No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Developing evasive action-based indicators for identifying pedestrian conflicts in less organized traffic environments: Developing Evasive Action-Based Indicators
Journal of Advanced Transportation
Abstract
There has been a growing interest in using surrogate safety measures such as traffic conflicts to analyse road safety from a broader perspective than collision data alone. This growing interest has been aided by recent advances in automated video-based traffic conflict analysis. The automation enables accurate calculation of various conflict indicators such as time-to-collision and post-encroachment time. These indicators rely on road users getting within specific temporal and spatial proximity from each other and therefore assume that proximity is a surrogate for conflict severity. However, this assumption may not be valid in many driving environments where close interactions between road users are common. The objective of this paper is to investigate the applicability of time proximity conflict indicators for evaluating pedestrian safety in less-organized traffic environments with a high mix of road users. Several alternative behavioural conflict indicators based on detecting pedestrian evasive actions are recommended to better measure traffic conflicts in such traffic environments. These indicators represent variations in the spatio-temporal gait parameters (step length, step frequency and walk ratio) immediately before the conflict point. A highly congested shared intersection in Shanghai, China, with frequent pedestrian conflicts is used as a case study. Traffic conflicts are analysed with the use of automated video-based analysis techniques. The results showed that evasive action-based indicators have higher potential to identify pedestrian conflicts and measure their severity in high mix less organized traffic environments than time proximity measures such as time-to-collision and post-encroachment time. Copyright
... This can be measured by the distance or the temporal proximity to the collision when the evasive action starts. For example, the Time To Collision (TTC) before the vehicle starts reacting [10], [11], [12] or the Post Encroachment Time (PET) corresponding to the TTC at the minimal approach distance [11], [12] have been used. The speed of the vehicle before it starts reacting can also be used, as proposed by [11]. ...
... This can be measured by the distance or the temporal proximity to the collision when the evasive action starts. For example, the Time To Collision (TTC) before the vehicle starts reacting [10], [11], [12] or the Post Encroachment Time (PET) corresponding to the TTC at the minimal approach distance [11], [12] have been used. The speed of the vehicle before it starts reacting can also be used, as proposed by [11]. ...
... This metric reflects the average degree and frequency of pedestrian velocity changes, due to avoidance maneuvers. Similarly, variations in walking speed and in pedestrian trajectory have been used by [10], [11] and [12] to measure pedestrian comfort. These metrics assume that changes in walking behavior to avoid the car require effort from pedestrians, which may cause discomfort. ...
The navigation of autonomous vehicles around pedestrians is a key challenge when driving in urban environments. It is essential to test the proposed navigation system using simulation before moving to real-life implementation and testing. Evaluating the performance of the system requires the design of a diverse set of tests which spans the targeted working scenarios and conditions. These tests can then undergo a process of evaluation using a set of adapted performance metrics. This work addresses the problem of performance evaluation for an autonomous vehicle in a shared space with pedestrians. The methodology for designing the test simulations is discussed. Moreover, a group of performance metrics is proposed to evaluate the different aspects of the navigation: the motion safety, the quality of the generated trajectory and the comfort of the pedestrians surrounding the vehicle. Furthermore, the success/fail criterion for each metric is discussed. The implementation of the proposed evaluation method is illustrated by evaluating the performance of a pre-designed proactive navigation system using a shared space crowd simulator under Robot Operating System (ROS).
... La sévérité d'une collision potentielle peut être mesurée par la distance ou la proximité temporelle de la collision lorsque l'action d'évitement commence. Par exemple, le TTC avant que l'AV ne commence à réagir a été utilisé par [109,169,209]. Le Post Encroachment Time correspondant au TTC à la distance d'approche minimale à été utilisé par [169,209]. ...
... Par exemple, le TTC avant que l'AV ne commence à réagir a été utilisé par [109,169,209]. Le Post Encroachment Time correspondant au TTC à la distance d'approche minimale à été utilisé par [169,209]. Une autre mesure est la vitesse de l'AV avant qu'il ne commence à réagir, comme proposé par [169]. ...
... Cette métrique reflète le degré moyen et la fréquence des changements de vitesse des piétons, dus aux manoeuvres d'évitement. De même, les variations de la vitesse de marche et de la trajectoire des piétons ont été utilisées par [169] et [209] pour mesurer le confort des piétons. Ces métriques supposent que les modifications du comportement de marche pour éviter la voiture demandent un effort de la part des piétons, ce qui peut être source d'inconfort. ...
Dans un futur proche, les véhicules autonomes ne seront plus limités aux routes et auront à naviguer dans des environnements urbains denses et dynamiques comme les espaces partagés. Dans de tels espaces, la séparation entre piétons et véhicules est réduite au minimum et les usagers doivent négocier leur passage sans règles de circulation explicites. Les piétons naviguent selon certaines normes sociales et s'attendent à ce que le véhicule autonome ait une navigation sécurisée, mais aussi efficace et conforme aux conventions sociales et urbaines. Pour cela, un élément clé de la navigation dans les espaces partagés repose sur la compréhension et l'anticipation des comportements piétons et de leurs interactions.Cependant, nous ne savons pas comment les piétons vont se comporter, car actuellement il est très rare que des véhicules autonomes naviguent dans le même espace que des piétons. Notre problématique est donc la suivante : comment anticiper les comportements piétons dans un espace partagé avec un véhicule autonome ?Cette thèse s'inscrit dans le cadre du projet ANR HIANIC (Human Inspired Autonomous Navigation In Crowds). Dans cette thèse, nous étudions le comportement des piétons en espace partagé avec un véhicule autonome en modélisant et en simulant des comportements piétons réalistes.Notre approche intègre des observations empiriques et des concepts issus des sciences sociales afin de proposer un modèle et un simulateur à base d'agents pour une application en robotique. À chaque étape, le modèle proposé a été évalué et validé par des simulations de plusieurs scénarios et des comparaisons avec des données réelles.Notre première contribution est un modèle des comportements piétons individuels dans un contexte d'espace partagé. Le modèle tient compte de la perception, de l'attention et de l'espace personnel des piétons pour simuler des foules peu denses en environnement ouvert.Notre deuxième contribution est un modèle de quatre relations sociales au sein des groupes de piétons (couples, amis, familles et collègues de travail). Le modèle permet de simuler à la fois le mouvement des groupes sociaux de piétons dans plusieurs contextes de foule et les comportements d'évitement des groupes par d'autres piétons.Notre troisième contribution est un modèle des comportements piétons en interaction avec un véhicule autonome en espace partagé.Le modèle permet de représenter des comportements piétons à la fois hétérogènes, précis et explicables dans plusieurs situations d'interaction. Le modèle peut être utilisé pour reproduire des scènes du monde réel et pour prédire les trajectoires des piétons autour d’un véhicule autonome en temps réel.Notre quatrième contribution est l'implémentation du modèle pour proposer le simulateur SPACiSS, "Simulator for Pedestrians and an Autonomous Car in Shared Spaces". SPACiSS est open source et permet de simuler les interactions entre les piétons et les véhicules dans différents espaces partagés. Avec l'intégration dans le cadre ROS, couramment utilisé en robotique, SPACiSS est conçu comme un environnement pour le test de systèmes de navigation autonome.Nous avons montré que la modélisation et la simulation à base d'agents peuvent contribuer à une intégration efficace entre les sciences sociales et la robotique. Cette association est prometteuse pour aborder des scénarios du monde réel.
... In addition to spatio-temporal proximity indicators, recent studies have increasingly utilized evasive action-based indicators to analyze non-motorized traffic conflicts. Tageldin et al. [21][22][23] demonstrated that indicators based on evasive maneuvers are more effective than spatio-temporal indicators in quantifying traffic conflicts between two-wheelers and pedestrians under mixed traffic conditions. Guo et al. [24] investigated the powered twowheeler conflicts in less-organized traffic environments, finding that the yaw rate ratio (YRR) can effectively measure the severity of conflict between e-bikes and bicycles on shared paths. ...
... Traffic conflicts are generally categorized into angle conflicts, head-on conflicts, and rear-end conflicts [22]. Video observations in the study area indicate that rear-end conflicts are the most prevalent, while head-on conflicts occur less frequently, and angle conflicts are almost nonexistent. ...
The rapid growth of e-bikes has intensified traffic conflicts on slow-moving shared paths in China. This study analyzed traffic safety of pedestrians and non-motorized vehicles and examined the factors influencing conflict severity utilizing traffic conflict techniques. Video-based surveys were conducted on six shared paths in Shenzhen, and conflict trajectory was extracted by Petrack software (Version 0.8). The minimum Time to Collision and Yaw Rate Ratio were selected as conflict indicators. Fuzzy c-means clustering was employed to classify conflicts into three severity levels: 579 potential conflicts, 435 minor conflicts, and 150 serious conflicts. Nineteen feature variables related to road environment, traffic operation, conflict sample information, and conflict behavior were considered. A SMOTE random forest model was constructed to explore critical influencing factors systematically. The results identified ten key factors affecting conflict severity. The increase in conflict severity is associated with the rise in pedestrian proportion and flow, and the decrease in e-bike proportion and flow. Male participants and pedestrians are more likely to engage in serious conflicts, while illegal lane occupation and wrong-way travel further elevate the severity level. These findings can provide references for traffic engineers and planners to enhance the safety management of shared paths and contribute to sustainable non-motorized transport.
... In the definition of evasive action-based measures, conflicts have been associated with such occurrences as the appearance of brake lights or the unexpected changing of lanes or direction [12]. Several identified evasive actions, such as braking or lane-changing, are often precautionary measures rather than indicators of imminent danger for collisions [13][14][15]. Consequently, a strong correlation between collisions and conflicts might not be established if conflicts are only defined based on observed evasive actions. Moreover, in scenarios where vehicles in two incidents share similar driving states but differ significantly in vehicle proximity, they could be assigned the same level of conflict risk based on the assessment of evasive actions taken. ...
... 14The variables of time related to speed and direction for two vehicles. ...
Accurate identification and analysis of traffic conflicts through surrogate safety measures (SSMs) are crucial for safety evaluation in road systems. Existing SSMs for conflict identification and analysis mostly consider the temporal–spatial proximity of conflicts without taking into account the severity of potential collisions. This makes SSMs unsuitable for traffic safety evaluation in complex road environments. In order to address the shortcomings above, this study first introduces a new SSM called the Potential Conflict Risk Index (PCRI). To validate the effectiveness of PCRI, the inD dataset is adopted for conflict identification comparison between time-to-collision (TTC) and PCRI. Using PCRI, this study conducts a conflict analysis in the freeway merging areas based on the data from the Outer Ring Expressway Dataset (ORED), accounting for differences between cars and trucks. The comparative results between TTC and PCRI show that PCRI can provide a more comprehensive identification of conflicts and a more accurate identification of the moment with the highest conflict risk. The results of conflict analysis suggest that conflicts occur more frequently in situations involving trucks, and these conflicts commonly occur in closer proximity to the on-ramp at freeway merging areas. The findings from this study can improve the accuracy of conflict identification under different conflict patterns, enhancing the specificity of traffic safety measures and ultimately ensuring the safety of road systems.
... (p. 5)" In line with this, Tageldin and Sayed (2016) showed that traffic conflict indicators based on sudden evasive action were better at identifying pedestrian conflicts and estimating their severity than traditional proximity indicators like time to collision. Bagdadi and Várhelyi (2011) further demonstrated that jerky, abrupt road user behavior is indicative of increased crash risk. ...
... The requires that "The conflict resulting from the trajectory of the conflict partners is not premeditated (planned) by at least one conflict partner", and is conceptually similar to existing conflict metrics based on "jerky" behavior (Tageldin and Sayed (2016); Bagdadi and Várhelyi (2011)). However, combining the type of surprise metrics described here with spatiotemporal proximity is a novel ...
The quantitative measurement of how and when we experience surprise has mostly remained limited to laboratory studies, and its extension to naturalistic settings has been challenging. Here we demonstrate, for the first time, how computational models of surprise rooted in cognitive science and neuroscience combined with state-of-the-art machine learned generative models can be used to detect surprising human behavior in complex, dynamic environments like road traffic. In traffic safety, such models can support the identification of traffic conflicts, modeling of road user response time, and driving behavior evaluation for both human and autonomous drivers. We also present novel approaches to quantify surprise and use naturalistic driving scenarios to demonstrate a number of advantages over existing surprise measures from the literature. Modeling surprising behavior using learned generative models is a novel concept that can be generalized beyond traffic safety to any dynamic real-world environment.
... To identify conflicts, a TTC threshold of 2s is usually used, but a TTC between 1.6 and 2 is considered to have a low collision risk (Sayed & Zein 1999). Therefore, it was decided to use a maximum TTC threshold of 1.5 seconds and a maximum PET threshold of 1.5 seconds to capture those severe conflicts that are more likely to be surrogates for crashes (Milosavljevic 2018;Tageldin & Sayed 2016). ...
... Furthermore, severe vehicle-to-pedestrian conflicts are reported for three different thresholds combinations for TTC and PET that are based on investigations by other researchers. For example, in Wu et al., the TTC parameter was set at 2.7 seconds and the PET was set at 8 seconds (Wu et al. 2017), while in (Milosavljevic 2018;Tageldin & Sayed 2016) a maximum TTC threshold of 1.5 seconds and a maximum PET threshold of 1.5 seconds were used. Table 5 Delay and vehicle-to-pedestrian conflicts for ten scenarios at the Gerrard and Church intersection (% change compared to Scenario 10 in parentheses) ...
