<rss version="2.0" xmlns:atom="https://www.w3.org/2005/Atom">
  <channel>
    <title>Transport Research International Documentation (TRID)</title>
    <link>https://trid.trb.org/</link>
    <atom:link href="https://trid.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>Transport Research International Documentation (TRID)</title>
      <url>https://trid.trb.org/Images/PageHeader-wTitle.jpg</url>
      <link>https://trid.trb.org/</link>
    </image>
    <item>
      <title>When does visual distraction become dangerous in car-following? Evidence from naturalistic driving study data with causal inference on time-to-collision and braking intensity</title>
      <link>https://trid.trb.org/View/2659609</link>
      <description><![CDATA[Visual distraction is a major contributor to crash risk, particularly in car-following situations that demand continuous monitoring and rapid response. Although prior research using simulators and Naturalistic Driving Study (NDS) data has advanced the understanding, evidence remains limited on how visual distraction increases risk in real-world contexts and under which conditions it is amplified. Visual distraction is not an isolated factor, but a context-dependent phenomenon shaped by roadway conditions, traffic dynamics, and external stimuli. Beyond measuring its overall effect, it is essential to identify the circumstances in which visual distraction becomes especially hazardous. To address this gap, this study applies causal inference methods to NDS data. A Causal Forest was used to estimate the causal effect of visual distraction on two safety indicators: time-to-collision (TTC) and braking intensity. Subsequently, mediation analysis using Double Machine Learning (DML) was applied to disentangle the extent to which visual distraction mediates driving risk from the portion attributable directly to roadway and traffic conditions, thereby clarifying the indirect behavioral pathways versus structural design effects. Results show that visual distraction significantly reduces TTC, indicating heightened conflict seriousness, whereas its effect on braking intensity was not statistically significant. Mediation analysis further revealed that the effect of visual distraction on TTC varied across contexts, with stronger effects under high traffic density, ADAS-equipped vehicles, wider sidewalks, and fewer lanes. These findings underscore the importance of integrated safety strategies that mitigate visual distraction while also accounting for roadway design, traffic environment, and vehicle technologies in shaping driver behavior and risk.]]></description>
      <pubDate>Wed, 25 Feb 2026 13:58:58 GMT</pubDate>
      <guid>https://trid.trb.org/View/2659609</guid>
    </item>
    <item>
      <title>A Comparative Analysis of Lane-Change Data Extraction Methods Based on Naturalistic Driving Data Sets</title>
      <link>https://trid.trb.org/View/2613116</link>
      <description><![CDATA[Ensuring the extraction of complete and accurate lane-change trajectory data is crucial for the safety of autonomous driving systems. This paper explores methods for extracting lane-change data from natural driving data. In regard to the selection and preprocessing of data, wavelet denoising was chosen to process the NGSIM data set. With regard to the extraction of lane-change data, this paper puts forth a methodology for the identification of complete single-discretionary lane-change data. Then, the extraction methods are classified into two categories: fixed threshold-based and clustering-based. An analysis is then conducted to ascertain the relative merits and limitations of each method. The findings indicate that the current methods have shortcomings in identifying the start and end points of lane changes, particularly in the context of unconventional lane-change trajectories. In future research, different methods should be used to separately identify and extract start and end points based on varying trajectory shapes or characteristics.]]></description>
      <pubDate>Fri, 20 Feb 2026 15:28:22 GMT</pubDate>
      <guid>https://trid.trb.org/View/2613116</guid>
    </item>
    <item>
      <title>Personality traits of drivers who exhibit risky driving behavior: Analysis using a real vehicle test in SCL-90</title>
      <link>https://trid.trb.org/View/2632903</link>
      <description><![CDATA[The personality traits of drivers are related to their driving safety levels. Existing research on personality traits and drivers’ safety levels largely lacks observation of driver behavior and relies on self-defined concepts, such as aggressive or cautious categories, rather than personality traits recognized in the field of psychology. The relationship between personality traits and accident-related dangerous driving behaviors is unclear. Hence, there are no diagnostic personality psychology methods to evaluate drivers’ safety attributes. We conducted a 51-group naturalistic driving trial on a mountainous highway to investigate the influence of underlying personality traits on risky driving behavior in a nonprofessional driving population. We employed the Symptom Checklist-90 to measure nine underlying personality traits characterizing drivers’ psychological well-being, quantified real-time