Single-Variable Scenario Analysis of Vehicle-Pedestrian Potential Crash Based on Video Analysis Results of Large-Scale Naturalistic Driving Data

Tian, Renran; Li, Lingxi; Yang, Kai; Jiang, Feng; Chen, Yaobin; Sherony, Rini

doi:10.1007/978-3-319-21070-4_30

Cited by 6 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To achieve this goal, a large-scale naturalistic driving study was carried out to collect driving scene videos from 110 vehicles during one-year period (Tian, 2015), with all vehiclepedestrian and vehicle-bicyclist encountering cases identified (Yang, et al, 2013A, 2013B) (Liu, et al, 2016), examined, and labeled to create a comprehensive pedestrian and bicyclist behavior database. Among these cases, a random sample of potential conflict cases were analyzed in this study to calculate all targeted variables.…”

Section: Research Objectivesmentioning

confidence: 99%

“…However, considering the nature of crash records typically being built on memories of event witnesses and reconstruction of the crash scenes, important scenario parameters describing the starting point of these events and pre-collision pedestrian movements are usually excluded. Thus, it is valuable to use naturalistic driving data in supplementing critical details which are usually missing and/or biasedly recorded in these databases (Tian, et al, 2015) (Sherony, et al, 2016).…”

mentioning

confidence: 99%

See 1 more Smart Citation

Pedestrian Moving Patterns during Potential Conflicts with 110 On-Road Driving Vehicles

Tian¹,

Chien²,

Chen

et al. 2019

Proceedings of the Human Factors and Ergonomics Society Annual

Self Cite

View full text Add to dashboard Cite

As the most commonly seen vulnerable road users, protection and interaction with pedestrians are key functionalities in vehicle active safety and self-driving research areas. Development and evaluation of such systems require deeper understanding of pedestrian behaviors, especially motion patterns, in different driving environments. Traditionally, most of the pedestrian movement studies rely on fixed roadside cameras in specific road locations with higher pedestrian density, like intersections and junctions. Although these studies can provide information to describe pedestrian walking behavior and vehicle-pedestrian interactions in micro and macro levels, there are two main limitations. Firstly, pedestrian movement data are rarely collected from the vehicle’s point of view, which makes some critical variables difficult to be collected related to pedestrian initial appearance situation. Secondly, insufficient data are acquired to cover low- pedestrian-density road environments like mid-block, rural areas, and small un-controlled intersections. In this study, we focus on three important pedestrian movement variables including appearance distances, initial time-to-collision, and crossing speed under different driving and road scenarios. Based on a large-scale naturalistic driving study, crossing pedestrians were randomly captured in the scene videos from 110 passenger cars when potential ego-vehicle-to-pedestrian conflicts appeared during a one-year period. Motion data of these pedestrians were then analyzed to calculate the targeted behavior measurements, with the empirical results reported.

show abstract

Section: Research Objectivesmentioning

confidence: 99%

mentioning

confidence: 99%

Pedestrian Moving Patterns during Potential Conflicts with 110 On-Road Driving Vehicles

Tian¹,

Chien²,

Chen

et al. 2019

Proceedings of the Human Factors and Ergonomics Society Annual

Self Cite

View full text Add to dashboard Cite

show abstract

“…From the TASI 110-car naturalistic driving video database, the labeled pedestrian and bicyclist potential conflict database was first retrieved [17,18,19]. This potential conflict database is composed of 15-second video clips representing the cases marked as potential conflict with higher encountering risks [18], containing about 5000 pedestrian and bicyclist videos.…”

Section: Sherony Et Al / Sae Int J Trans Safety / Volume 4 Issue mentioning

confidence: 99%

“…Multi-step data analysis was performed on the collected 94TB of data to detect all the encounters with pedestrians and bicyclists, assign scenario labels, and then calculate comprehensive road and driving features. For more details about the TASI 110-car naturalistic driving dataset, please refer to the earlier publications [17,18,19].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Pedestrian/Bicyclist Limb Motion Analysis from 110-Car TASI Video Data for Autonomous Emergency Braking Testing Surrogate Development

Sherony¹,

Tian

Chien

et al. 2016

SAE Int. J. Trans. Safety

Self Cite

View full text Add to dashboard Cite

Many vehicles are currently equipped with active safety systems that can detect vulnerable road users like pedestrians and bicyclists, to mitigate associated conflicts with vehicles. With the advancements in technologies and algorithms, detailed motions of these targets, especially the limb motions, are being considered for improving the efficiency and reliability of object detection. Thus, it becomes important to understand these limb motions to support the design and evaluation of many vehicular safety systems. However in current literature, there is no agreement being reached on whether or not and how often these limbs move, especially at the most critical moments for potential crashes. In this study, a total of 832 pedestrian walking or cyclist biking cases were randomly selected from one large-scale naturalistic driving database containing 480,000 video segments with a total size of 94TB, and then the 832 video clips were analyzed focusing on their limb motions. We modeled the pedestrian/bicyclist limb motions in four layers: (1) the percentages of pedestrians and bicyclists who have limb motions when crossing the road; (2) the averaged action frequency and the corresponding distributions on when there are limb motions; (3) comparisons of the limb motion behavior between crossing and non-crossing cases; and (4) the effects of seasons on the limb motions when the pedestrians/bicyclists are crossing the road. The results of this study can provide empirical foundations supporting surrogate development, benefit analysis, and standardized testing of vehicular pedestrian/ bicyclist detection and crash mitigation systems.

show abstract

Modeling Vehicle-Pedestrian Encountering Risks in the Natural Driving Environment Using Machine Learning Algorithms

Gandhi

Luo

Tian

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Single-Variable Scenario Analysis of Vehicle-Pedestrian Potential Crash Based on Video Analysis Results of Large-Scale Naturalistic Driving Data

Cited by 6 publications

References 12 publications

Pedestrian Moving Patterns during Potential Conflicts with 110 On-Road Driving Vehicles

Pedestrian Moving Patterns during Potential Conflicts with 110 On-Road Driving Vehicles

Pedestrian/Bicyclist Limb Motion Analysis from 110-Car TASI Video Data for Autonomous Emergency Braking Testing Surrogate Development

Modeling Vehicle-Pedestrian Encountering Risks in the Natural Driving Environment Using Machine Learning Algorithms

Contact Info

Product

Resources

About