<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>Hybrid data-model driven trajectory prediction on highways: Integrating anticipatory interaction awareness and personalized driving preferences</title>
      <link>https://trid.trb.org/View/2604704</link>
      <description><![CDATA[Accurate trajectory prediction of human-driven vehicles (HDVs) in mixed traffic environments is critical for enabling safe and efficient interactions between connected autonomous vehicles (CAVs) and human-driven vehicles on highways. The coexistence of HDVs and CAVs introduces complex dynamics: HDVs exhibit heterogeneous driving preferences influenced by driver behavior patterns, while CAVs’ planned trajectories create anticipatory interactions that reshape HDV motion patterns. Existing methods often overlook the dual challenges of personalized driving preference modelling and anticipatory interaction awareness, particularly in scenarios where CAV trajectories are dynamically integrated into HDV prediction frameworks. To address these challenges, we propose a hybrid data–model driven framework that integrates physics-based behavioral calibration with data-driven interaction modeling. At the core of this framework is the Kepler Optimization-based Temporal Attention Fusion Transformer Network (KO-TAFTN), which enables unified modeling of dynamic historical interactions, anticipatory interactions, and static driving preferences. A driving preference extraction module first derives individualized behavioral traits using Kepler-based physical modeling. These preferences are encoded into context vectors via a static encoder and static enhancement layers, and then incorporated into the network. To improve interpretability and robustness, a variable selection module is applied to evaluate the relevance of input features. A dynamic encoder and a temporal attention fusion module jointly capture and fuse historical and anticipatory interactions by modeling temporal dependencies. Finally, a multimodal trajectory prediction module generates diverse candidate trajectories that reflect potential future motion patterns of HDVs. Experiments demonstrate that the proposed framework consistently outperforms benchmark methods in mixed traffic environments, particularly under complex and congested scenarios. Visualization results further validate the effectiveness of integrating human driving preferences and anticipatory interaction cues. These findings underscore the framework’s potential to improve interaction safety and trajectory accuracy during the transitional evolution of mixed traffic systems.]]></description>
      <pubDate>Mon, 22 Dec 2025 16:07:15 GMT</pubDate>
      <guid>https://trid.trb.org/View/2604704</guid>
    </item>
    <item>
      <title>Composite safety potential field for highway driving risk assessment</title>
      <link>https://trid.trb.org/View/2562336</link>
      <description><![CDATA[In the era of rapid advancements in vehicle safety technologies, driving risk assessment has become a focal point of attention. Technologies such as collision warning systems, advanced driver assistance systems (ADAS), and autonomous driving require driving risks to be evaluated proactively and in real time. To be effective, driving risk assessment metrics must not only accurately identify potential collisions but also exhibit human-like reasoning to enable safe and seamless interactions between vehicles. Existing safety potential field models assess driving risks by considering both objective and subjective safety factors. However, their practical applicability in real-world risk assessment tasks is limited. These models are often challenging to calibrate due to the arbitrary nature of their structures, and calibration can be inefficient because of the scarcity of accident statistics. Additionally, they struggle to generalize across both longitudinal and lateral risks. To address these challenges, the authors propose a composite safety potential field framework, namely C-SPF, involving a subjective field to capture drivers’ risk perception about spatial proximity and an objective field to quantify the imminent collision probability, to comprehensively evaluate driving risks. Different from existing models, the C-SPF is calibrated using abundant two-dimensional spacing data from trajectory datasets, enabling it to effectively capture drivers’ proximity risk perception and provide a more realistic explanation of driving behaviors. Analysis of a naturalistic driving dataset demonstrates that the C-SPF can capture both longitudinal and lateral risks that trigger drivers’ safety maneuvers. Further case studies highlight the C-SPF’s ability to explain lateral driver behaviors, such as abandoning lane changes or adjusting lateral position relative to adjacent vehicles, which are capabilities that existing models fail to achieve.]]></description>
      <pubDate>Fri, 11 Jul 2025 10:00:26 GMT</pubDate>
      <guid>https://trid.trb.org/View/2562336</guid>
    </item>
    <item>
      <title>Trust calibration through perceptual and predictive information of the external context in autonomous vehicle</title>
      <link>https://trid.trb.org/View/2435045</link>