The provision of leading pedestrian intervals (LPI) has emerged in recent years to achieve safety equitability for pedestrians at signalized intersections. LPI is a way to provide the pedestrian walk interval a few seconds before starting the circular green indication to adjacent parallel traffic. Although the safety benefit of LPI is indisputable, there are fundamental questions that need to be addressed for the optimal deployment of this strategy. First, can significant safety benefits for pedestrians be achieved while maintaining a satisfactory operational level of service for vehicles? Second, what are the application circumstances most conducive to achieving the greatest safety benefits for pedestrians? Third, how can a jurisdiction effectively assess contemplated treatments to achieve optimal deployment? This exploratory paper addresses these three fundamental questions by reviewing relevant literature before presenting the research from the application of microsimulation to fifteen Toronto intersections where LPIs have been implemented. The microsimulation involved using a recently released module for accommodating LPI phasing in the PTV Vistro software. To directly address the first and second questions, vehicle-to-pedestrian conflicts and vehicle delay were estimated for ten scenarios that allowed for the provision of, and variability in the LPI interval, right turn volumes, right turn on red provision, pedestrian and vehicle volumes, and crossing width. The results suggest that significant safety benefits can be achieved for pedestrians while maintaining a satisfactory level of service for vehicles. They further suggest that potential LPI deployments need to be assessed on a case-by-case basis since the effects of LPI can be significantly impacted by the influencing factors investigated. Statistical models were developed to quantify the effects of LPI implementation on vehicle-to-pedestrian conflicts after controlling for pedestrian and turning vehicle volumes. The results of this exploratory investigation, though interesting and consistent with the literature and logical considerations, may not be generalizable in a strict sense. Nevertheless, the study does provide a blueprint for investigating the design, traffic, and operational factors that can influence the impact of LPI on pedestrian safety without detrimentally impacting vehicle level of service.
... What was considered as a breakthrough in this domain was the automation of trajectory extraction for vehicles and pedestrians directly from videos (Saunier and Sayed, 2006). Later, several works used trajectories extracted from videos to analyze before-and-after vehicle-pedestrian conflicts (Ismail et al., 2010) in street designs with elements of shared space (Kaparias et al., 2013), in less organized traffic environments (Tageldin and Sayed, 2016), and in the context of vehicle-bicycle conflicts at intersections (Sayed et al., 2013). ...
... Errors and failures in the process of detection will propagate to the process of tracking, which will later lead to wrong conclusions in the analysis (Sayed et al., 2013). (2) The majority of the non-deep learning studies mentioned above focused on vehicle-pedestrian interactions (Ismail et al., 2009(Ismail et al., , 2010Kaparias et al., 2013;Tageldin and Sayed, 2016;Ni et al., 2016) and only a few studies on vehicle-bicycle interactions (Sayed et al., 2013). However, in real-world traffic situations at big intersections, other heterogeneous road users are often involved. ...
Learning how road users behave is essential for the development of many intelligent systems, such as traffic safety control, intelligent transportation systems, and self-driving cars. However, automated and accurate recognition of road users’ behavior is still one of the bottlenecks in realizing such systems in city traffic that is—compared to other types of traffic—especially dynamic and full of uncertainties. There are some urban environments which make detection and prediction of road users’ behavior particularly challenging, e. g., temporarily shared spaces of intersections for vehicle turning or shared spaces as a traffic design. The former allow vehicles to turn and interact with other crossing road users, the latter intended to make different types of road users share the space, therefore reducing the dominance of vehicles and improving pedestrian movement and comfort. Direct interactions between vehicles and vulnerable road users (VRUs, e. g., pedestrians
and cyclists) and ambiguous traffic situations (e. g., road users negotiating usage of the road) make road users’ behavior stochastic and difficult to predict.
With the development of deep learning techniques and the availability of large-scale real traffic data, this thesis proposes deep conditional generative models for automated interaction detection in the temporarily shared spaces of intersections, as well as for trajectory prediction in shared spaces as a traffic design. Models based on Conditional Variational Auto-Encoder (CVAE) are trained to map deterministic input (e. g., a sequence of video data or a segment of an observed trajectory) to many possible outputs of road users’ behavior, characterized by their interaction or their movement in the next seconds.
This thesis makes two main contributions to the research on modeling road users’ behavior in shared spaces using deep learning approaches:
(1) The interaction detection model takes the information of road users’ type, position, and motion—all of which have been automatically extracted by deep learning object detectors and optical flow from video data—as input, and generates a frame-wise probability that represents the dynamics of interaction between a turning vehicle and any VRUs involved. The model’s efficacy was proven by testing on real-world datasets acquired from two different intersections. It achieved an F1-score above 0.96 at a right-turn intersection in Germany and 0.89 at a left-turn intersection in Japan, both with very busy traffic flows.
(2) Various factors and state-of-the-art deep learning architectures are investigated for trajectory prediction. In this thesis, three frameworks based on CVAE are proposed for accurate multi-path trajectory prediction of heterogeneous road users in shared spaces as a traffic design. The latent space of the CVAE is trained for encoding stochastic behavior patterns and the multi-sampling process from the trained latent space enables the frameworks to generate not only one deterministic future trajectory, but multiple possible future trajectories for each road user. The first framework focuses on studying multiple contexts, namely motion, interaction, pedestrian grouping, and different types of environmental scene context for trajectory prediction. The second and third frameworks focus on exploring dynamic context (i. e., motion and interaction) using attention
mechanisms, and improving the models’ generalizability—predicting trajectories of heterogeneous road users in various shared spaces that have not been used to train the models. All of the frameworks, but the second and third in particular, showed superior performance on various popular open-source datasets and benchmarks. The last two frameworks even took first place (in different submission times) in one of the most widely recognized open challenges (TrajNet online test) by reducing the overall average and final displacement errors of the predicted trajectories in the next 4.8 seconds to 0.353 meters and 1.179 meters, respectively.
... The literature on traffic conflicts, which aims to systematically evaluate objective risk based on observed interactions without collisions, tends to frame risk as ''conflict severity,'' where more severe conflicts represent greater likelihood of a collision (8,9,15,(22)(23)(24)(25)(26). The severity of a conflict is usually assessed as qualitative levels (e.g., low, medium, high) by human observers or quantitative bins based on objective conflict indicators such as time to collision (TTC) or post encroachment time (PET). ...
... No single conflict measure best captures all types of interactions and risks, but PET has been used successfully and frequently in the past for non-motorized traveler interactions (22,24). We use a threshold of 5 s PET to define interactions, based on past research (7,9,10,24). Pedestrian interactions with only one direction of motor vehicle and bicycle traffic at each location were considered, so that the vehicle approach was included in the camera scene. ...
Assessments of interactions between road users are crucial to understanding comfort and safety. However, observers may vary in their perceptions and ratings of road user interactions. The objective of this paper is to examine how perceptions of yielding, comfort, and safety for pedestrian interactions vary among observers, ranging from members of the public to road safety experts. Video clips of pedestrian interactions with motor vehicles and bicycles were collected from 11 crosswalks and shown to three groups of participants (traffic safety experts, an engaged citizen advisory group, and members of the general public) along with questions about yielding, comfort, and risk of injury. Experts had similar views of yielding and comfort to the other two groups, but a consistently lower assessment of injury risk for pedestrians in the study. Respondent socio-demographics did not relate to perceptions of yielding, comfort, or risk, but self-reported travel habits did. Respondents who reported walking more frequently rated pedestrian comfort as lower, and respondents who reported cycling more frequently rated risk as lower for pedestrian interactions with both motor vehicles and bicycles. Findings suggest small groups of engaged citizens can provide useful information about public perspectives on safety that likely diverge from expert assessments of risk, and that sample representation should be assessed in relation to travel habits rather than socio-demographics.
... The intensity of swerving can be measured by the angular velocity of the vehicle's rotation (yaw rate), and these indicators usually perform better than proximity indicators in measuring the conflict severities of motorcycles, bicycles and pedestrians in less-organized traffic environments (Tageldin et al., 2015;Guo et al., 2018). For vehicle-pedestrian conflicts, pedestrian step frequency, step length and their changes can also be used (Tageldin and Sayed, 2016). It is noted that the traffic conflict definition introduces uncertainty related to the type/strength of evasive action to be taken to avoid a collision. ...
... This uncertainty causes objective conflict indicators to only represent a partial image of the true severity of traffic events (Ismail et al., 2011). It may also explain why subjective assessment of traffic conflicts by human observers may, in some cases, capture more sophisticated severity aspects than objective conflict indicators (Tageldin and Sayed, 2016). ...
Limitations of crash data and crash-based methods have given rise to the study of alternate measures of safety that are not predicated on the occurrence of a crash such as traffic conflicts. The popularity of these alternative safety measures will likely play a prominent role in road safety analysis in the forthcoming era of connected and autonomous vehicles because of the vast amount of real-time vehicle data that are likely to be available. While traffic conflicts and crashes share the same failure mechanism in the driving process, which allows models of crash frequency and severity to be applied to model conflict frequency and severity, modeling traffic conflicts has new challenges because of their distinct characteristics. This paper provides a comprehensive review of research studies that have used traffic conflicts as a safety measure, and to present conceptual and methodological issues associated with these studies. It is found that although substantial progress has been made in the modeling methodologies of traffic conflicts over the years, more research efforts are needed. Some promising directions for future research are outlined and discussed.
... Because of their unpredictable trajectories and poor visibility [69], researchers have found that right-turn motions are hazardous for pedestrians. Even though SSMs have been thoroughly researched in rich countries, research is still being done to modify these models to fit the diverse and mixed traffic situations of emerging nations [70], therefore resolving the shortcomings of the current models for these settings. ...
This study looks at how Surrogate Safety Measures (SSMs) are used to evaluate traffic safety at signalized crossings when there is mixed traffic. Real-time flexibility, integration with Connected and Automated Vehicles (CAVs), and data-driven modeling are critical areas that still require investigation. Most SSM models were developed for human-driven vehicles and require modifications for automated environments. Future research should explore how CAVs influence key conflict metrics such as Time-to-Collision and Post Encroachment Time. The integration of Internet of Things technologies, LiDAR, and sensor-based systems presents opportunities to enhance real-time traffic safety assessments. A significant challenge is the lack of standardized SSM thresholds, leading to inconsistencies in risk assessment models. Addressing this requires large-scale meta-analyses and adaptive threshold-setting methodologies. Additionally, traditional SSM methodologies may not fully capture complex traffic interactions. Machine learning and deep learning techniques can enhance safety assessments by dynamically adjusting to real-world conditions. This review systematically analyzes the limitations of current SSM methodologies and proposes AI-driven frameworks for improved predictive accuracy, contributing to proactive safety interventions and more intelligent traffic management strategies.
... 12.68) (Meng and Weng, 2011). Fundamentally, pedestrian collision avoidance would depend on conflicting vehicle kinematics, vehicle characteristics (such as size and type), and the pedestrian's detailed gait parameters (e.g., step length and frequency) (Tageldin and Sayed, 2016) and behavioral cues (e.g., attention and head rotation) (Losada et al., 2023). However, due to data limitations and the absence of robust quantitative links between these factors and pedestrian braking capacity, it is assumed that pedestrian braking capacity is similar to vehicles, where the driver perceives pedestrian braking behavior as analogous to vehicles' braking behaviour. ...
... Chen et al. [20] analyzed the conficts between right-turning vehicles and pedestrians at signalized intersections using PET and RTTC, discovering that pedestrian-vehicle conficts are concentrated near the crosswalks at the exit lanes. Tageldin et al. [24] also analyzed pedestrian-vehicle conficts at intersections using TTC and PET, fnding that these conficts are unevenly distributed across the crosswalk. Fu et al. [2] categorized pedestrianvehicle interactions at intersections into primary and secondary interactions, revealing signifcant diferences in the average speed and acceleration of right-turning vehicles between primary and secondary interactions, while PET showed no signifcant diference. ...
Due to the high incidence of pedestrian collisions at intersections and the vulnerable status of pedestrians, an increasing number of researchers are focusing on pedestrian safety. Currently, researchers tend to apply surrogate safety measures (SSMs) in pedestrian safety analysis, with Time to Collision (TTC) and Post Encroachment Time (PET) being the most widely used. However, their application is subject to certain conditions and limitations. Therefore, this study proposes a novel traffic conflict indicator called improved Time Advantage (improved TAdv), designed to identify pedestrian-vehicle conflicts without collision trajectories. A conflict model for right-turning vehicles and pedestrians at signalized intersections is developed to analyze the impact of various factors on the probability of pedestrian-vehicle conflicts. The results indicate that the improved TAdv offers advantages over traditional traffic conflict indicators. Meanwhile, the findings emphasize the spatial disparities of right-turning pedestrian-vehicle conflicts and analyze the impact of vehicle speed at different locations on the likelihood of such conflicts, offering new insights into pedestrian safety analysis.
... Non-statistical methods determine conflict severity based on changes in evasive maneuvers, such as deceleration, jerk, yaw rate, braking, steering, and acceleration. Several studies have demonstrated the usefulness of evasive maneuver-based metrics in measuring the severity of traffic conflicts between motor vehicle users (Bagdadi and Várhelyi, 2013;Guo et al., 2018;Tageldin and Sayed, 2016;Tageldin et al., 2015;Zaki and Sayed, 2013). However, it is difficult to characterize the relationship between the influence of multiple factors in this way. ...
