before and after data, required to acquire statistically significant quantities of data. Significant advances have been made in the ability to characterize transportation system safety with tools such as the Highway Safety Manual (HSM) predictive methods, however these tools require that crash modification factors (CMFs) have been produced for the infrastructure treatments in question, and for those models to have been calibrated to local conditions.]]></description>
      <pubDate>Thu, 28 Aug 2025 18:18:57 GMT</pubDate>
      <guid>https://rip.trb.org/View/2594039</guid>
    </item>
    <item>
      <title>Proactive Assessment of Pedestrian and Bicycle Safety at Intersections Through Video Analysis of Right Turn on Red Maneuvers</title>
      <link>https://rip.trb.org/View/2401739</link>
      <description><![CDATA[Intersections are critical points in urban traffic systems where different modes of transportation converge, including vehicles, pedestrians, and cyclists. However, despite efforts to improve safety measures, intersections remain sites of significant risk, particularly for vulnerable road users such as pedestrians and cyclists. One significant contributing factor to this risk is the practice of right turn on red (RTOR) maneuvers, which can introduce conflicts between turning vehicles and pedestrians or cyclists crossing the intersection. RTOR maneuvers at intersections represent a convenience for motorists but often pose significant safety risks for pedestrians and cyclists. Therefore, there is a pressing need to develop innovative strategies that leverage advanced technology, particularly video data analysis, to better understand and mitigate the risks associated with RTOR maneuvers at intersections. Utilizing safety surrogate measures, such as post-encroachment time (PET), offers distinct advantages over relying solely on crash history data. These measures provide real-time insights into potential safety risks, offering a more comprehensive understanding of near-misses and risky behaviors, and enabling a proactive approach to intersection safety management that can effectively mitigate hazards and improve overall safety outcomes. By utilizing PET, extracted from continually collected video data available to the research team, this research also seeks to capture the severity of RTOR-related near misses, providing valuable insights for targeted safety interventions.]]></description>
      <pubDate>Mon, 08 Jul 2024 14:54:09 GMT</pubDate>
      <guid>https://rip.trb.org/View/2401739</guid>
    </item>
    <item>
      <title>Identifying Near-Misses and Reducing Conflict through D-FYA at Signalized Intersections</title>
      <link>https://rip.trb.org/View/2287568</link>
      <description><![CDATA[Near-misses (a.k.a. traffic conflicts in other literature) refer to vehicles and/or pedestrians that almost collide. Collisions are often avoided by exceptional maneuvers (e.g., hard braking). According to a previous study, about 2,000 near-misses lead to a real crash. Traditionally, we rely on traditional crash records to identify the causes for collisions and to provide mitigating measures. This approach, ironically, will work only after many collisions occur. It would be much better if we can capture and analyze the near-misses and prevent collisions. 
Identifying the near-misses requires efficient vehicle and pedestrian tracking technologies. Compared with video-based or radar-based detectors, LIDAR sensors have three major advantages in perceiving vehicles and pedestrians. First, the performance will not deteriorate in dark and foggy conditions because the LIDAR sensing is based on active laser beam firing and reflection whereas video-based detectors are rely on comparing the objects and their backgrounds, becoming difficult in the above conditions. Second, LIDAR performance will not deteriorate in perceiving slow and still objects such as pedestrians or still vehicles while the radar-based traffic detectors have been shown to perform poorly when detecting slow or still objects. Third, the “ceiling” of LIDAR sensing accuracy is higher than the existing traffic detectors. The LIDAR sensing technology is being applied in autonomous vehicles to quickly perceive surrounding objects, requiring high adaptiveness and accuracy. As such, when the LIDAR sensors are applied to highway applications, they are anticipated to be more effective for increasingly complicated traffic conditions. As such, the LIDAR sensing technologies are likely to become a permanent solution to improve traffic safety in Utah and elsewhere.
]]></description>
      <pubDate>Wed, 08 Nov 2023 15:56:24 GMT</pubDate>
      <guid>https://rip.trb.org/View/2287568</guid>
    </item>
    <item>
      <title>Real-time Safety Diagnosis System for Connected Vehicles with Parallel Computing Architecture (Project O6)</title>
      <link>https://rip.trb.org/View/2004707</link>
      <description><![CDATA[The ongoing STRIDE F4 project – Automatic Safety Diagnosis in Connected Vehicle Environment – is to construct a computational pipeline of a near-crash diagnosis system to identify near-crash events by processing the Basic Safety Messages (BSMs) generated in the Connected Vehicle (CV) environment on the individual level. The in-vehicle system identifies outliers by analyzing BSMs from nearby vehicles and comparing with each individual driver’s past normal driving pattern provided by the Traffic Management Center (or a cloud server). The speed of data processing and transmission at both the cloud system and the in-vehicle system can be quite demanding. For the near-crash warning signal to be generated promptly in real-time environment, parallel computing is indispensable. The parallel computing technology can be incorporated into both the cloud system and the in-vehicle system.