      <description><![CDATA[Maintaining an appropriate level of trust is critical for driving safety in autonomous vehicles. While enhancing the driver’s situation awareness (SA) of system information in autonomous driving is known to significantly promote trust calibration, it remains unclear whether enhancing the driver’s SA of the external context during driving contributes to this calibration. This study addresses this gap by improving SA of the external context during Level 3 (L3) driving automation across various driving environments. Driving contexts were manipulated using distinct road conditions containing low, medium, or high contextual risks. To enhance driver’s SA of the driving context, the authors redesigned the in-vehicle central control panel to display real-time perceptual and predictive information about the external driving context. The authors hypothesized that SA of driving contexts would facilitate trust calibration rather than merely enhancing trust, allowing trust to adjust to appropriate levels under different driving conditions. Experiment 1 examined the impact of perceptual information about the road, traffic infrastructure, and surrounding vehicles on drivers’ trust. The authors found that driver’s trust decreased with increased contextual risk only when the reconfigured panel was used, while the number of accidents was not affected. Experiment 2 investigated the effect of predictive information about the external context on drivers’ trust by marking safe and dangerous zones around driver’s vehicle with green and red areas, respectively. The authors revealed that the predictive information calibrated the trust according to road conditions and increased overall trust levels, while the number of accidents was not affected. Together, these findings suggest that enhancing perception and prediction of external contexts helps drivers align their trust with contextual risk levels in L3 driving automation without compromising driving safety.]]></description>
      <pubDate>Fri, 11 Oct 2024 09:32:01 GMT</pubDate>
      <guid>https://trid.trb.org/View/2435045</guid>
    </item>
    <item>
      <title>Technology to Ensure Equitable Access to Automated Vehicles for Rural Areas (06-004) [supporting dataset]</title>
      <link>https://trid.trb.org/View/2310061</link>
      <description><![CDATA[Project Description: The role of multimodal sensor datasets in training autonomous vehicle machine learning algorithms is crucial. While there are several existing datasets available, the majority of them focus on urban road scenarios. This paper introduces the Rural Road Detection Dataset (R2D2), which aims to overcome this limitation by providing a comprehensive collection of labeled point clouds specifically for object detection and semantic segmentation of rural roads. The dataset encompasses diverse rural environments and road types, creating a challenging learning environment for machine learning algorithms. With over 10,000 labeled point clouds obtained from various locations, R2D2 serves as a valuable resource for researchers and practitioners working towards safer and more efficient transportation systems in rural areas. The authors anticipate that their dataset will expedite the progress of autonomous driving in remote regions, bringing us closer to a future where all roads, regardless of their rural nature, can be navigated with safety and efficiency.  Data Scope: The dataset contains: (1)  LIDAR Point clouds: 10.5K LIDAR Intensity Images; (2) Stereo Images: 5K images with depth maps; (3)  Semantic Annotations for Point-Clouds: 10.5K Point-Clouds with point wise labels; (4) Object Detections labels: 5K 2D Bounding Box labels for camera images; and (5) Calibration parameters: Intrinsic and Extrinsic calibration parameters.]]></description>
      <pubDate>Wed, 27 Dec 2023 10:29:56 GMT</pubDate>
      <guid>https://trid.trb.org/View/2310061</guid>
    </item>
    <item>
      <title>Mixed Traffic Flow State Detection: A Connected Vehicles-Assisted Roadside Radar and Video Data Fusion Scheme</title>
      <link>https://trid.trb.org/View/2230981</link>
      <description><![CDATA[An increasing number of connected vehicles (CVs) driving together with regular vehicles (RVs) on the road is an inevitable stage of future traffic development. As accurate traffic flow state detection is essential for ensuring safe and efficient traffic, the level of road intelligence is being enhanced by the mass deployment of roadside perception devices, which is capable of sensing the mixed traffic flow consisting of RVs and CVs. In this background, the authors propose a roadside radar and camera data fusion framework to improve the accuracy of traffic flow state detection, which utilizes relatively more accurate traffic parameters obtained from real-time communication between CVs and roadside unit (RSU) as calibration values for training the back propagation (BP) neural network. Then, with the perception data collected by roadside sensors, the BP neural network-based data fusion model is applied to all vehicles including RVs. Furthermore, considering the changes of road environments, a dynamic BP fusion method is proposed, which adopts dynamic training by updating samples conditionally, and are applied to fuse traffic flow, occupancy and RVs speed data. Simulation results demonstrate that for CVs data and all vehicles (including RVs) data, the proposed dynamic BP fusion method is more accurate than single sensor detection, entropy based Bayesian fusion method and traditional BP fusion without training by CVs. It can achieve smaller error, and the accuracies of vehicle speed, traffic flow, and occupancy are all above 95%.]]></description>
      <pubDate>Mon, 28 Aug 2023 09:19:11 GMT</pubDate>
      <guid>https://trid.trb.org/View/2230981</guid>
    </item>
    <item>
      <title>FSRDD: An Efficient Few-Shot Detector for Rare City Road Damage Detection</title>