Motorized vehicles (MV) and non-motorized vehicles (NMV) are mixed in the intersection center area (ICA). This mixing leads to complicated interactions between vehicles, which seriously affects traffic safety, especially at mixed intersections of high density. To deep understanding of the interaction course between motorized and non-motorized vehicles in ICAs.Two intersections with a high density of interaction behavior between motorized and non-motorized vehicles were investigated through high-resolution traffic video. Firstly, to extract high-precision trajectories from roadside video, we proposed a new trajectory extraction framework that integrates Yolov7, Deepsort, and the trajectory reconstruction algorithm, which integrated the social force model and particle filtering (SFPF) proposed in our previous research. Second, 183 complete interaction events between motorized and non-motorized vehicles were extracted based on the surrogate safety indicator TTC, and latent variables affecting the course of interaction behavior between motorized and non-motorized vehicles were defined based on turning direction, kinetic state, surrounding environment, signal light, vehicle action behavior, and types of NMV. Third, an ordered logit model was built to study the interactions. Analyzing the significance of the model showed that the following variables have a significant effect on the severity of the conflict (p<0.05or lower): the turning directions of the two vehicles, their speeds, steering behaviors, the distance between the conflict point and the vehicle, the signal plan, and the surrounding environment. The study contributes to developing active safety control and driver assistance strategies.
... GPS map data was also collected from the vehicle-mounted camera, including the exact location of the participant as well as the vehicle's speed and acceleration rate in G's. Video data were coded as a traffic conflict if there was "an observable event that could lead to a crash unless one of the involved parties changed its movement by slowing down, changing lanes, or accelerating to avoid a collision" (Cafiso & Di Silvestro, 2011;Tageldin & Sayed, 2016). It was observed that the drivers attempted to avoid collisions by performing evasive maneuvers such as accelerating, decelerating, changing the vehicle's movement path, or stopping. ...
Talking with passengers while driving may impair driver performance. However, little is known about the extent to which this factor affects driver behavior. Since distracted driving accounts for most pedestrian accidents, the present study seeks to demonstrate the danger posed to pedestrian crossing safety by talking with passengers. Based on a real-world driving approach, the current study examines how talking with a passenger affects a driver’s behavior when encountering pedestrians. A study of 41 participants’ driving behavior at different locations was conducted, and 2,922 conflicts with pedestrians were coded according to four conditions (without or with passengers, and at marked crossings or at other locations). Using binary regression, it was found that this factor near pedestrian crossing areas causes significant dysfunction in driving yielding behavior. Distraction negatively affects driver-yielding behavior when drivers engage in speeding, resulting in a fivefold decrease in evasive maneuvers. In addition, a multinomial regression model indicated that drivers would reduce their evasive maneuvers by more than 60%, with lane changes accounting for most of these maneuvers. Also, the present study concluded that pedestrian crossing gaps should be at least two times larger when drivers converse with passengers than when they do not do so. This study may contribute to the development of legislation, policy, countermeasures, and future research aimed at reducing distracted driving.
... Past studies have shown that proximity-based indicators fail to capture the evasive actions (Tageldin and Sayed, 2016). An evasive action generally involves a significant change either in the speed or in the direction of the road user to avoid the collision. ...
... Tageldin et al. (2015) use an automated video-based analysis technique to detect jerk rates, in order to measure traffic conflicts as indicators of safety. Tageldin and Sayed (2016) suggest that evasive action-based indicators, which represent variations in the spatio-temporal gait parameters (i.e., step length, step frequency and walk ratio), are possible indicators of pedestrian conflicts. ...
Large-scale telematics data enable a high-resolution inference of road network’s safety conditions and driver behavior. Although many researchers have investigated how to define meaningful safety surrogates and crash predictors from telematics, no comprehensive study analyzes the driver behavior derived from large-scale telematics data and relates them to crash data and the road networks in metropolitan cities. This study extracts driver behavior indices (e.g., speed, speed variation, hard braking rate, and hard acceleration rate) from large-scale telematics data, collected from 4000 vehicles in New York City five boroughs. These indices are compared to collision frequencies and collision rates at the street level. Moderate correlations were found between the safety surrogate measures and collision rates, summarized as follows: (i) When normalizing crash frequencies with traffic volume, using a traffic AADT model, safety-critical regions almost remain the same. (ii) The correlation magnitude of hard braking and hard acceleration varies by road types: hard braking clusters are more indicative of higher collision rates on highways, whereas hard acceleration is a stronger hazard indicator on non-highway urban roads. (iii) Locations with higher travel times coincide with locations of high crash incidence on non-highway roads. (iv) However, speeding on highways is indicative of collision risks. After establishing the spatial correlation between the driver behavior indices and crash data, two prototype safety metrics are proposed: speed corridor maps and hard braking and hard acceleration hot-spots. Overall, this paper shows that data-driven network screening enabled by telematics has great potential to advance our understanding of road safety assessment.
... The video data collected from the vehicle-mounted camera also provided GPS map data (i.e., pinpointing the participant's exact location), as well as the vehicle's speed and rate of acceleration in G's. When coding the video data, a traffic conflict was defined as "an observable event which would end in a crash unless one of the involved parties change its movement by slowing down, changing lanes, or accelerating to avoid collision" (Cafiso et al., 2017;Tageldin and Sayed, 2016). In these conflicts, the drivers were observed to attempt to prevent the collision by performing an evasive maneuver such as acceleration, deceleration, changing the movement path, or stopping the vehicle. ...
A significant proportion of global road crashes are attributed to unsafe driving behaviors. The current study aimed to explore potential differences in driving behaviors across experienced and novice drivers using two separate approaches; a questionnaire study and an instrumented vehicle study (IVS). The analysis of 260 questionnaires and 1,372 traffic interactions within the IVS revelated that driving experience affects driving performance for different driving tasks. Factor analysis of the questionnaire data revealed the impact of driving errors, lapses, violations, and aggressive violations on the behavior of novice and experienced drivers. Behavioral models of novice and experienced drivers encountering other road users were determined using binary logistic regression. The results showed that novice drivers were more likely to engage in driving violations while experienced drivers were more likely to engage in aggressive violations. Unauthorized speeding, zigzag movements, using a mobile phone while driving, and unauthorized overtaking on roads were the most frequent driving violations by novice drivers. The most frequent aggressive violations by experienced drivers were tempting other drivers to create a race and chasing other drivers. These findings may be used as a framework to facilitate safer driving behaviors by reducing errors, lapses, violations and aggressive violations, and facilitating safety-promoting attitudes.
... The alternative model (Figure 1b) defines the conflict as the outcome of evasive actions. This track is followed if conflict is defined on the basis of TTCmin, but also many other outcomebased indicators, such as Post-Encroachment Time, PET (Allen et al. 1978), as well as indicators describing the intensity of the evasive action itself (Tageldin & Sayed 2016;Bagdadi 2013;Gettman et al. 2008;af Wåhlberg 2004). Being defined this way, conflicts land on a parallel track with collisions since by the moment the conflict is identified we can be sure that the collision has already been avoided (knowing that the lowest TTC value is TTCmin, we also know that it cannot go down to zero to become a crash). ...
The study presents a simple and easy to implement method for detection of the evasive action start in traffic interactions. The method is based on comparison of the studied trajectory with a reference set of ‘unhindered’ trajectories, interpreting the start of evasive action as the moment when no more similarities can be found. The suggested algorithm performs well for primary interactions when road users arrive in an unhindered state. It fails, however, in case of secondary interactions. Explorative application of the method on a large dataset of normal and conflict traffic situations concludes that traffic conflicts occur more frequently in secondary interactions, presumably due to higher cognitive load on the involved road users. Despite the limitations, the method can be used both for the safety studies based on traffic conflicts and for more general quantification and visualisation of the road user behaviour.
... Bagdadi (2013) defined the conflict severity as a combination of Delta-V, TA, and the maximum average deceleration indicators, in which the two latter indicators estimate the effectiveness of deceleration as an evasive manoeuvre. Tageldin and Sayed (2016) investigated the applicability of time proximity indicators, such as PET and TTC, to evaluate pedestrian safety in less-organised traffic environments with heterogeneous traffic complexities. The authors recommended evasive action-based indicators representing variations in the spatiotemporal gait parameters (i.e., step length and frequency) as complementary measures to evaluate pedestrian traffic conflict severity (similar to Medina et al., 2008). ...
Interactions of motorised vehicles with pedestrians have always been a concern in traffic safety. The major threat to pedestrians comes from the high level of interactions imposed in uncontrolled traffic environments, where road users have to compete over the right of way. The interactions become more complex with the variety of user types and their available conflict resolution strategies. In this research, a conflict risk evaluation model is developed to assess the safety level of pedestrian conflict with other road users. Surrogate safety indicators are employed to measure road users’ temporal and spatial proximity during a conflict. The thresholds are determined through the application of various methods (i.e., intersection point, p-tile, maximum between-class variance, and minimum cross-entropy method) to separate potential critical conflicts against normal traffic conditions, on the basis of the conflict risk evaluation model. An F-score method is used to select the optimal threshold given by various applied methods. Two data sets of shared space and mid-block were used to develop and validate conflict risk evaluation models for the interaction of pedestrians with vehicles (passenger cars) and light vehicles (two- or three-wheel vehicles) separately. The proposed model can potentially be used as a real-time conflict risk evaluation model to improve traffic safety.
... Kathuria and Vedagiri (2020) and Kadali and Vedagiri (2016) suggested that 11% to 18.5% of the pedestrian-vehicle interactions at unsignalised crossings in India were critical. Tageldin and Sayed (2016) reported more than 950 pedestrian-vehicle, pedestrian-bicycle, and pedestrianmotorcycle conflicts per hour at a congested intersection in Shanghai, China, where the pedestrian volume is 1200 pedestrians per hour. Table 1 shows that overall mobile phone use was higher among pedestrians who experienced a conflict (19%) than those who did not (13.2%). ...
Pedestrian distraction related to mobile phone use has become an emerging road safety issue. While much research has investigated the effects of pedestrian distraction on risky crossing behaviours, little is understood about potential risk compensation behaviours and the relationships between pedestrian distraction and safety outcomes, such as near-misses or involvement in a conflict situation. This paper aims to explore pedestrian distraction and its associations with other risky crossing behaviours and safety outcomes, using field observation data in Hanoi, Vietnam. Associations between pedestrian distraction and risky behaviours or conflict situations were identified using log-binomial and logistic regression. Results show that around 14.4% of pedestrians were distracted while crossing. Interestingly, while distracted pedestrians were more likely not to observe traffic while crossing, they were less likely to violate traffic signals or cross outside the marked crosswalks. These findings suggest distracted pedestrians may have risk compensation behaviours. Results also indicated a relatively small yet significant effect of mobile phone distraction on conflict situations.
... In normal braking situation, the maximum jerk is around -8 m/s 3 , while critical situation can be -9.9 to -12.6 m/s 3 [27]. However, the jerk at -8 m/s 3 is chosen as a maximum threshold to express the critical situation and to prevent biologically infeasible deceleration ramp up [26,28]. ...
The rear-end collision is one of the top motorcycle crashes in many ASEAN countries. Monitoring incoming motorcycle from behind by using inside rearview mirror can reduce risk of this accident. However, there is no any protocol for assessing the visibility of rearview mirror in being followed by a motorcycle scenario. This project aims to investigate how lead driver detects the following motorcycle and develop the assessment procedure.
The procedure for measuring the visibility of the rearview mirror used a test object to simulate the lamp position of motorcycle in longitudinal direction behind the tested vehicle. The images in the rearview mirror were photographed by using a camera mounted on a 50 th percentile Hybrid III male mockup model as driver representative of the view observation. The criteria of assessment came from current riding characteristic of motorcyclists which keep their headway very close to the vehicle in front at 0.5 s with speed range from 10 km/h to 60 km/h. To pass the test, the lamp of the test object has to be seen in the mirror in between 1.4 m and 8.4 m longitudinal and in between-100 cm and 100 cm lateral of area behind the vehicle.
Four vehicle models, which were selected from the top-selling vehicles in ASEAN market, were tested to demonstrate the proposed procedure. Three of the four models failed the test and the remaining model got 2.14 points out of maximum 4 points. However, if the simulation of luggage blocking is used, the score will be reduced to 1.34 points. The results could be interpreted that the standard rearview mirror could not detect a following motorcycle in current situation. This proposed procedure can promote technologies that can increase the rearview visibility. Further assessment procedure for preventing rear end collision in the future is also proposed. Also, this proposed procedure can be integrated into ASEAN NCAP test protocol for advanced rear visualization version 2019 as a supplement.
... For example, the Time To Collision (TTC) is proposed in some studies [Kap+13; Pas20; TS16]. Whereas, the Post Encroachment Time (PET) corresponding to the TTC at the minimal approach distance have been used in other works [Pas20;TS16]. Another metric proposed by [Pas20] for the collision severity is the AV speed before the AV starts avoiding the collision. ...
Application of deep learning to geometric 3D data poses various challenges for researchers. The complex nature of geometric 3D data allows to represent it in different forms: occupancy grids, point clouds, meshes, implicit functions, etc. Each of those representations has already spawned streams of deep neural network models, capable of processing and predicting according data samples for further use in various data recognition, generation, and modification tasks.Modern deep learning models force researchers to make various design choices, associated with their architectures, learning algorithms and other specific aspects of the chosen applications. Often, these choices are made with the help of various heuristics and best practice methods discovered through numerous costly experimental evaluations. Probabilistic modeling provides an alternative to these methods that allows to formalize machine learning tasks in a meaningful manner and develop probability-based training objectives. This thesis explores combinations of deep learning based methods and probabilistic modeling in application to geometric 3D data.The first contribution explores how probabilistic modeling could be applied in the context of single-view 3D shape inference task. We propose a family of probabilistic models, Probabilistic Reconstruction Networks (PRNs),which treats the task as image conditioned generation and introduces a global latent variable, encoding shape geometry information. We explore different image conditioning options, and two different training objectives based on Monte Carlo and variational approximations of the model likelihood. Parameters of every distribution are predicted by multi-layered convolutional and fully-connected neural networks from the input images. All the options in the family of models are evaluated in the single-view 3D occupancy grid inference task on synthetic shapes and according image renderings from randomized viewpoints. We show that conditioning the latent variable prior on the input images is sufficient to achieve competitive and state-of-the-art single-view 3D shape inference performance for point cloud based and voxel based metrics, respectively. We additionally demonstrate that probabilistic objective based on variational approximation of the likelihood allows the model to obtain better results compared to Monte Carlo based approximation.The second contribution proposes a probabilistic model for 3D point cloud generation. It treats point clouds as distributions over exchangeable variables and use de Finetti’s representation theorem to define a global latent variable model with conditionally independent distributions for coordinates of each point. To model these point distributions a novel type of conditional normalizing flows is proposed, based on discrete coupling of point coordinate dimensions. These flows update the coordinates of each point sample multiple times by dividing them in two groups and inferring the updates for one group of coordinates from another group and, additionally, global latent variable sample by the means of multi-layered fully-connected neural networks with parameters shared for all the points. We also extend our Discrete Point Flow Networks (DPFNs) from generation to single-view inference task by conditioning the global latent variable prior in a manner similar to PRNs from the first contribution. Resulting generative performance demonstrates that DPFNs produce sets of samples of similar quality and diversity compared to state of the art based on continuous normalizing flows, but are approximately 30 times faster both in training and sampling. Results in autoencoding and single-view inference tasks show competitive and state-of-the-art performance for Chamfer distance, F-score and earth mover’s distance similarity metrics for point clouds.