safety levels using three risky driving behaviors, and conducted data envelope analysis. Thus, we identified associations between underlying personality traits and degree of riskiness and determined typical risky driving behaviors related to underlying personality traits that significantly affected safety. Obsessive-compulsive symptoms and paranoia were the most reliable indicators of drivers engaging in risky driving behaviors—the former having increased speeding risks, the latter at greater risk of exhibiting road rage. Professionals in the field of psychology can identify these indicators to diagnose potential personality traits that seriously affect driving safety. These assessments are applicable to driver psychological training and management, and early intervention programs for traffic accident-prone drivers to reduce traffic accidents.]]></description>
      <pubDate>Fri, 09 Jan 2026 16:58:53 GMT</pubDate>
      <guid>https://trid.trb.org/View/2632903</guid>
    </item>
    <item>
      <title>Traj-LLM: A New Exploration for Empowering Trajectory Prediction With Pre-Trained Large Language Models</title>
      <link>https://trid.trb.org/View/2591853</link>
      <description><![CDATA[Predicting the future trajectories of dynamic traffic actors is a cornerstone task in autonomous driving. Though existing notable efforts have resulted in impressive performance improvements, a gap persists in scene cognitive and understanding of complex traffic semantics. This paper proposes Traj-LLM, the first to investigate the potential of using pre-trained Large Language Models (LLMs) without explicit prompt engineering to generate future motions from vehicular past trajectories and traffic scene semantics. Traj-LLM starts with sparse context joint encoding to dissect the agent and scene features into a form that LLMs understand. On this basis, we creatively explore LLMs' strong understanding capability to capture a spectrum of high-level scene knowledge and interactive information. To emulate the human-like lane focus cognitive function and enhance Traj-LLM's scene comprehension, we introduce lane-aware probabilistic learning powered by the Mamba module. Finally, a multi-modal Laplace decoder is designed to achieve scene-compliant predictions. Extensive experiments manifest that Traj-LLM, fueled by prior knowledge and understanding prowess of LLMs, together with lane-aware probability learning, transcends the state-of-the-art methods across most evaluation metrics. Moreover, the few-shot analysis serves to substantiate Traj-LLM's performance, as even with merely 50% of the dataset, it surpasses the majority of benchmarks relying on complete data utilization. This study explores endowing the trajectory prediction task with advanced capabilities inherent in LLMs, furnishing a more universal and adaptable solution for forecasting agent movements in a new way.]]></description>
      <pubDate>Wed, 05 Nov 2025 10:02:34 GMT</pubDate>
      <guid>https://trid.trb.org/View/2591853</guid>
    </item>
    <item>
      <title>The influence of built environment on speeding behavior – a naturalistic approach</title>
      <link>https://trid.trb.org/View/2604270</link>
      <description><![CDATA[The built environment elements in urban areas can have a significant impact on the performance of transport systems, including road safety. The primary objective of this research is to investigate the influence of the built environment on speeding behavior, as an indicator of road safety performance, using the city of Curitiba, Brazil, as the study’s setting. The built environment comprises physical features within the city, such as development patterns and roadway designs, and can be categorized into six groups: density, diversity, design, destination accessibility, distance to transit, and demographics. The Geographically Weighted Regression (GWR) statistical model was employed to explore the correlation between built environment variables and the occurrence of speeding in a spatially nonstationary scenario. Additionally, Moran’s I and Local Moran statistical methods were applied to investigate the spatial autocorrelation of speeding within the city. Data on speeding and location were collected through the application of a Naturalistic Driving Study. The measure of speeding was based on free-flow situations, considering the opportunity in which drivers could speed. In this study, the database included 1002 trips, 381.45 h of driving, and 9,443.83 km of travel within Curitiba and its metropolitan area from 2019 to 2021. The GWR model was applied using Curitiba’s traffic analysis zones (TAZs) as the zonal level. GWR reduced residual spatial autocorrelation relative to the global linear model; however, the global model achieved a lower AICc. Only the variable “proportion of arterial roads” showed a statistically significant correlation with speeding at a 95 % confidence level, with an inverse correlation observed across 100 % of the TAZs. Furthermore, it was observed that speeding behavior in Curitiba exhibits spatial autocorrelation, justifying the use of the GWR model. Low-Low and High-High spatial clusters were identified, with statistically significant differences observed between them, including average income, density of commercial and service units, density of intersections, density of speed cameras, and traffic signal density. The characteristics of arterial roads in Curitiba, including infrastructure and traffic control, may be influencing the reduction of speeding behavior.]]></description>