      <link>https://trid.trb.org/View/2082359</link>
      <description><![CDATA[Road damage detection (RDD) is indispensable for safe autonomous driving. Existing RDD models focus on designing feature representations following expert knowledge. However, collecting and labeling all types of samples is time-consuming and leads to insufficient training data. To alleviate the adverse effect of few training samples, a novel few-shot road damage detector (FSRDD) is proposed in this paper to detect rare road damages. The proposed FSRDD includes three stages. First, fully annotated abundant base classes are leveraged to train a base detector, where ghost attention (GA) and proposal feature metric (PFM) modules are developed to eliminate the redundant information and measure the proposal features, respectively. Second, the recognition branch of the detector is fine-tuned using a few samples of all classes. Finally, the test set is inferred with the help of an offline scale-aware prototypical calibration block (SPCB). Extensive experiments show that the authors' FSRDD achieves 10-shot rare road damage detection with 33.4% and 12.9% mAP50 on RDD and CNRDD datasets, respectively, significantly outperforming state-of-the-art methods.]]></description>
      <pubDate>Tue, 25 Apr 2023 09:49:41 GMT</pubDate>
      <guid>https://trid.trb.org/View/2082359</guid>
    </item>
    <item>
      <title>Examination of the effectiveness of multiple training methods on supporting drivers’ better understanding towards level 2 automated vehicle systems</title>
      <link>https://trid.trb.org/View/1572720</link>
      <description><![CDATA[Vehicle automation aims to improve driving safety and reduce the workload and stress of a human driver by transferring the driving task from the driver to control devices. As all automated vehicle control systems have limitations, improper expectations and trust can decrease the effectiveness and safety benefits of an automated system. If a driver over-trusts the system, he or she may use the automation when conditions exceed the capacity of the system, compromising safety. Although previous studies suggest that providing accurate knowledge to users may help train them to establish appropriately calibrated trust of automated systems, limited effort has been done to examine how training can improve drivers’ understanding of the systems, and to develop appropriate training programs for this purpose. The main objective of this study is to develop and evaluate the efficacy of different training methods to promote a driver’s safe operation and trust calibration of automated vehicle (AV) systems. Two AV systems were examined in this study: Adaptive Cruise Control (ACC) and Lane Keeping Assist (LKA). Both knowledge- and skill-based training strategies were developed to describe the AV system functionalities and limitations associated with the two systems. These strategies were later evaluated to determine their impact on producing an accurate mental models each AV system, and were compared with training based on use of the owners’ manual. It was hypothesized that both knowledge and skill-based training leads to better understanding of system limitations and functionalities, compared to exclusive reliance on the owners’ manual.]]></description>
      <pubDate>Fri, 01 Mar 2019 15:51:09 GMT</pubDate>
      <guid>https://trid.trb.org/View/1572720</guid>
    </item>
    <item>
      <title>Effects of a cognitive training and a virtual reality immersion on older adults' cognition and driving abilities</title>
      <link>https://trid.trb.org/View/1460318</link>
      <description><![CDATA[The main objective of this dissertation was to assess the effectiveness of a computerized cognitive training program, combined or not with a driving simulator immersion, on calibration, cognitive and driving performances and subjective well-being of 88 drivers aged 70 and older who were over or under-estimators of their cognitive abilities. Weekly interventions were conducted during three months. An active control group was engaged in a reading magazine activity. Our results show that the cognitive training and the active control activity both improved visual selective attention abilities and driving style (regarding speed adaptation and safety distances). Almost half of the sample (41/88, with 2/3 who were under-estimators), corrected their calibration bias. The driving simulator immersion did not influence the transfer of cognitive training benefits on road. The benefits of the intervention on the self-regulation process to promote the preservation of the driving activity in safe conditions, which is essential for the successful ageing, are discussed. The second objective was to assess the effectiveness of a countermeasure to reduce the simulator sickness, which is a major limitation to the driving simulator use for the older drivers. The effects of the neck galvanic cutaneous stimulation were assessed and did not reduce simulator sickness symptoms of older drivers, contrarily to young drivers. New research avenues are suggested in order to consider the driving simulator as an intervention tool to promote the driving safety.]]></description>
      <pubDate>Fri, 17 Mar 2017 10:37:10 GMT</pubDate>
      <guid>https://trid.trb.org/View/1460318</guid>
    </item>
    <item>
      <title>Challenges facing simulator use in transportation research: lessons from a road safety case study</title>
      <link>https://trid.trb.org/View/1335704</link>