... Pedestrian temporal violation behavior is attributed to a higher propensity of pedestrian crashes as such behavior is not expected to be encountered from other road users (2,3). Pedestrian temporal violation behavior is defined as committing crossing at improper traffic signal phases, for example, during the pedestrian red signal or ''Don't Walk'' signal (4)(5)(6). ...
This study aims to model pedestrian temporal violation behavior at signalized crosswalks. Video data of pedestrian crossing behavior were collected from three locations in China and were used to investigate the effect of several factors on pedestrian temporal violation behavior. The temporal violation behavior was analyzed using the relationship between pedestrian waiting duration and their endurance probabilities. A fully parametric duration model with Weibull distribution was used to model the temporal violation behavior, and the cluster-specific heterogeneity among the three study sites was accounted for using random intercepts. Six variables were identified to significantly affect the violation behavior: pedestrian gender and phone distraction status, location type, pedestrian volume, day of the week, and time of the day. The results show that pedestrians are likely to disobey traffic regulations when there are longer waiting durations. Male pedestrians have a higher violation tendency than females. Pedestrians distracted by their phones have longer waiting durations than undistracted pedestrians. Signalized road segment crosswalks are associated with higher temporal violation propensity than signalized intersection crosswalks. Pedestrians are more likely to commit violations at higher pedestrian densities. Weekdays are associated with shorter waiting durations and higher violation tendency than weekends. Pedestrians are more likely to violate traffic regulations in the morning than at midday and in the evening. These findings give insights into the pedestrian crossing behavior to better accommodate pedestrians and improve safety.
... Examples of temporal proximity measures include time headway and Time-to-Collision (TTC) whilst spatial proximity measures include examples such as Proportion of Stopping Distance (PSD). The key advantage of spatial and/or temporal proximity measures is their ability to capture the severity of a traffic conflict interaction in an objective and quantifiable way [13]. However, this assumption may not be applicable in all driving environments, for example, in congested traffic settings where road users are predisposed to regular vehicular interactions and are frequently near each other. ...
With the ever-increasing advancements in the technology of driver assistant systems, there is a need for a comprehensive way to identify traffic conflicts to avoid collisions. Although significant research efforts have been devoted to traffic conflict techniques applied for junctions, there is dearth of research on these methods for motorways. This paper presents the validation of a traffic conflict prediction algorithm applied to a motorway scenario in a simulated environment. An automatic video analysis system was developed to identify lane change and rear-end conflicts as ground truth. Using these conflicts, the prediction ability of the traffic conflict technique was validated in an integrated simulation framework. This framework consisted of a sub-microscopic simulator, which provided an appropriate testbed to accurately simulate the components of an intelligent vehicle, and a microscopic traffic simulator able to generate the surrounding traffic. Results from this framework show that for a 10% false alarm rate, approximately 80% and 73% of rear-end and lane change conflicts were accurately predicted, respectively. Despite the fact that the algorithm was not trained using the virtual data, the sensitivity was high. This highlights the transferability of the algorithm to similar road networks, providing a benchmark for the identification of traffic conflict and a relevant step for developing safety management strategies for autonomous vehicles.
... For example, the Time To Collision (TTC) is proposed in some studies [Kap+13; Pas20; TS16]. Whereas, the Post Encroachment Time (PET) corresponding to the TTC at the minimal approach distance have been used in other works [Pas20;TS16]. Another metric proposed by [Pas20] for the collision severity is the AV speed before the AV starts avoiding the collision. ...
The current trend in electric autonomous vehicles design is based on pre-existing models of cities which have been built for cars. The carbon footprint of cities cannot be reduced until the overall requirement for vehicles is reduced and more green and pedestrianized zones are created for better livability. However, such green zones cannot be scaled without providing autonomous mobility solutions, accessible to people with reduced mobility. Such solutions need to be capable of operating in spaces shared with pedestrians, which makes this a much harder problem to solve as compared to traditional autonomous driving. This thesis serves as a starting point to develop such autonomous mobility solutions. The work is focused on developing a navigation system for autonomous vehicles operating around pedestrians. The suggested solution is a proactive framework capable of anticipating pedestrian reactions and exploiting their cooperation to optimize the performance while ensuring pedestrians safety andcomfort.A cooperation-based model for pedestrian behaviors around a vehicle is proposed. The model starts by evaluating the pedestrian tendency to cooperate with the vehicle by a time-varying factor. This factor is then used in combination with the space measurements to predict the future trajectory. The model is based on social rules and cognitive studies by using the concept of the social zones and then applying the deformable virtual zone concept (DVZ) to measure the resulting influence in each zone. Both parts of the model are learnt using a data-set of pedestrians to vehicle interactions by manually annotating the behaviors in the data-set.Moreover, the model is exploited in the navigation system to control both the velocity and the local steering of the vehicle. Firstly, the longitudinal velocity is proactively controlled. Two criteria are considered to control the longitudinal velocity. The first is a safety criterion using the minimum distance between an agent and the vehicle’s body. The second is proactive criterion using the cooperation measure of the surroundingagents. The latter is essential to exploit any cooperative behavior and avoid the freezing of the vehicle in dense scenarios. Finally, the optimal control is derived using the gradient of a cost function combining the two previous criteria. This is possible thanks to a suggested formulation of the cooperation model using a non-central chi distribution for the distance between the vehicle and an agent.A smooth steering is derived using a proactive dynamic channel method for the space exploration. The method depends on evaluating the navigation cost in a channel (sub-space) using a fuzzy cost model. The channel with the minimum cost is selected, and a human-like steering is affected using a Quintic spline candidate path between channels. Finally, the local steering is derived using a sliding mode path follower.The navigation is evaluated using PedSim simulator under ROS in pedestrian-vehicle interaction scenarios. The navigation is tested with different pedestrian density and sparsity. The proactive framework managed to navigate the vehicle producing smooth trajectories while maintaining the pedestrians’ safety and reducing the travel time in comparison with traditional reactive methods (Risk-RRT).
... Behavioural stress responses can be both beneficial and harmful, 4 and may result in poorer task performance (Hills et al., 2019). Observed behavioural responses such as speed, gait, or path deviations have been used in traffic safety research to identify traffic conflicts based on "evasive maneuvers" (Kim et al., 2014;Tageldin and Sayed, 2016). ...
Understanding perceptions of safety and comfort (PSC) while walking or cycling is essential to accommodating and encouraging active travel, but current measures of PSC, primarily surveys, suffer from validity and reliability issues. Physiological markers of stress like electrodermal activity and heart rate variability have been proposed as alternative, objective measures of PSC. This paper presents a literature summary and conceptual framework examining the use of physiological stress markers during walking and cycling. The existing studies of active traveller stress markers report inconsistent findings and account for limited controls. We propose a comprehensive conceptual framework to describe the array of dynamic stimuli experienced during active travel, with complex appraisals and multidimensional stress responses that feedback to travel behaviour and stimuli exposure, and culminate in a set of physiological outcomes triggered by activation of the autonomic nervous system – all moderated by numerous personal and trip-related factors. The key challenge of inferring traffic-related fear or discomfort from physiological markers measured on-road is potential confounding effects of: (1) non-traffic factors that induce or modify stress responses, (2) traffic factors that induce stress responses not associated with safety or comfort, and (3) personal and environmental factors that directly influence physiological measurements outside of a stress response. No physiological stress marker has yet been shown to be reliable for on-road active travellers, particularly not for inter-subject comparisons. Physiological markers have the potential to provide high-resolution, objective information about pedestrian and cyclist PSC, but further research, particularly controlled experiments, and more precise study framing are needed to ensure validity and address moderating and confounding factors.
... Traffic environments at urban road intersections are complex and diverse due to frequent traffic conflicts between vehicles [1]. The safety of intersections in urban road networks is important because they are key nodes in the urban road network, where all types of traffic participants (e.g., motor vehicles and pedestrians) must meet and disperse. ...
Urban road intersections are one of the key components of road networks. Due to complex and diverse traffic conditions, traffic conflicts occur frequently. Accurate traffic conflict detection allows improvement of the traffic conditions and decreases the probability of traffic accidents. Many time-based conflict indicators have been widely studied, but the sizes of the vehicles are ignored. This is a very important factor for conflict detection at urban intersections. Therefore, in this paper we propose a novel time difference conflict indicator by incorporating vehicle sizes instead of viewing vehicles as particles. Specially, we designed an automatic conflict recognition framework between vehicles at the urban intersections. The vehicle sizes are automatically extracted with the sparse recurrent convolutional neural network, and the vehicle trajectories are obtained with a fast-tracking algorithm based on the intersection-to-union ratio. Given tracking vehicles, we improved the time difference to the conflict metric by incorporating vehicle size information. We have conducted extensive experiments and demonstrated that the proposed framework can effectively recognize vehicle conflict accurately.
... Despite being suitable to replicate multiple traffic dynamics, agent-based models are complex to calibrate and do not capture risk-taking behavior in an explicit manner. Some efforts to incorporate some of these dynamics, especially the evasive maneuvers' dynamics, into different aforementioned modeling frameworks have also been attempted with different levels of success (21)(22)(23)(24)(25)(26)(27). Finally, the ''data-driven approaches'' models of Yuan and Daamen are classified as ''data-in-the-loop'' models and ''data-in-the-model'' models. ...
Pedestrians are among the travelers most vulnerable to collisions that are associated with high fatality and injury rates. The increasing rate of urbanization and mixed land-use construction make walking (along with other non-motorized travel) a predominant transportation mode with a wide variety of behaviors expected. Because of the inherent safety concerns seen in pedestrian transportation infrastructures, especially those with conflicting multimodal movements expected (crosswalks, transit platforms, etc.), it is important that pedestrian behavior is modeled as a risk-taking stochastic behavior that may lead to errors and thus collision formation. In previous work, the complexity and cost associated with building pedestrian models in a cognitive-based environment weighted down the construction of simulation tools that can capture pedestrian-involved collisions, including those seen in shared space environments. In this paper, a tool that will help evaluate the safety of pedestrian traffic is initiated: an extended modeling framework of pedestrian walking behavior is adopted while incorporating different physiological, physical, and decision-making elements. The focus is on operational decisions (i.e., path choices defined by longitudinal and lateral trajectories) with a pre-specified set of origins and destinations. The model relies on the prospect theory paradigm where pedestrians evaluate their acceleration and directional alternatives while considering the possibility of colliding with other “particles.” Using a genetic algorithm method, the new model is calibrated using detailed trajectory data. This model can be extended to model the interactions between a variety of different modes that are present in different mixed land-use environments.
... Several studies have attempted to integrate different measures for conflict analysis, and most of them are for vehicle conflicts (Ismail et al., 2011;Zheng et al., 2018;Wang et al., 2019;Cavadas et al., 2020). Moreover, the evasive action based measures, which have been proven to be more suitable for vulnerable road users (Tageldin et al., 2015;Tageldin and Sayed, 2016), have been rarely integrated with proximity measures for pedestrian conflict analysis. ...
Pedestrians confront risky situations at unsignalized crosswalks when they are consecutively interacting with motorized vehicles and non-motorized vehicles while crossing. This study aims to investigate the safety of pedestrians with a new perspective that focuses on consecutive conflicts occurring during pedestrian crossing. Based on about 9 h video data collected by an unmanned aerial vehicle from six unsignalized crosswalks of a roundabout, consecutive conflicts were identified, and an integrated severity index that combines post encroachment time, jerk and yaw rate ratio was proposed to measure the severity of consecutive conflicts. Moreover, bivariate logistic models that account for and not account for the correlation between the pedestrian-motorized vehicle (P-MV) conflict and the pedestrian-non-motorized vehicle (P-NV) conflict of a consecutive conflict were developed, and speed-, count-, time to zebra-related factors and other factors of involved road users were considered in the models. A total of 899 consecutive conflicts were identified and on average one in six pedestrians encountered consecutive conflicts. The bivariate logistic modeling results show that the model accounting for the correlation significantly outperform its counterpart. A negative correlation is found between the severities of P-MV conflict and P-NV conflict, and the P-NV conflict is more likely to be the serious one. It is also found that speed of motorized vehicle and time to zebra for the first conflicting subject are the common factors that affect the severities of both P-NV conflicts and P-MV conflicts, while speed of pedestrian, speed of non-motorized vehicle, number of motorized vehicles, number of non-motorized vehicles, group and direction of pedestrians have significant effects on the severity of either P-MV conflicts or P-NV conflicts.