      <pubDate>Wed, 15 Oct 2025 12:30:50 GMT</pubDate>
      <guid>https://trid.trb.org/View/2604270</guid>
    </item>
    <item>
      <title>Using Video Analytics to Automatically Annotate Driver Behavior and Context in Naturalistic Driving Data</title>
      <link>https://trid.trb.org/View/2499264</link>
      <description><![CDATA[Naturalistic driving data provides a wealth of information for researchers studying driver behavior and distracted driving. However, manually annotating the videos to extract the data is costly and time consuming. This project’s research team set out to develop a system to analyze videos from the second Strategic Highway Research Program Naturalistic Driving Study dataset and automatically produce annotations and descriptors for events, behavior, and driving scenarios related to transportation safety.(1) The project had four objectives: characterize high-level driver behavior, such as eating or using a cellphone; classify the environment outside the vehicle, such as the position of roadway objects, work zones, and intersections; understand interactions and dependencies between drivers and the surrounding environment, such as looking at a billboard or a passing vehicle; and demonstrate how the video analytics techniques used in this study can help human factors researchers address research questions in novel ways. The researchers developed and tested advanced computer vision algorithms, including deep neural network (DNN)-based methods to capture spatial and temporal information embedded in the naturalistic driving videos. The DNN models included convolutional neural network models for image recognition and transformer-based models to process sequential data. All the codes developed as part of this project are open sourced.]]></description>
      <pubDate>Mon, 31 Mar 2025 08:54:35 GMT</pubDate>
      <guid>https://trid.trb.org/View/2499264</guid>
    </item>
    <item>
      <title>A unified probabilistic approach to traffic conflict detection</title>
      <link>https://trid.trb.org/View/2485303</link>
      <description><![CDATA[Traffic conflict detection is essential for proactive road safety by identifying potential collisions before they occur. Existing methods rely on surrogate safety measures tailored to specific interactions (e.g., car-following, side-swiping, or path-crossing) and require varying thresholds in different traffic conditions. This variation leads to inconsistencies and limited adaptability of conflict detection in evolving traffic environments, particularly as the integration of autonomous driving systems adds complexity. Consequently, there is an increasing need for consistent detection of traffic conflicts across interaction contexts. To address this need, the authors propose a unified probabilistic approach in this study. The proposed approach establishes a unified framework of traffic conflict detection, where traffic conflicts are formulated as context-dependent extreme events of road user interactions. The detection of conflicts is then decomposed into a series of statistical learning tasks: representing interaction contexts, inferring proximity distributions, and assessing extreme collision risk. The unified formulation accommodates diverse hypotheses of traffic conflicts and the learning tasks enable data-driven analysis of factors such as motion states of road users, environment conditions, and participant characteristics. Jointly, this approach supports consistent and comprehensive evaluation of the collision risk emerging in road user interactions. The authors demonstrate the proposed approach by experiments using real-world trajectory data. A unified metric for indicating conflicts is first trained with lane-change interactions on German highways, and then compared with existing metrics using near-crash events from the U.S. 100-Car Naturalistic Driving Study. The authors' results show that the unified metric provides effective collision warnings, generalises across distinct datasets and traffic environments, covers a broad range of conflict types, and captures a long-tailed distribution of conflict intensity. In summary, this study provides an explainable and generalisable approach that enables traffic conflict detection across varying interaction contexts. The findings highlight its potential to enhance the safety assessment of traffic infrastructures and policies, improve collision warning systems for autonomous driving, and deepen the understanding of road user behaviour in safety–critical interactions.]]></description>
      <pubDate>Tue, 28 Jan 2025 14:05:58 GMT</pubDate>
      <guid>https://trid.trb.org/View/2485303</guid>
    </item>
    <item>
      <title>Naturalistic driving study data applied to road infrastructure: A systematic review</title>
      <link>https://trid.trb.org/View/2480467</link>