      <description><![CDATA[Over the last decade, the use of simulation in road and transportation research has become commonplace. Whilst commercial off-the-shelf (COTS) products are frequently used, these devices are often designed primarily for use in training. They can therefore present challenges for use in the research environment, particularly for issues concerned with experimental control, data acquisition and fidelity. Transport researchers need to interact with simulator manufacturers more and more often to ensure that a fit-for-purpose product is developed to support unique research needs, and too often anecdotal stories associated with problems with interaction, coordination and communication are retold. This paper provides a case study of the design, development and implementation of road simulator that was procured to investigate how a novel in-vehicle warning technology influenced driving behaviour at rail level crossings. Disparate and conflicting expectations between the research team and simulator developer impacted the fidelity and tractability of the simulation, and unexpected issues arose with data collection, fusion and equipment calibration. Key lessons for future research practice are drawn from this case study. The paper emphasises the importance of interdisciplinarity for informing the design of simulation-based research and proposes concrete methods to innovate interaction and collaboration in similar research.]]></description>
      <pubDate>Wed, 17 Dec 2014 10:56:44 GMT</pubDate>
      <guid>https://trid.trb.org/View/1335704</guid>
    </item>
    <item>
      <title>Research challenges and findings from a driver training pilot study in China</title>
      <link>https://trid.trb.org/View/1250821</link>
      <description><![CDATA[The George Institute in Australia and China collaborated on a pilot study, funded by the FIA Foundation, to develop and evaluate a driver education and training program to reduce novice driver crashes in China. The program was established and implemented in Beijing during 2010-2011. A randomised control trial was conducted with recently licensed drivers recruited through official driving schools. Block randomisation was used to randomise participants to the intervention (n=64) or control group (n=63), with the latter receiving roadside assistance memberships to similar monetary value as the intervention. The intervention included a DVD education program on novice-specific risks plus six hours of in-vehicle training focused on maintaining a safety gap around the vehicle. Participants completed baseline and follow-up questionnaires at approximately four months apart. Recruitment proved challenging, however, once involved, all participants bar one continued through to follow-up. Relatively equal distribution by age, gender and other characteristics was achieved. Very low incidence of any risk taking was reported. Trained participants were significantly more likely to transition to driving-related employment positions and also, but not significantly, reported greater average driving exposure, driving risks and inflated perceptions of their ability than controls. There was no difference in crash involvement (9 participants or 14 per cent in each group). Findings may reflect cultural differences in research familiarity and reporting as much as actual outcomes of the intervention; notwithstanding limitations in participant numbers and self-reported methods applied. The potential for the program to lead to over-calibration and increased risk cannot be discounted. More effort is needed in future studies to build rapport with participants and to have local endorsements to support truthful responding without consequences. Representative research and more local data on novice crash and offence issues is also needed to increase our understanding of novice driver risks in China and how to best address them.]]></description>
      <pubDate>Tue, 21 May 2013 10:42:44 GMT</pubDate>
      <guid>https://trid.trb.org/View/1250821</guid>
    </item>
    <item>
      <title>Calibration as side effect? Computer-based learning in driver education and the adequacy of driving-task-related self-assessments</title>
      <link>https://trid.trb.org/View/1244193</link>
      <description><![CDATA[To reduce the high risk of young, novice drivers being involved in traffic accidents, there have been several attempts to utilize computers for driver education. Previous studies have shown promising results concerning the benefits of using computers for the acquisition of driving-task-related cognitive skills. However, these studies' findings are inconclusive regarding whether using computers for driver education affects drivers' calibration skills. Underdeveloped calibration skills are considered to be an important reason explaining why young, novice drivers are at a higher risk of being involved in an accident relative to other drivers. To examine the effects of computer-based learning in driver education on drivers' calibration skills, the authors provided student drivers (N = 38) with two different types of learning material (computer-based vs. paper-based, approximately 90 min in duration). Two days later, they presented them with a driving simulator task. Right before the test, the participants were asked to predict the likelihood that they would be able to successfully implement their newly acquired competencies. The authors chose "anticipatory recognition of hazardous traffic situations" as the learning objective to examine both facets of calibration: accuracy of assessing driving tasks (situational or risk awareness) and accuracy