... The work by Ismail et al., [19], one of the early studies using computer vision techniques, automatically analyzed pedestrian-vehicle conflicts at an intersection using the extracted trajecotries from video data. Later on, similarly, several works used trajectories extracted from videos to analyze before-and-after vehicle-pedestrian conflicts [20], in street designs with elements of shared space [21], in less organized traffic environments [22], and vehicle-bicycle conflicts at intersections [18]. The work carried out by Ni et al., [23] analyzed pedestrian-vehicle interaction patterns using indicators of Time-to-collision (TTC) [24] and Gap Time (GT) [19]. ...
Intersections where vehicles are permitted to turn and interact with vulnerable road users (VRUs) like pedestrians and cyclists are among some of the most challenging locations for automated and accurate recognition of road users' behavior. In this paper, we propose a deep conditional generative model for interaction detection at such locations. It aims to automatically analyze massive video data about the continuity of road users' behavior. This task is essential for many intelligent transportation systems such as traffic safety control and self-driving cars that depend on the understanding of road users' locomotion. A Conditional Variational Auto-Encoder based model with Gaussian latent variables is trained to encode road users' behavior and perform probabilistic and diverse predictions of interactions. The model takes as input the information of road users' type, position and motion automatically extracted by a deep learning object detector and optical flow from videos, and generates frame-wise probabilities that represent the dynamics of interactions between a turning vehicle and any VRUs involved. The model's efficacy was validated by testing on real--world datasets acquired from two different intersections. It achieved an F1-score above 0.96 at a right--turn intersection in Germany and 0.89 at a left--turn intersection in Japan, both with very busy traffic flows.
... The body of research on this topic suggests that a number of aspects, such as route choice, placement, walking speed and positioning in relation to other known and unknown pedestrians, could be taken into consideration in evaluations of energy-efficient outdoor lighting for pedestrians. Extracting this kind of microscopic data on pedestrian movement behaviour, as well as road users, in general, is a relatively common method used in urban traffic studies and has been achieved via simple, semi-automated [48,49] or fully automated [50,51] tools. Theoretically, with proper camera calibration and sufficient video resolution, very accurate measurements of road users' positions and speeds are possible [52]. ...
When daylight hours are limited, pedestrians are dependent on appropriate outdoor lighting. Although new city lighting applications must consider both energy usage and pedestrian responses, current methods used to capture pedestrian walking behaviour during dark conditions in real settings are limited. This study reports on the development and evaluation of a video-based method that analyses pedestrians’ microscopic movements (VAPM—video analysis of pedestrian movements), including placement and speed, in an artificially lit outdoor environment. In a field study utilising between-subjects design, 62 pedestrians walked along the same path under two different lighting applications. VAPM accurately discriminated pedestrians’ microscopic movements in the two lighting applications. By incorporating methodological triangulation, VAPM successfully complemented observer-based assessments of pedestrians’ perceptions and evaluations of the two lighting applications. It is suggested that in evaluations of pedestrian responses to city lighting applications, observer-based assessments could be successfully combined with an analysis of actual pedestrian movement while walking in the lit environment. However, prior to employing a large-scale application of VAPM, the methodology needs to be further adapted for use with drones and integration into smart city lighting systems.
... In the field of test and comparison the performance of two safety measures for crashes, several studies used ρs for this purpose [12][13][14]. This correlation coefficient is predominantly used as a nonparametric alternative to a traditional coefficient of correlation and can be utilized under general conditions. ...
Comparative study for safety performance measures of
signalized intersection sites.
Raghad Zeki Abdul Majeed1 and Dr. Hussein A Ewadh2
1 PhD Student in Highway and Transportation Engineering Department, Faculty of Engineering, Al-
Mustansiriyah University, Baghdad, Iraq;
2 Professor Road and Traffic Engineering, Civil Engineering Department, College of Engineering,
University of Babylon, Babylon, Iraq.
Email: raghad.zeki@yahoo.com
Abstract. The decision on which safety performance measure to use for diagnosis of
sites with probable to profit from safety improvements is a major factor in making the
right decision for the guidance of resources. This study concentrated on introducing a
quantitative comparison of sites according to four crash-based methods suggested in
the Highway Safety Manual (HSM). Archived crash data for three years are obtained
to conduct a comparison course. As a case study, the analysis and measure of safety
for 9 four-leg signalized intersections in Baghdad city have been done independently
using Empirical Bayes method (EB-method), observed Crash Frequency (CF), Crash
Rate (CR) and Empirical Bayes Adjustment (EB-adj.) as safety performance measures
in HSM. In this paper the EB-method is used as a benchmark for comparison. The
safety measures are evaluated through rank correlation analysis while hazard location
identification results are compared through the use of values rank-based mean
absolute error. Quantitative evaluation tests showed that each of EB-adj and observed
CF correlated well with EB-method while CR method exhibited poor performance in
comparison with EB-method and was the worst in hazard location identification. This
result is quite confusing since many agencies still depend on the CR method in traffic
safety analyses.
Pedestrian streets, also known as streets closed to motorized traffic, serve to promote active modes of transportation. This concept offers the potential to enhance safety for the most vulnerable road users while concurrently reducing air pollution. The present study aims to evaluate the safety of interactions between pedestrians and cyclists, focusing on three pedestrian streets within the city of Montreal. Video data was collected during the day in the summer of 2021 . Following camera calibration, a total of 80 h of data was analyzed. Each road user detected and tracked was categorized as either a “pedestrian” or “cyclist”. The analysis involves the computation of indicators for individual cyclists (speed and acceleration) and for their interactions with pedestrians (distance and time to collision [TTC]). Two multivariate regression models were estimated to analyze the relationship between TTC or the cyclist speed as the dependent variables and several other factors. The findings from the safety analysis reveal a discernible variation in safety indicator values between distinct sites, even those situated on the same thoroughfare, independent of regulatory measures. The statistical analysis indicates that elevated TTC values correspond to high acceleration and increased distances between pedestrians and cyclists. Moreover, high TTC values are associated negatively with the density of pedestrians within the camera’s field of view. In contrast, concerning speed, high values are linked to low TTC and distances, together with elevated acceleration values.
Objective:
Motorized vehicles (MV) and non-motorized vehicles (NMV) are mixed in the intersection center area (ICA). This mixing leads to complicated interactions between vehicles, which seriously affects traffic safety, especially at mixed intersections of high density. To deep understanding of the interaction course between motorized and non-motorized vehicles in ICAs.
Methods:
Two intersections with a high density of interaction behavior between motorized and non-motorized vehicles were investigated through high-resolution traffic video. Firstly, to extract high-precision trajectories from roadside video, we proposed a new trajectory extraction framework that integrates Yolov7, Deepsort, and the trajectory reconstruction algorithm, which integrated the social force model and particle filtering (SFPF) proposed in our previous research. Second, 183 complete interaction events between motorized and non-motorized vehicles were extracted based on the surrogate safety indicator TTC, and latent variables affecting the course of interaction behavior between motorized and non-motorized vehicles were defined based on turning direction, kinetic state, surrounding environment, signal light, vehicle action behavior, and types of NMV. Third, an ordered logit model was built to study the interactions.
Results:
Analyzing the significance of the model showed that the following variables have a significant effect on the severity of the conflict (p < 0.05 or lower): the turning directions of the two vehicles, their speeds, steering behaviors, the distance between the conflict point and the vehicle, and the surrounding environment. The vehicles entering the ICA 10 s before the end of the signal phase have a higher probability of having a serious crash event while making the interaction.
Conclusions:
The study contributes to developing active safety control and driver assistance strategies.
Facilitating proactive pedestrian safety management, the application of extreme value theory (EVT) models has gained popularity due to its extrapolation capabilities of estimating crashes from their precursors (i.e., conflicts). However, past studies either applied EVT models for crash risk analysis of autonomous vehicle–pedestrian interactions or human-driven vehicle–pedestrian interactions at signalised intersections. However, our understanding of human-driven vehicle–pedestrian interactions remains elusive because of scant evidence of (i) EVT models’ application for heterogeneous traffic conditions, (ii) appropriate set of determinants, (iii) which EVT approach to be used, and (iv) which conflict measure is appropriate. Addressing these issues, the objective of this study is to investigate pedestrian crash risk analysis in heterogeneous and disordered traffic conditions, where drivers do not follow lane disciplines. Eleven-hour video recording was collected from a busy pedestrian crossing at a midblock location in India and processed using artificial intelligence techniques. Vehicle-pedestrian interactions are characterised by two conflict measures (i.e., post encroachment time and gap time) and modelled using block maxima and peak over threshold approaches. To handle the non-stationarity of pedestrian conflict extremes, several explanatory variables are included in the models, which are estimated using the maximum likelihood estimation procedure. Modelling results indicate that the EVT models provide reasonable estimates of historical crash records at the study location. From the EVT models, a few key insights related to vehicle–pedestrian interactions are as follows. Firstly, a comparison of EVT models shows that the peak over threshold model outperforms the block maxima model. Secondly, post encroachment time conflict measure is found to be appropriate for modelling vehicle–pedestrian interactions compared to gap time. Thirdly, pedestrian crash risk significantly increases when they interact with two-wheelers in contrast with interactions involving buses where the crash risk decreases. Fourthly, pedestrian crash risk decreases when they cross in groups compared to crossing individually. Finally, pedestrian crash risk is positively related to average vehicle speed, pedestrian speed, and five-minute post encroachment time counts less than 1.5 s. Further, different block sizes are tested for the block maxima model, and the five-minute block size yields the most accurate and precise pedestrian crash estimates. These findings demonstrate the applicability of extreme value analysis for heterogeneous and disordered traffic conditions, thereby facilitating proactive safety management in disordered and undisciplined lane conditions.
E-scooters aspire to provide flexibility to their users while covering the first/last mile of a multimodal trip. Yet, their dual travel behavior, i.e., utilizing both vehicles’ roadways and pedestrians’ sidewalks, creates new challenges to transport modelers. This study aims to model e-scooter riding behavior in comparison to traditional urban transport modes, namely car and walking. The new modeling approach is based on perceived safety that is influenced by the road environment and affects routing behavior. An ordinal logistic model of perceived safety is applied to classify road links in a 7-point Likert scale. The parametric utility function combines only three basic parameters: time, cost, and perceived safety. First/last mile routing choices are modeled in a test road network developed in Athens, Greece, utilizing the shortest-path algorithm. The proposed modeling approach proved to be useful, as the road environment of an urban area is heterogenous in terms of safety perceptions. Indeed, the model outputs show that the flexibility of e-scooters is limited in practice by their low-perceived safety. To avoid unsafe road environments where motorized traffic dominates, e-scooter riders tend to detour. This decision-making process tool can identify road network discontinuities. Nevertheless, their significance regarding routing behavior should be further discussed.
Virtual reality (VR) technology emerges as a promising tool for investigating human perception and behavior in highly controlled, immersive, and risk-free environments. This study proposed to apply simulated VR technology to investigate the interactions between perceived crash risk and behavior patterns in a road crossing with changes in the safety-related environmental attributes. In the context of the 8-meter-wide segment in a residential block, 35 VR environments with variations of six environmental attributes were generated. Two hundred participants were recruited for the experiment. The measured behavioral outcomes were 1) waiting and reaction time in the decision phase before crossing and 2) crossing speed and gait variability in the crossing phase. Random effect regression and multi-level structural equation models were constructed to test the study hypotheses. The results demonstrated that environmental attributes, including barriers to visibility (coefficient = 0.446), geometric patterns (coefficient = −0.625), and pavement signs (coefficient = −0.502), were associated with the pedestrians’ perceived risk, but the influence varied by street types. In addition, changes in the perceived threats to pedestrians were found to mediate the environment-crossing behavior relationship (coefficient of the indirect effect = 0.679). Those who perceive higher crash risk took longer to decide to start walking at a crosswalk and tended to walk in haste while crossing the road. Using VR technology, the present study addressed an inter-relationship between environmental characteristics, cognition, and crossing behavior, contributing to better knowledge on road safety interventions to reduce the risk of pedestrian-involved crashes.
In recent years, significant research has focused on traffic safety evaluations at unsignalized intersections due to complex and heterogeneous traffic movements as well as driver behaviour at such locations. However, at unsignalized T-intersections, priority traffic rules are less respected, which creates more conflicts. Further, multiple traffic movements such as right turns and through movements with varied driver behaviour results in increases the severity of conflicts. Many research studies have focused on the proactive safety measures in traffic safety evaluations as compared to crash-based analysis. Also, it is observed that Time to Collision (TTC) and Post Encroachment Time (PET) are the predominant types of surrogate safety measures in traffic safety evaluations. From the existing research outcomes, it is understood that these surrogate safety measures may give a better understanding of chain events for crash occurrences, collision mechanisms, and resulting consequences. However, further research is required in order to understand the suitability of such surrogate safety measures based on the complexity of heterogeneous traffic as well as driver behaviour with considerations of turning vehicles, particularly at T-intersections. In this context, this paper critically reviews the recent developments in Surrogate Safety Measures (SSM) and their applications at unsignalized intersections, with a particular focus on the T-intersection. This paper also brings attention to T-intersection safety evaluation with SSM in a developing country context. The outcome of the present study is more useful in the evaluation of traffic safety at T-intersections and suitable safety indicators for the evaluation.
Road accidents are one of the leading fatality rates around the globe. Even with advancements in technologies, still, the reduction in accidents is minimal. This review intends to evaluate the studies on automated prevention for road accidents. In this manner, the PRISMA statement was executed, and data extraction was made from the Scopus database. By initial screening, 75 articles were extracted, while performing the detailed screening only 22 left for final assessment. The keywords combinations were "Automation" AND "Road Accidents" with year limitation from 2011 to mid-2021. The extracted articles were focusing on automation aiding drivers, automation in automated vehicles, automation in crash reduction and automation aiding pedestrians, cyclists and mixed vehicles. The adaptation of these automation techniques in road safety can reduce the fatality rate, especially in developing countries where the rate is high and requires a major transformation in infrastructure.