      <description><![CDATA[Naturalistic driving studies (NDS) have great potential to characterize the road infrastructure factors influencing everyday driving. A systematic review was undertaken to evaluate the objectives, data processing, and analyses in best-practice applications of NDS data to road infrastructure. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines, a systematic search of seven databases was conducted on 27 June 2023 (PROSPERO CRD42023434948). Fifty-three English-language, peer-reviewed studies were analyzed on the basis of the primary infrastructure category reflected in the research aims. Studies described curves (14), turns at intersections (8), intersections (6), multi-modal treatments (6), ramps (4), work zones (4), charging (2), and other factors (9). Each study was assessed for the risk of methodological bias using amended National Heart, Lung, and Blood Institute templates for Quality Assurance. 74% of studies were assessed to be of ’Good’ quality, 13% of ‘Fair’ quality, and 13% of ‘Poor’ quality. Road infrastructure was characterized by external video (38%) complemented by non-NDS sources including satellite imagery (21%) and government data (19%). Data preparation was required in 91% of studies to extract meaningful variables (e.g. manual video coding) and/or link multiple datasets. Analysis predominantly determined correlations between aspects of driver behavior (speed, trajectory, etc.) and infrastructure factors (geometry, lane configuration, etc.). The methods employed were broadly applicable, but required considerable subject-specific adaptation for non-NDS datasets and/or time-consuming video coding. The incorporation of road infrastructure factors in NDS research can continue to be improved by reducing the computational cost of sample processing. Encouraged by the adaptability of the identified methods, NDS research has the potential to benefit from the consideration of road infrastructure factors in a Safe System context. The analytical requirements for all components of the Safe System should be considered when planning future NDS data collections and/or analysis.]]></description>
      <pubDate>Mon, 27 Jan 2025 15:11:03 GMT</pubDate>
      <guid>https://trid.trb.org/View/2480467</guid>
    </item>
    <item>
      <title>Context-Aware Attention Encoder-Decoder Network for Connected Heavy-Duty Vehicle Aggressive Driving Identification Under Naturalistic Driving Conditions</title>
      <link>https://trid.trb.org/View/2414162</link>
      <description><![CDATA[Driving behavior analysis and identification are of great significance for improving traffic safety and reducing fuel consumption. Existing methods primarily focused on the driving behavior of light-duty vehicles based on analysis methods using simulation or questionnaire data, while heavy-duty vehicles, which bear significant responsibility for fatal accidents, are seldom investigated. This study develops a context-aware attention encoder-decoder deep framework for aggressive driving identification, utilizing real massive multi-source heterogeneous data collected under naturalistic driving conditions. The proposed framework incorporates vehicle-related, driving-related, weather-related and environment-related data. Through the BiLSTM based encoder-decoder deep architecture, high-level representations of driving behavior are learned from driving signals layer by layer, and temporal dependencies are captured from both forward and backward directions. By learning context-aware personalized latent semantic vectors at different time step, the model is capable of adaptively focusing on the important information for prediction. To the authors' knowledge, the aggressive driving of heavy-duty vehicles under connected and naturalistic driving conditions has been rarely explored. This study contributes to the current understanding in this field. The proposed framework is evaluated based on the multi-source heterogeneous driving behavior data generated from over 13000 vehicles in a connected environment under naturalistic driving conditions. Empirical results from extensive experiments validate that the proposed model outperforms competing models, providing a promising approach with high effectiveness and robustness for aggressive driving behavior identification.]]></description>
      <pubDate>Tue, 07 Jan 2025 11:17:50 GMT</pubDate>
      <guid>https://trid.trb.org/View/2414162</guid>
    </item>
    <item>
      <title>Leader–follower identification with vehicle-following calibration for non-lane-based traffic</title>
      <link>https://trid.trb.org/View/2471091</link>
      <description><![CDATA[Most car-following models were originally developed for lane-based traffic. Over the past two decades, efforts have been made to calibrate car-following models for non-lane-based traffic. However, traffic conditions with varying vehicle dimensions, intermittent following, and multiple leaders often occur, making subjective Leader–Follower (LF) pair identification challenging. In this study, the authors analyze Vehicle Following (VF) behavior in traffic with a lack of lane discipline using high-resolution microscopic trajectory data collected in Chennai, India. The paper’s main contributions are threefold. Firstly, three criteria are used to identify LF pairs from the driver’s perspective, taking into account the intermittent following, lack of lane discipline due to consideration of lateral separation, and the presence of in-between vehicles. Secondly, the psycho-physical concept of the regime in the Wiedemann-99 model is leveraged to determine the traffic-dependent “influence zone” for LF identification. Thirdly, a joint and consistent framework is proposed for identifying LF pairs and estimating VF parameters. The proposed methodology outperforms other heuristic-based LF identification methods from the literature in terms of quantitative and qualitative performance measures. The proposed approach can enable robust and more realistic LF identification and VF parameter calibration with practical applications such as level of service (LOS) analysis, capacity, and travel time estimation.]]></description>