of driving-task-related self-assessments (self-efficacy, state awareness). The analysis of participant's gaze data confirmed the expectation that student drivers who used computer-based learning material would not only detect situation-specific hazard cues sooner but would also demonstrate better comprehension of the information they perceived. Contrary to expectations, the computer-based learning did not lead to more accurate predictions of test performance. However, it increased the insecurities of the participants, thereby reducing the risk that these student drivers would overestimate their own competence. Because using computers helps student drivers to develop better hazard-perception skills and more defensive self-efficacy expectations, the implementation of computers in driver education is more likely to support safe behavioral patterns in traffic than conventional methods.]]></description>
      <pubDate>Thu, 14 Mar 2013 12:45:07 GMT</pubDate>
      <guid>https://trid.trb.org/View/1244193</guid>
    </item>
    <item>
      <title>Goals for Driver Education. Application of the GDE Theoretical Framework on Elderly Drivers</title>
      <link>https://trid.trb.org/View/890779</link>
      <description><![CDATA[Throughout the last 20 years, a number of evaluation studies have shown that driver education and training has not always resulted in the expected outcome. Several studies have shown none or even a negative effect on road safety. This has led VTI to focus research on why this is the case and how driver education and training could be developed in order to meet the expectations of improved safety. One of the conclusions that have been reached is that basic and advanced training should not focus too much on vehicle maneuvering issues. It should also address higher order skills such as making decision about when, where and how trips should be made, making drivers aware of risks that typically affect the type of drivers they belong to and provide possibilities to assess their own limitations and abilities as drivers. As a result of this research, the Goals for Driver Education (GDE) matrix was developed, which is a theoretical framework for defining what competencies different types of drivers need in order to be safe drivers. The GDE matrix has a hierarchical approach and defines 4 levels of competences where abilities on the lower levels are dependent of abilities on a higher level. The first level defines the basic car control skills. The second level concerns the application of these abilities to driving in traffic where traffic rules and other road users must be taken into consideration. The third level is trip related, which means abilities to make decisions about when, where and how trips should be made. On the fourth level, aspects that are not directly traffic related, but rather related to individual goals and social life and preconditions are at focus. These four levels have been divided into three dimensions, defining on each level the (1) necessary knowledge and skills, (2) specific risks that are connected to each level, and (3) how to assess own status with respect to abilities, knowledge and understanding – a calibration of own abilities in order to avoid over- and under-confidence. In this paper it is shown how the matrix may be used for development of driver education and training for elderly drivers.]]></description>
      <pubDate>Fri, 19 Jun 2009 09:28:44 GMT</pubDate>
      <guid>https://trid.trb.org/View/890779</guid>
    </item>
    <item>
      <title>SAFE DRIVING AND THE TRAINING OF CALIBRATION: LITERATURE REVIEW</title>
      <link>https://trid.trb.org/View/706463</link>
      <description><![CDATA[It is assumed that the essential issue in safe driving is not so much the development of specific skills, but the ability to balance task demands and skills accurately. This balancing of demands and capabilities is also known as 'calibration'. This paper explores theories relating to calibration and investigates whether, and how, to incorporate the issue of calibration in formal driving instruction. The literature thus far supports the thesis of the paper that calibration is a core issue in safe driving. Inexperienced drivers show less awareness than experienced drivers of the actual realities of road system operation, and less awareness of their own role. Calibration is conceptualised as not just momentary demand regulation, but also as behavioural regulation on the basis of anticipated events (hazards). It is theorized that the problem with young drivers lies both in the anticipatory realm and in momentary demand regulation. Miscalibration can cause for instance: small safety margins, excessive speed, aggressive driving, short following distances, and the performance of risky manoeuvres. Current driver training does not prevent miscalibration, and may even stimulate miscalibration. This is related to the fact that the training does not incorporate enhancing learning conditions for the driver after qualification. A correct calibration of task demands and coping abilities largely depends on the amount of practice and the amount and quality of feedback that a driver receives. It is suggested that driver training should incorporate methods to match self-assessed ability to actual ability. Drivers should learn to actively search for, and use, the feedback that the driving environment provides them with.]]></description>
      <pubDate>Wed, 06 Mar 2002 00:00:00 GMT</pubDate>
      <guid>https://trid.trb.org/View/706463</guid>
    </item>
  </channel>
</rss>