Designing the crosswalks at signalized intersections aims to facilitate pedestrians' safe crossing by preventing conflicts with motorized vehicles. However, pedestrian safety remains at high risk due to the signal violations by pedestrians at the intersections in India. The present study examines the violation of pedestrians at twelve signalized intersections in India. The study results show that approximately 44.6% of pedestrians arriving during the red phase violate the signals. The violation prevalence is high in the latter half of the red-light period. With the help of a multiple linear regression model, the relationship between various explanatory variables (i.e., individual characteristics and traffic conditions) and the violation prevalence has been established. Further, the pedestrian-vehicle conflicts during signal violation have been classified based on the yielding behaviour of pedestrians and road users. The yielding behaviour of the road users is found to vary in the beginning, middle, and end of the red phase. The study utilized the conflict indicator ‘Post Encroachment Time’ (PET) to determine the proximity of conflicts during violation. The severity levels of the conflict were determined with the help of the Swedish Traffic Conflict Technique (STCT) using ‘Time to Accident’ (TA) and ‘Conflicting Speed’ (CS). A generalized ordered logit model has been utilized to identify the factors (i.e., traffic conditions, personal characteristics, and situational characteristics) that affect the severity level of pedestrian-vehicle conflicts. The results highlight the need for engineering countermeasures, education, and enforcement to deal with the non-compliance behaviour of pedestrians.
Traffic simulation models have been utilized recently in safety evaluations by calculating traffic conflict indicators from simulated road user trajectories. However, existing simulation models (1) do not accurately capture road users’ behaviour and evasive actions, and (2) do not consider road users’ intelligence and rationality. This study proposes a safety-oriented microsimulation framework for modeling conflict interactions between motorcyclists and pedestrians in mixed traffic conditions. Motorcyclists are modeled as utility-based rational and intelligent decision-makers using a Markov Decision Process (MDP). Continuous Inverse Reinforcement Learning (IRL) is proposed to recover the motorcyclists’ reward function using their actual trajectories. The recovered motorcyclists’ reward function provides inferences into their behavior in conflict interactions. The motorcyclists’ optimal policies (sequences of decisions) are estimated using the Actor-Critic Deep Reinforcement Learning. The results show that the model simulated motorcyclist trajectories and the evasive actions with high accuracy. Moreover, the predicted PETs correlated well with corresponding field-measured conflicts.
In developing nations, road traffic crashes involving pedestrians have become a foremost worry. Presently, most of the road safety assessment projects and selection of interventions are still restricted to traditional methods that depend on historical crash data. However, in low and middle-income countries such as India, the availability, reliability, and accuracy of crash data are uncertain. Alternatively, Post Encroachment Time (PET) has added attention as a proximal indicator to examine pedestrian-vehicular potential crashes and address pedestrian risk under mixed traffic conditions. Hence, it will be meaningful to examine if the PET is a good substitute for pedestrian-vehicular crashes and if so, what built environment and pedestrian-level factors influence PET. In this background, the present study establishes a mathematical association between the average PET value of the urban road network level and actual crashes. Afterward, multiple linear regression models are developed to study the impact of the built environment, traffic parameters, and pedestrian-level attributes on PET. The outcomes indicate that vehicle speed, lack of enforcement, absence of traffic signal (for traffic as well as pedestrians), land use type, slum population, inadequate sight distance, pedestrian’s state of crossing, and pedestrian’s risky crossing behaviour substantially affect the average PET at road network-level.
Relative to safety assessment using data from observed crashes, conflict-based road safety assessment can potentially provide additional insights into crash causation processes. Despite numerous review studies on this topic, the application context of conflict measures has been generally overlooked. This study conducts a systematic review of conflict-based safety measures with a specific focus on the context of their applications. This study employs the Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (PRISMA) guidelines of systematic review and meta-analysis to review conflict measures used for the safety assessment of intersections over the past ten years (2010-19). A total of 386 studies are systematically reviewed to identify conflict measures used for various contexts, including intersection types, traffic operating conditions, study types, and the purpose of the study.
The systematic review indicates that temporal proximity measures, specifically time-to-collision and post-encroachment time, are the most widely used conflict measures regardless of the application context. Other families of conflict measures such as spatial proximity, kinematic, mixed and combinations of measures have also been applied depending on the context. Using the extracted data from relevant studies, linear regression models were developed for time-to-collision and post-encroachment time thresholds at signalized intersections and time-to-collision thresholds at unsignalized intersections. The thresholds are found to be associated with traffic environment types, sources of conflict data and the application purpose of conflict measures. The findings of this study identify several critical gaps in the literature that can help guide future research directions in the conflict-based safety assessment of transport facilities. Critical gaps include the scarcity of validation studies for conflict measures, the lack of suitable techniques to estimate crash risk by severity types, the primary focus on signalized intersections (leaving studies of other facility types underrepresented), and the lack of suitable conflict measures for vulnerable road users.
Shared spaces are gaining popularity worldwide alongside the promotion of walking and cycling. Pedestrians, conventional bicycles, and electric bicycles (e-bikes) currently coexist in many shared spaces, with the prevalence of e-bikes increasing gradually in recent years. Shared spaces can provide a lower-stress experience for users because they are separated from motorized traffic, but frequent interactions among users raise safety concerns. This study sought to investigate conflict behaviours and characteristics among pedestrians, conventional bicycles, and e-bikes in shared spaces. Video data covering 12 h from three locations in Shenzhen city was analyzed. A total of 1748 pedestrians, 1748 conventional bicycles, and 930 e-bikes were observed, while 337 traffic conflicts were identified using the Dutch Objective Conflict Technique for Operation and Research (DOCTOR) method. Shared-space crash data recorded between April of 2013 and September of 2019 was used to validate and complement the conflict analysis. Friedman test was used to compare the conflict behaviours and characteristics among different groups. The analysis showed a positive relationship between the traffic volume and the number of conflicts but an inverse relationship between the traffic ratios and conflict ratios. Evasive actions, including swerving, decelerating, accelerating, and their combinations, were analyzed for various conflict types and severities, with swerving found to be the most common, especially in slight conflicts. Compared with slight conflicts, conventional bicycles and e-bikes exhibited low-speed characteristics in serious conflicts. These results indicated that high traffic volume and traffic complexity are the main factors that affect conflicts. It was proposed that the conflict coefficient be used to measure shared-space safety. In addition, the video observation study and crash data analysis suggested that conflicts and crashes between e-bikes and pedestrians are high-occurrence events, and that pedestrians are often exposed to higher injury risks.
The assessment of pedestrian safety is often conducted in a reactive way by analyzing pedestrian crash data. However, in a developing country, the availability of reliable crash data is a major challenge. Therefore, without relying solely on reactive approach, it is essential to combine some methods that can proactively assess pedestrian safety. In this background, the present study proposed a methodology combining both reactive and proactive approaches to assess pedestrian safety at urban intersections in Kolkata, India. The method developed in the present study utilizes a combination of the historical crash data analysis, the analysis of pedestrian-vehicular conflict (i.e., pedestrian-vehicular post-encroachment time), along with pedestrians’ risk perception toward the built environment and traffic parameters, to identify the key factors influencing pedestrian safety and to recognize potential risk-prone intersections in Kolkata.
Based on the combined reactive and proactive assessments, there is evidence that high vehicle volume, pedestrian-vehicular volume ratio, absence of traffic signal, inaccessibility of the pedestrian crosswalk, absence of police personnel, higher approaching speed, commercial area, inadequate sight distance, presence of slum population, and higher population density near the intersection significantly increase the risk of pedestrian crashes. Finally, using this combined proactive-reactive approach, the present study also identifies and ranks twenty-five high risk-prone intersections for pedestrians. This is a significant step toward scientific decision making and allowing the use of information beyond historical crash records.
Shared space is a concept of urban street design which implies the creation of a level surface within the whole road reserve and is aimed at encouraging different road users to interact spontaneously and to negotiate priority with each other. To build successful shared spaces, traffic engineers can rely at present on specific guidelines as well as technical reports. Nevertheless, there is no method available to compute the performance of shared spaces in terms of Level Of Service (LOS).
In order to address this gap, a new indicator of traffic quality for pedestrians is being developed. This measure of performance considers aspects of comfort related to the crossing, which pedestrians use to go from one side of the roadway to the other. During this movement, discomfort is generated by the necessity to solve the conflicts with vehicles. Therefore, factors which potentially influence comfort are mathematically formulated. Later, the performance indicator can be calibrated on the basis of the opinion of a group of respondents, who evaluated real-world crossing movements in video sequences. The effectiveness and usability of the developed indicator is demonstrated in an exemplary case study. A shared street in the district of Bergedorf, Hamburg (D) is selected and filmed. To reproduce the interaction of road users and the mechanism of space negotiation, an innovative modeling approach based on social force model (SFM) is proposed. The model is calibrated and implemented in a Java-based simulation tool. Alternative shared space scenarios, as well as conventional ones with space segregation, are simulated.
The goal of this dissertation is to establish a method to evaluate the performances of shared spaces through traffic microsimulation. This method includes the data survey and acquisition, the definition of performance indicators, the development of a microsimulation approach, the calibration of the motion model on the basis of real-world data and finally the execution of simulations to collect the results. In addition, this work shows the necessity to employ a comfort-based indicator for pedestrian traffic quality in shared spaces. The benefits of this approach, with respect to conventional efficiency-based indicators as time delay, is properly shown in real-world situations and successively demonstrated by help of statistical methods.
Limitations associated with traditional collision-based safety analysis techniques led to a growing interest in the use of surrogate safety mea-sures such as the traffic conflict technique. This interest was facilitated by advances in automated video-based data collection methods that helped to overcome the reliability issues associated with manual collection of data on traffic conflicts. Various objective conflict indicators that mea-sure various spatial and temporal aspects of user proximity are available to measure the severity of traffic events. These time-proximity conflict measures assume that proximity is a surrogate for conflict severity. How-ever, this assumption may not be valid in many driving environments. The objective of this paper is to investigate whether time-proximity con-flict measures can be a good indicator of safety in less-organized traffic environments with highly mixed road users. A case study of motorcycle conflicts in a highly congested shared intersection in Shanghai, China, was used as a case study. Traffic conflicts were analyzed with the use of automated video-based analysis techniques. Several traffic conflict indi-cators designed to detect evasive actions, such as deceleration, jerk, and yaw rate, were recommended as better able to measure traffic conflicts in such traffic environments. The results showed that indicators that measured evasive actions had higher potential to identify motorcycle conflicts in highly mixed, less-organized traffic environments than did time-proximity measures such as the time to collision.
A better understanding of pedestrian behavior is needed to enhance existing pedestrian simulation models. A study was conducted on microscopic pedestrian behavior during several interactions, with pedestrian walking speed and gait parameters (step frequency and length) as variables. Pedestrian trajectories at a signalized intersection in Vancouver, British Columbia, Canada, were extracted from video recordings by means of computer vision techniques. Walking speed and gait parameters were estimated by analyzing pedestrian speed profiles. The study provided detailed analysis of seven interactions. The variations in walking speed and gait parameter values across group size and gender during the seven interactions were also investigated. Results showed that, for some of the studied interactions, pedestrian speed alone may not be adequate to describe pedestrian behavior and that gait parameters can help to provide better understanding of pedestrian behavior in these particular interactions. Furthermore, a specific set of parameters was identified that can be extracted from the results and can be used to calibrate a microscopic pedestrian behavior-modeling platform currently being developed.
This study investigates the feasibility of using the spatiotemporal parameters of gait step frequency and step length-as cues for classifying pedestrians according to their gender and age. The gait parameters are automatically extracted from the pedestrian walking speed profile. Computer vision techniques are used for the automatic detection and tracking of pedestrians in an open (uncontrolled) environment. The classification is undertaken by using a simple k nearest neighbor algorithm. For demonstration, two case studies are used: Vancouver, British Columbia, Canada, and Oakland, California. For gender, correct classification rates of 78% and 81% were achieved for the Vancouver and Oakland case studies, respectively. Gender classification for the Vancouver case study considered pedestrians walking alone or in groups, and the Oakland case study gender classification considered only pedestrians walking alone. Pedestrian age classification resulted in a correct classification rate of 86% for the Oakland case study. Another classification measure, the kappa statistic, showed that the classification results were statistically significant beyond what is expected by chance. The method has the advantages of relying only on the pedestrian speed profile and using a simple classification algorithm.
A traffic conflict is usually composed of a chain of events in which at
least one of the involved road users performs some sort of evasive action
to avoid a collision. An evasive action usually involves powerful braking,
which leads to sudden, negative change in acceleration (deceleration). The
temporal dynamics (variation over time) of the acceleration of a vehicle is
represented by the jerk profile. More formally, jerk is the derivative of the
acceleration. In the case of an evasive action by braking, the jerk profile is
characterized by strong, negative values. This study examined two issues
in the quest to understand the benefits of evasive action analysis. The first
issue was whether jerk profiles can be used to identify critical traffic events
(conflicts). The second issue addressed the validity of the assumption that
the deceleration profile is inadequate as a stand-alone measure for conflict
identification. Automated video analysis was used to collect traffic data
and analysis was applied on two data sets with distinct traffic patterns. The
study revealed a significant difference between the jerk behavior of the
groups of drivers involved in conflictive and normal traffic interactions.
It also showed instances in which automated jerk evaluation was successful
in finding conflicts undetected by conventional conflict indicators. The
same could not be demonstrated for the road users’ deceleration behavior.
These findings support earlier studies on the shortcomings of the use of
deceleration data for conflict identification.