      <pubDate>Mon, 16 Dec 2024 11:59:20 GMT</pubDate>
      <guid>https://trid.trb.org/View/2471091</guid>
    </item>
    <item>
      <title>A multimodal data-driven approach for driving risk assessment</title>
      <link>https://trid.trb.org/View/2404323</link>
      <description><![CDATA[Real-time assessment and short-term warning of driving risks are critical for AI-assisted vehicles to significantly improve the safety and reliability of mobility. However, existing methods do not comprehensively consider these factors, making it difficult to achieve more accurate risk assessments. Aiming at this problem, this paper proposes a new driving risk assessment framework by integrating multimodal data. First, based on naturalistic driving experiments, the authors collected multimodal data encompassing human-vehicle–road factors. Then, using the Latent Dirichlet Allocation (LDA) model, they identified three risk levels based on driving behavior features: normal driving, longitudinal risky driving, and lateral risky driving. To better understand the spatiotemporal importance of multiple factors, a spatiotemporal dual-channel neural network based on a multi-layer attention mechanism (MLA-DCNN) is developed. This model has a spatiotemporal dual-channel structure, which can integrate “low-level” historical sequences and “high-level” extract statistical features of multiple features. In addition, it adopts three layers of attention mechanism, respectively used to capture the differences of features in temporal, spatial, and extracted-level dimensions. Results reveal that the LDA model is more effective than traditional clustering methods in uncovering latent patterns of driving risk. The proposed model achieved an impressive accuracy of 91.04%, demonstrating higher risk assessment capabilities than the other alternative models. In addition, the multilayer attention enhances the interpretability of the model and is able to capture the spatiotemporal importance of different factors across various road environments. This method can be applied to connected and automated vehicles (CAVs) using multimodal natural driving data collected by in-vehicle sensors. It enhances the risk warning capabilities of driving assistance systems, and the multidimensional importance analysis also supports decision-making for traffic management authorities.]]></description>
      <pubDate>Tue, 30 Jul 2024 16:26:59 GMT</pubDate>
      <guid>https://trid.trb.org/View/2404323</guid>
    </item>
    <item>
      <title>DERNet: Driver Emotion Recognition Using Onboard Camera</title>
      <link>https://trid.trb.org/View/2389107</link>
      <description><![CDATA[Driver emotion is considered an essential factor associated with driving behaviors and thus influences traffic safety. Dynamically and accurately recognizing the emotions of drivers plays an important role in road safety, especially for professional drivers, e.g., the drivers of passenger service vehicles. However, there is a lack of a benchmark to quantitatively evaluate the performance of driver emotion recognition performance, especially for various application situations. In this article, the authors propose an emotion recognition benchmark based on the driver emotion facial expression (DEFE) dataset, which consists of two splits: training and testing on the same set (split 1) and different sets (split 2) of drivers. These two splits correspond to various application scenarios and have diverse challenges. For the former, a driver emotion recognition network is proposed to provide a competitive baseline for the benchmark. For the latter, a novel driver representation difference minimization loss is proposed to enhance the learning of common representations for emotion recognition over different drivers. Moreover, the minimum required information for achieving a satisfactory performance is also explored on split 2. Comprehensive experiments on the DEFE dataset clearly demonstrate the superiority of the proposed methods compared to other state-of-the-art methods. An example application of applying the proposed methods and a voting mechanism to real-world data collected in a naturalistic environment reveals the strong practicality and readiness of the proposed methods. The codes and dataset splits are publicly available at https://github.com/wdy806/CDERNet/.]]></description>
      <pubDate>Thu, 27 Jun 2024 14:03:57 GMT</pubDate>
      <guid>https://trid.trb.org/View/2389107</guid>
    </item>
    <item>