First, the amount of BSMs received by the cloud server from the CVs could be massive up to several hundreds of GBs/sec. The data collection, data updating and warning massage broadcasting at the cloud server and the in-vehicle system can be carried out in a parallel fashion by using parallel computing. Second, vehicles are equipped with small computers to analyze the BSMs from all nearby vehicles. The in-vehicle data processing can also be accelerated by parallel computing.
The research team proposes to continue their current research using the parallel computing technology to accelerate the data processing and analysis in both the cloud system and the in-vehicle system. The group has extensive experience in parallel computing in solving large-scale fluid flow problems using the Message Passing Interface (MPI) library and the OpenCL technology. These technologies make the most out of today’s heterogeneous computing systems equipped with multi-core CPUs and GPUs. The team would like to leverage their existing parallel computing practice and adapt it to the traffic safety message processing and analysis.]]></description>
      <pubDate>Wed, 10 Aug 2022 15:07:53 GMT</pubDate>
      <guid>https://rip.trb.org/View/2004707</guid>
    </item>
    <item>
      <title>Rapid Safety Assessment Tool for Non-Conventional Roadway Designs and Emerging Technologies: Innovative Artificial Intelligence Application </title>
      <link>https://rip.trb.org/View/1877735</link>
      <description><![CDATA[Traditional safety assessment methodologies are profoundly dependent on reactive crash data. The Highway Safety Manual (HSM) approach recommends 3-5 worth of crash data before and after the implementation of safety countermeasures. Waiting for such a long period of time might not be feasible to address safety issues at various roadway facilities. Furthermore, with the advent of emerging transportation technologies, rapid safety assessment tools will be required. This research provided a proof-of-concept for the development of a proactive road safety assessment framework which could be utilized for a rapid evaluation of problematic intersections, non-conventional designs, newly adopted countermeasures, as well as emerging transportation technologies. The framework is based on leveraging advanced Artificial Intelligence (AI) and machine vision to identify Surrogate Measures of Safety (SMoS) in near real-time from video cameras installed at intersections. The study established a relationship between the types of crashes and their contributing factors utilizing SMoS. Video analytics used machine vision and object detection algorithms to identify motion paths and trajectories for different road users. From estimated users’ trajectories for vehicles and pedestrians, near crashes, known as traffic conflicts, were extracted by identifying critical thresholds for the SMoS such as Time-To-Collision (TTC), Post-encroachment Time (PET), and Deceleration Rate to Avoid a Crash (DRAC). Based on the SMoS identified, and the dominating conflict patterns, safety countermeasures could be recommended. The developed methodology of this study is a first step to cost-effectively assist transportation agencies evaluating hazardous locations and safety countermeasures without the need to wait for traditional crash data.
]]></description>
      <pubDate>Thu, 09 Sep 2021 17:14:08 GMT</pubDate>
      <guid>https://rip.trb.org/View/1877735</guid>
    </item>
    <item>
      <title>Near Miss Information Visualization Application for BIM</title>
      <link>https://rip.trb.org/View/1360643</link>
      <description><![CDATA[The primary objective of this research is to create a tool functional within Building Information Modeling (BIM) to visualize and analyze near miss information.  Sub objectives of this research will include creating algorithms and user-interfaces for safety managers to navigate and analyze reported near miss data.  Based on the findings, the research team will also identify and assemble a set of best practices for visualizing safety information for construction projects.  An implementation guide for integrating safety information into BIM will also be developed based on research findings. The proposed research contributes to improving the capabilities of safety information visualization for decisions made by management personnel.  By enabling construction safety managers to input, review and analyze safety information (specifically near misses) into an existing project BIM, hazardous situations and set of conditions can be identified and mitigated before an injury, illness or fatality occurs.]]></description>
      <pubDate>Sat, 11 Jul 2015 01:00:27 GMT</pubDate>
      <guid>https://rip.trb.org/View/1360643</guid>
    </item>
  </channel>
</rss>