Overview Cultural paradigms determine both how we view road safety and the actions we take to improve it. While there may be many different competing paradigms for dominance, one paradigm can be seen as dominant. The dominant paradigm has changed a number of times over the century of motorization. The Finnish scholar, Valde Mikkonen (1997) has characterized conceptual development in road safety as a slow evolution punctuated by brief revolutions, which he sees as leading to a new dominant paradigm. The 1960s paradigm shift A paradigm shift of earthquake proportions that took place in the 1960s is responsible for the current high level of road safety in U.S. and Canada. A number of developments prepared the way for a major shift in thinking and action. Roadway development had made great progress, with the birth of the Interstate system. Cars had made technical advances in style and comfort, and especially engine performance; the mid-60s muscle cars still stand among the fastest production cars. The straight-line acceleration of these cars was not, however, well balanced with overall engineering sophistication, especially in terms of occupant protection. Also in the 1960s, emerging fuel economy concerns led to new lines of compact cars, which were even less safe. The 1960s paradigm shift was also facilitated by some basic theoretical and scientific developments following World War II. A number of new and improved intellectual tools contributed to the shift, including systems analysis, decision theory, and epidemiology. The human factors engineering that developed out of military aviation during the war and the innovative biomechanics work of Hugh DeHaven and Col. John Stapp were also critical.
Improving pedestrian safety at intersections remains a critical issue. Although several types of safety countermeasures, such as reforming intersection layouts, have been implemented, methods have not yet been established to quantitatively evaluate the effects of these countermeasures before installation. One of the main issues in pedestrian safety is conflicts with turning vehicles. This study aims to develop an integrated model to represent the variations in the maneuvers of left-turners (left-hand traffic) at signalized intersections that dynamically considers the vehicle reaction to intersection geometry and crossing pedestrians. The proposed method consists of four empirically developed stochastic sub-models, including a path model, free-flow speed profile model, lag/gap acceptance model, and stopping/clearing speed profile model. Since safety assessment is the main objective driving the development of the proposed model, this study uses post-encroachment time (PET) and vehicle speed at the crosswalk as validation parameters. Preliminary validation results obtained by Monte Carlo simulation show that the proposed integrated model can realistically represent the variations in vehicle maneuvers as well as the distribution of PET and vehicle speeds at the crosswalk.
Various indicators of objective conflict have been proposed in the literature to measure the severity of traffic events. Objective conflict indicators measure various spatial and temporal aspects of proximity on the premise that proximity is a surrogate for severity. These aspects of severity may be partially overlapping and in some cases independent. Two sets of conflict indicators were used in a study conducted to demonstrate that integration of the severity cues provided by each conflict indicator could be performed to reflect better the true, yet unobservable, severity of traffic events. The first set of conflict indicators required the presence of a collision course common to the interacting road users. The second set measured severity in mere temporal proximity between road users. The study proposes a methodology with which to aggregate the event-level measurements of conflict indicators into a safety index. First, individual conflict indicator measurements are mapped into severity intervals [0, 1]. Second, these severity indices are aggregated to a safety index that includes both individual severities and exposure. The methodology is applied on individual measurements of pedestrian-vehicle conflicts.
Pedestrians are vulnerable road risers, and despite their limited representation in traffic events, pedestrian-involved injuries and fatalities are overrepresented in traffic collisions. However, little is known about pedestrian exposure to the risk of collision, especially when compared with the amount of knowledge available for motorized traffic. More data and analysis are therefore required to understand the processes that involve pedestrians in collisions. Collision statistics alone are inadequate for the study of pedestrian-vehicle collisions because of data quantity and quality issues. Surrogate safety measures, as provided by the collection and study of traffic conflicts, were developed as a proactive complementary approach to offer more in-depth safety analysis. However, high costs and reliability issues have inhibited the extensive application of traffic conflict analysis. An automated video analysis system is presented that can (a) detect and track road users in a traffic scene and classify them as pedestrians or motorized road users, (b) identify important events that may lead to collisions, and (c) calculate several severity conflict indicators. The system seeks to classify important events and conflicts automatically but can also be used to summarize large amounts of data that can be further reviewed by safety experts. The functionality of the system is demonstrated on a video data set collected over 2 days at an intersection in downtown Vancouver, British Columbia, Canada. Four conflict indicators are automatically computed for all pedestrian-vehicle events and provide detailed insight into the conflict process. Simple detection rules on the indicators are tested to classify traffic events. This study is unique in its attempt to extract conflict indicators from video sequences in a fully automated way.
Introduced the statistic kappa to measure nominal scale agreement between a fixed pair of raters. Kappa was generalized to the case where each of a sample of 30 patients was rated on a nominal scale by the same number of psychiatrist raters (n = 6), but where the raters rating 1 s were not necessarily the same as those rating another. Large sample standard errors were derived.
This paper reports the results of a study that explored the relationship between fatal crashes and conflict rates at mid-block on 14 locations in Delhi, India. All locations had a mix of motorized and non-motorized traffic. The sites were selected to represent low, medium and high fatality rates. The analysis was done in two stages. The first stage used recent 3-year statistics along the entire street. The second stage focused on each fatal crash for mid-block segments on those streets. Peak-hour traffic at 14 selected locations was videotaped. Trained observers recorded traffic compositions at mid-block, average space mean speeds by mode and conflicts by type, reactor mode and cause mode. After converting raw conflict counts to rates, site ranking went from high to low conflict rate sites. The studies showed a weak crash-conflict association. Conflict data for various sites were compared for different combinations of conflicts. The comparison revealed that the presence of only a few non-motorized modes is enough to cause conflicts between motorized vehicles and on-road non-motorized vehicles. The study did not provide a conclusive relationship between mid-block conflicts and fatal crash sites. However, the conflict study provided useful insights into the interaction between different traffic entities in the traffic streams of 14 sites. An important conclusion of this study is that a traffic-planning emphasis on studying conflict rates may not result in reducing fatality rates on urban roads along mid-block segments.
The aim of this research was to describe age-related changes in locomotor adjustments during obstructed gait and expand and build from the current body of literature describing single obstacle avoidance strategies by including trials in which the subjects stepped over two identical obstacles placed in series. We observed young adults (YA: N = 8; aged 23.1 +/- 2.0 years) and older adults (OA: N = 8; aged 76.1 +/- 4.3 years) as they walked along a 5 m long instrumented pathway (GAITRite) and stepped over one or two obstacles that were scaled to their lower leg length. Infrared markers, tracked using the Optotrak motion analysis system (60 Hz; Northern Digital Inc, Canada), were fixed to subjects' trunk and feet, and several anatomical landmarks were digitized for each segment (e.g. toes). Data analyses included lead and trail toe clearance values, take-off and landing distance, step time, length, width and velocity, and three-dimensional trunk angles. Both age groups were able to successfully complete the obstacle avoidance task, and the presence of a second obstacle did not affect clearance strategies of either OA or YA. OA crossed the obstacles with a reduced step velocity and stepped closer to the trailing edge, although take-off distances were not different between the age groups. Additionally, OA used similar ranges of trunk motion as YA when crossing the obstacle, but did so while using smaller step lengths and step widths compared to YA, effectively, using a narrower base of support. Together, these findings suggest that older adults adopted a more cautious crossing strategy in that they reduced their crossing step velocity. However, other aspects of the avoidance strategy used by the older adults, specifically the shortened landing distances and the use of similar ranges of trunk motion within a narrowed BOS, could potentially put them at risk for tripping or imbalance when stepping over an obstacle.
Intelligent Transportation Systems need methods to automatically monitor the road traffic, and especially track vehicles. Most research has concentrated on highways. Traffic in intersections is more variable, with multiple entrance and exit regions. This paper describes an extension to intersections of the feature-tracking algorithm described in [1]. Vehicle features are rarely tracked from their entrance in the field of view to their exit. Our algorithm can accommodate the problem caused by the disruption of feature tracks. It is evaluated on video sequences recorded on four different intersections.
The factorization method described in this series of reports requires an algorithm to track the motion of features in an image stream. Given the small inter-frame displacement made possible by the factorization approach, the best tracking method turns out to be the one proposed by Lucas and Kanade in 1981. The method defines the measure of match between fixed-size feature windows in the past and current frame as the sum of squared intensity differences over the windows. The displacement is then defined as the one that minimizes this sum. For small motions, a linearization of the image intensities leads to a Newton-Raphson style minimization. In this report, after rederiving the method in a physically intuitive way, we answer the crucial question of how to choose the feature windows that are best suited for tracking. Our selection criterion is based directly on the definition of the tracking algorithm, and expresses how well a feature can be tracked. As a result, the criterion is optima...
Automated video analysis is advocated as a robust tool for traffic data collection. A challenging task that is often associated with automated methods is the interpretation of the traffic scene. In particular, the recognition of the road-users types is necessary to learn traffic scenarios and understand behaviour patterns within each road-user class. The objective of this paper is to present and evaluate a generic road-users classification procedure. The classification relies on the motion pattern attributes associated with the trajectories of each road-user type, namely, vehicles, pedestrians and cyclists. A novel approach for features selection is proposed where singular spectrum analysis identifies the basic harmonics (speed variation patterns) characterizing the movements behaviour. Constrained based classification is then applied on the selected features to categorize the road-users. Performance evaluation of the proposed classification is presented. Validation of the procedure is undertaken using real world data sets collected at a conventional four-legged intersection in the City of Surrey, British Columbia. Satisfactory results were demonstrated and evaluated through several performance measures. The main benefit of this research is to apply classification as a first step in the activity and behaviour recognition of road-users in traffic scenes.
Following initial investigation of the G.M. (brakelight) conflict recording technique during 1972/73 in various Canadian cities, it was concluded that this procedure — while relatively easy to apply had some very fundamental conceptual drawbacks as an indicator of expected accident rate.
This paper addresses the automated analysis of pedestrians' conformance behavior and data collection in an intersection with a perceived high rate of traffic conflicts involving pedestrians. The intersection is located in the Downtown Eastside of Vancouver, British Columbia, Canada. Despite implementation of countermeasures such as reducing the speed limit, safety issues linger at the intersection. The intersection is characterized by a high pedestrian volume and considerable crossing violations by pedestrians, which elevate the safety risk and disrupt vehicle traffic flow. A challenge in performing pedestrian road safety analysis is the shortfall of reliable data. Recent advances in automated detection of pedestrians through the use of computer vision expanded the range of applications in traffic safety. In this study, an automated system for identifying pedestrian crossing nonconformance to traffic regulations by using pattern matching was developed and tested. The results show satisfactory accuracy in detecting both spatial and temporal violations, with the detection rate for violations being more than 84% correct. The automated collection of pedestrian data on crossing speed and counts is also demonstrated with high accuracy. The availability of these data is important for diagnosing the safety issues at the intersection and for justifying capital for implementing pedestrian safety countermeasures.
Changes in pedestrian-vehicle traffic conflicts in urban streets redesigned according to the principles of shared space were investigated with a recently developed pedestrian vehicle conflict analysis (PVCA) method. In a first step, the PVCA method was revised to reflect more accurately the features of shared space; this revision included the definition of a systematic process for identifying conflict occurrences on the one hand and the full quantification of the conflict severity grading process on the other. Then the refined PVCA method was applied to a case study in London with video data from periods before and after the redevelopment of the Exhibition Road site from a conventional dual carriageway to a modern design with some elements of shared space. The results of the comparative analysis indicated a general decrease in traffic conflict rates as a result of the redesign but also highlighted specific issues that may require additional analysis.
An emphasis on active modes of transportation, that is, walking and cycling, has recently been renewed amid concerns for the environment and public health. However, the focus of research and practice that these modes have traditionally received is secondary to that received by motorized modes. As a consequence, the data on pedestrians (in particular, microscopic data) required for analysis and modeling are lacking. For instance, accurate data on the length of individual stride are not available in the transportation literature. This paper proposes a simple method to extract frequency and length of pedestrian stride automatically from video data collected nonintrusively in outdoor urban environments. The walking speed of a pedestrian oscillates during each stride; the oscillation can be identified through the frequency analysis of the speed signal. The method was validated with real-world data collected in Rouen, France, and Vancouver, Canada, where the root mean square errors for stride length were 6.1 and 5.7 cm, respectively. A method to distinguish pedestrians from motorized vehicles is proposed and used to analyze the 50 min of the Rouen data set to provide the distributions of stride frequency and length.
The potential for using computer vision techniques to solve several shortcomings associated with traditional road safety and behavior analysis is demonstrated. Surrogate data such as traffic conflicts provide invaluable information that can be used to understand collision-contributing factors and the collision failure mechanism better. Recent advances in computer vision techniques have encouraged the use of proactive safety surrogate measures such as detection of conflicts and violations. The objective of this study is to demonstrate the automated safety diagnosis of pedestrian crossing safety issues by using computer vision techniques. The automated safety diagnosis is applied at a major signalized intersection in downtown Vancouver, British Columbia, Canada, at which concerns had been raised regarding the high conflict rate between vehicles and pedestrians as well as the elevated number of traffic violations (i.e., jaywalking). This study is unique in its attempt to extract conflict indicators and detect violations from video sequences in a fully automated way. This line of research benefits safety experts because it provides a prompt and objective safety evaluation for intersections. The research also provides a permanent database for traffic information that can be beneficial for a sound safety diagnosis as well as for developing safety countermeasures.