      <title>Connected and Automated Vehicle (CAV) Testing Scenario Design and Implementation using Naturalistic Driving Data and Augmented Reality</title>
      <link>https://trid.trb.org/View/2343288</link>
      <description><![CDATA[Testing and evaluation is a critical step in the development and deployment of connected and automated vehicle (CAV) technology. Testing standards for human-driven vehicles, such as Federal Motor Vehicle Safety Standards (FMVSS), were established a long time ago. However, current standards cannot be applied to CAVs, because they often assume the presence of a human driver, who conducts the driving tasks. It is very important to develop test procedures and identify applicable test scenarios (user cases) for CAVS to evaluate the “intelligence” of the vehicle. The intelligence level indicates whether a CAV can drive safely and efficiently without human intervention. The newly released Automated Driving Systems Guideline 2 has made it very clear that the new automated driving systems need validation methods and to be tested by incorporating behavior competencies. In this research, a unified framework is designed to solve the entire test scenario library generation (TSLG) problem, where a novel method is proposed for the library generation question. Theoretical analysis provides justifications of the proposed method regarding both evaluation accuracy and efficiency. Specifically, the proposed method obtains unbiased index estimation of performance metrics (i.e., accuracy) with a fewer number of required tests (i.e., efficiency). The three case studies verify the proposed methodology and the results show that the evaluation process can be accelerated by 10³ times compared with the naturalistic driving data (NDD) evaluation method, with the same accuracy.]]></description>
      <pubDate>Mon, 26 Feb 2024 08:51:30 GMT</pubDate>
      <guid>https://trid.trb.org/View/2343288</guid>
    </item>
    <item>
      <title>A Review of Driving Style Recognition Methods From Short-Term and Long-Term Perspectives</title>
      <link>https://trid.trb.org/View/2321622</link>
      <description><![CDATA[Driving style recognition provides an effective way to understand human driving behaviors and thereby plays an important role in the automotive sector. However, most works fail to consider the influence of deploying the recognition results on the vehicle side, which requires real-time recognition performance. To facilitate the application of driving styles in automotive, the authors survey related advances in driving style recognition along short- and long-term pipelines. The authors first defined short- and long-term driving styles and then described the input data used by the recognition models and related data-processing techniques. Furthermore, the authors also revisited existing evaluation metrics for different recognition algorithms. Finally, the authors discussed the potential applications of driving style recognition in intelligent vehicles.]]></description>
      <pubDate>Fri, 26 Jan 2024 10:02:59 GMT</pubDate>
      <guid>https://trid.trb.org/View/2321622</guid>
    </item>
    <item>
      <title>How drivers perform under different scenarios: Ability-related driving style extraction for large-scale dataset</title>
      <link>https://trid.trb.org/View/2317605</link>
      <description><![CDATA[The extraction and analysis of driving style are essential for a comprehensive understanding of human driving behaviors. Most existing studies rely on subjective questionnaires and specific experiments, posing challenges in accurately capturing authentic characteristics of group drivers in naturalistic driving scenarios. As scenario-oriented naturalistic driving data collected by advanced sensors becomes increasingly available, the application of data-driven methods allows for a exhaustive analysis of driving styles across multiple drivers. Following a theoretical differentiation of driving ability, driving performance, and driving style with essential clarifications, this paper proposes a quantitative determination method grounded in large-scale naturalistic driving data. Initially, this paper defines and derives driving ability and driving performance through trajectory optimization modelling considering various cost indicators. Subsequently, this paper proposes an objective driving style extraction method grounded in the Gaussian mixture model. In the experimental phase, this study employs the proposed framework to extract both driving abilities and performances from the Waymo motion dataset, subsequently determining driving styles. This determination is accomplished through the establishment of quantifiable statistical distributions designed to mirror data characteristics. Furthermore, the paper investigates the distinctions between driving styles in different scenarios, utilizing the Jensen–Shannon divergence and the Wilcoxon rank-sum test. The empirical findings substantiate correlations between driving styles and specific scenarios, encompassing both congestion and non-congestion as well as intersection and non-intersection scenarios.]]></description>
      <pubDate>Tue, 16 Jan 2024 09:03:26 GMT</pubDate>
      <guid>https://trid.trb.org/View/2317605</guid>
    </item>
  </channel>
</rss>