The detection and understanding of nonconforming behavior (violations) can be useful in forming safety diagnoses and developing safety countermeasures. Traffic violations occur when road users, including pedestrians, seek increased mobility and disregard traffic laws and regulations. Such behavior can cause additional collision risks. This paper's objective is to demonstrate the automated identification of pedestrian crossing violations with computer vision techniques. Two types of violations are considered. The first is spatial violations: pedestrians cross an intersection in nondesignated crossing regions. The second is temporal violations: pedestrians cross an intersection during an improper signal phase. The methodology primarily relies on the identification of road users' trajectories and separating pedestrians with nonconforming behavior from those with conforming behavior. The methodology is demonstrated on two urban intersections, one in downtown Vancouver, Canada, the other in Kuwait City, Kuwait. The results show satisfactory accuracy in the detection of spatial and temporal violations, with an approximately 90% correct violation detection rate having been achieved in both case studies.
This study investigates changes in pedestrian speed behavior following the implementation of a scramble phase. The aim is to demonstrate the feasibility of automatic collection of pedestrian data; and to study pedestrian speed variations with respect to design changes to intersection crossings. The results show that the average crossing speed is higher after the implementation of the scramble phase. Within the scramble, the average crossing speed is higher for diagonal crossing than side crosswalk crossing. The average crossing speed is lower for pedestrians crossing during the Walk interval. Pedestrians have higher speed through the first half of the crosswalk.
New urban planning concepts are being redefined to emphasize walkability (a measure of how walker-friendly an area is) and to accommodate the pedestrian as a key road user. However, the availability of reliable information on pedestrian traffic remains a major challenge and inhibits a better understanding of many pedestrian issues. Therefore, the importance of developing new techniques for the collection of pedestrian data cannot be overstated. This paper describes the use of computer vision techniques for the automated collection of pedestrian data through several applications, including measurement of pedestrian counts, tracking, and walking speeds. An efficient pedestrian-tracking algorithm, the MMTrack, was used. The algorithm employed a large-margin learning criterion to combine different sources of information effectively. The applications were demonstrated with a real-world data set from Vancouver, British Columbia, Canada. The data set included 1,135 pedestrian tracks. Manual counts and tracking were performed to validate the results of the automated data collection. The results show a 5% average error in counting, which is considered reliable. The results of walking speed validation showed excellent agreement between manual and automated walking speed values (root mean square error = 0.0416 m/s, R-2 = .9269). Further analysis was conducted on the mean walking speed of pedestrians as it related to several factors. Gender, age, and the group size were found to influence the pedestrian mean walking speed significantly. The results demonstrate that computer vision techniques have the potential to collect microscopic data on road users at a degree of automation and accuracy that cannot be feasibly achieved by manual or semiautomated techniques.
In sustainable urban planning, non-motorised active modes of travel such as walking are identified as a leading driver for a healthy, liveable and resource-efficient environment. To encourage walking, there is a need for a solid understanding of pedestrian walking behaviour. This understanding is central to the evaluation of measures of walking conditions such as comfortability and efficiency. The main purpose of this study is to gain an in-depth understanding of pedestrian walking behaviour through the investigation of the spatio-temporal gait parameters (step length and step frequency). This microscopic-level analysis provides insight into the pedestrian walking mechanisms and the effect of various attributes such as gender and age. This analysis relies on automated video-based data collection using computer vision techniques. The step frequency and step length are estimated based on oscillatory patterns in the walking speed profile. The study uses real-world video data collected in downtown Vancouver, BC. The results show that the gait parameters are influenced by factors such as crosswalk grade, pedestrian gender, age and group size. The step length was found to generally have more influence on walking speed than step frequency. It was also found that, compared to males, females increase their step frequency to increase their walking speed.
This paper describes the application of the traffic conflict technique to estimate, traffic safety at intersections. Using data collected from 94 conflict surveys, traffic conflict frequency and severity standards for signalized and unsignalized intersections have been established. These standards allow for the relative comparison of the conflict risk at various intersections. An Intersection Conflict Index (ICI) measure was developed to summarize conflict risk and provide an indication regarding the relative risk of being involved in a conflict at an intersection. In addition, regression analysis was used to develop predictive models which relate the number of traffic conflicts to traffic volume and accidents. The regression analysis results indicate that: (i) the average hourly conflict rate (AHC) and the average hourly severe conflict rate (AHC 4+) correlated reasonably well with traffic volume for both signalized and unsignalized intersections, and (ii) strong relationships between accidents and conflicts were obtained for signalized intersections only. These research efforts are expected to further enhance the usefulness of the traffic conflict technique as a tool to evaluate the safety of intersections. Finally, a case study is presented as an example of the usefulness of traffic conflict analysis.
To quantitatively evaluate the safety of pedestrian crossing and to quantify the risk degree of traffic conflicts between crossing pedestrian and vehicle at signalized intersection, this paper conducted the risk degree model of crossing pedestrian and vehicle based on the features of pedestrian crossing behavior and the theory of conflict. The SMOK clustering algorithm based on the fuzzy cluster analysis method is used. Evaluation analysis process and results of risk degree model is demonstrated in the case studies of six signalized intersections in Nanjing City of China with the manual data collection method on video, which gives acquisition and calculation method of the evaluation indexes of TTC, PET, DST and validates the practicality of proposed model. The result shows that three indexes (TTC, PET, DST) play a significant role in clustering analysis and the frequency and severity of conflicts differs according to different types of conflict.
New urban planning concepts are being redefined to emphasize active modes of transportation such as walking and cycling. In this environment, a disparate mix of road users shares the same road. In order to address road-safety and level of service requirements for the different types of road-users, traffic information needs to be collected with high degree accuracy. The reliability and accuracy of the data collected can significantly affect the quality of analysis. In recent years, automated video analysis was established as a robust tool for data collection. Road-user trajectories obtained through automated computer vision are rich in information. They hold features that reveal the structure of the traffic scene and provide clues to the movement characteristics of the road-users. The objective of this paper is to present and evaluate a road-user classification procedure. The classification relies on the motion pattern attributes associated with the trajectories of each road-user type, namely, vehicles, pedestrians and cyclists. A novel approach for features selection is proposed where singular spectrum analysis (SSA) identifies the basic harmonics (speed variation patterns) characterizing the movement behaviors. Constrained based classification and spectral clustering are then applied on the selected features to categorize the road-users. Several case studies are used for the evaluation of the proposed classification. The case studies use real world data sets collected at a roundabout and a conventional four-legged intersection in Greater Vancouver, British Columbia. Very promising results were demonstrated and evaluated through several performance measures. The main benefit of this research is to apply classification as a first step in the activity and behavior recognition of road-users in traffic scenes. The proposed approach is useful as a complement to vision based classification, where road-users classes can be difficult due to partial occlusion or when the extraction of physical characteristics (descriptive features) of road-users does not provide enough details to reveal the correct type of a road-user.
This paper demonstrates an automated safety diagnosis approach for evaluating vehicle–bicycle conflicts using video analysis. The use of traffic conflicts for safety diagnosis is gaining acceptance as a surrogate for collision data analysis. Traffic conflicts can provide insight into the failure mechanism that leads to road collisions and do not require long observation periods. In the approach presented in this paper, traffic conflicts are automatically detected and their severity ranked using the Time to collision (TTC) safety indicator. As well, vehicle violations such as failure to respect yield signs are automatically identified. The safety diagnosis is also supplemented with automated classification and count of vehicles and bicycles. A case study is presented for diagnosing safety issues at a busy intersection downtown Vancouver, British Columbia. Vehicle–bicycle conflicts as well as vehicle rear-end and merging conflicts were identified and examined. The results showed a high exposure of cyclists to traffic conflicts and a significant driver non-compliance rate. Several countermeasures to mitigate the safety issues were presented and evaluated.
This paper describes the development and implementation of the conflict-based assessment of pedestrian safety (CAPS) methodology for the evaluation of pedestrian accessibility at complex intersections. Significant research has explored pedestrian access to modern roundabouts and other complex intersections, and a significant focus has been placed on accessibility for pedestrians who were blind. A majority of these studies relied on actual street crossings by study participants under the supervision of a trained orientation and mobility specialist. These crossing studies quantified risk from a measurement of intervention events, in which the orientation and mobility specialist had to physically stop the participant from crossing. Although such studies provide useful data on the crossing risk at a particular intersection, street crossings can be dangerous to the study participants and are time-consuming and expensive to conduct. The CAPS method emphasizes the use of conflict-based safety factors to quantify risk in a framework compatible with indicator studies. This method relates pedestrian crossing decisions to advanced measurements of vehicle dynamics to estimate lane-by-lane conflicts and identifies the grade of conflict on the basis of a five-criterion rating scale. The CAPS framework was applied to a study of crossings by blind pedestrians at a multilane roundabout. The resulting risk scores were calibrated from the actual orientation and mobility interventions observed during the study. The calibrated CAPS framework correctly matched all (high-risk) orientation and mobility intervention events and further identified other (lower-risk) pedestrian vehicle conflicts. The CAPS framework provides a more efficient, objective, and consistent safety assessment of pedestrian crossings in a research context, without the need for pedestrians to step into the roadway.
This study presents the results of a collision-based full Bayes (FB) before-after (BA) safety evaluation of a newly proposed design for channelized right-turn lanes. The design which is termed "Smart Channels" decreases the angle of the channelized right-turn to approximately 70°. Its implementation is usually advocated to afford drivers a better view of the traffic stream they are to merge with and to allow also for safer pedestrian crossing. The evaluation used data for three treatment intersections and several comparison sites in the city of Penticton, British Columbia. The evaluation utilized FB univariate and multivariate linear intervention models with multiple regression links representing time, treatment, and interaction effects as well as the traffic volumes effects. As well, the models were extended to incorporate random parameters to account for the correlation between sites within comparison-treatment pairs. The results showed that the implementation of the right-turn treatment has resulted in a considerable reduction in the severity and frequency of collisions. Another objective of the paper was to compare the results of the collision-based evaluation with the results of a traffic conflict-based evaluation of the same treatment intersections. The comparison showed remarkable similarity between the overall and the location specific reductions in conflicts and collisions which provides support for using traffic conflicts in BA studies. The results also provide positive empirical evidence that can support the validity of traffic conflict techniques.
The allocation of resources to build facilities amicable to pedestrians is governed by pedestrian activity at the location of interest. Collecting real-world data such as pedestrian counts at each point of interest is an expensive and time-consuming process. However, unlike trip generation models to estimate vehicle trips, the literature documents limited research to model and measure activity pertaining to pedestrian counts. The development and an assessment of models to measure pedestrian activity at signalized intersections are presented. Data collected at 176 signalized intersections in the city of Charlotte, North Carolina, are used to develop models to measure pedestrian activity by the time of day at signalized intersections. Pedestrian counts collected at the 176 intersections are used as a dependent variable. Factors such as demographic characteristics (population, household units), socioeconomic characteristics (income level, employment), land use characteristics (residential, commercial, industrial, etc.), network characteristics (number of approaches, number of lanes, speed limits, traffic volume, presence of medians), and the number of transit stops are extracted and estimated by using features available in a commercial geographic information system software program. These factors are used as independent variables. Multiple regression analysis through backward elimination of independent variables is used to develop the models. The developed models could be used by practitioners to measure pedestrian activity at a location if data are available. The measured pedestrian activity could be used in transportation planning, safety, and operational analyses.
Camera calibration is a necessary step in all video analysis applications to recover the real-world positions of concerned road users. Camera calibration can be performed based on feature correspondences between the real-world space and image space. In urban traffic scenes, the field of view may be too limited or video camera may not be accessible to allow reliable calibration based on vanishing points. A review of the current methods for camera calibration reveals little attention to these practical challenges that arise when studying urban intersections to support applications in traffic engineering. This study presents the development details of a robust camera calibration approach based on integrating a collection of geometric information found in urban traffic scenes in a consistent optimization framework. The developed approach was tested on six datasets obtained from urban intersections in British Columbia, California, and Kentucky. The results clearly demonstrated the robustness of the proposed approach.
One purpose of this paper is to evaluate the extent to which the pedestrian data collection efforts of transportation agencies in the U.S. are addressing pedestrian safety factors. A second purpose is to suggest how pedestrian data collection can be improved to facilitate the monitoring of these factors. Fifteen pedestrian safety issues are identified based on a literature review and examination of pedestrian-vehicle crashes in Utah. A year 2001 survey of U.S. transportation agencies indicated that 45 (75%) of the 60 respondents were counting pedestrians at various locations. Hand counting, the recording of pushbutton usage, and video cameras were being used to collect data. Automated systems, such as position sensors and image processing, were not being used to count pedestrians. The usage of advanced data collection technologies is not critical to the resolution of pedestrian safety concerns, although permanent counting installations might increase data collection efficiency. Only four of the 15 pedestrian safety issues were specifically being addressed by the agencies' data collection efforts. The agencies' existing methods could, however, be used to target seven additional safety factors. The development of a pedestrian data monitoring guide is recommended; an outline is proposed. Several agencies admitted that pedestrian volumes did not affect the resultant treatments. Evidently, some transportation agencies could benefit from direction on how to relate pedestrian demand and behavior data to safety improvements.
Ground reaction forces from two force plates are used to calculate the cyclic oscillations of the body centre of mass of subjects walking at preferred speed. Good approximations to the oscillations may be obtained from formulae containing just the first- and second-order Fourier coefficients of the ground reaction forces taken over a complete walking cycle. The symmetric components of the oscillations have consistent mutual phase relations for normal subjects, so that the amplitudes alone can be used as sufficient parameters to characterize the body centre of mass oscillations. The technique enables detection of small but consistent gait asymmetries. The walking speed strongly influences some of the symmetric gait parameters but, for normal subjects at least, the walking speed does not affect the asymmetric parameters.
This paper presents the results of a before-after (BA) safety evaluation of a newly proposed design for channelized right-turn lanes. The new design, termed "Smart Channels", decreases the angle of the channelized right turn to approximately 70°. The implementation of these modified right-turn channels is usually advocated to allow for safer pedestrian crossing. However, the benefits also extend to vehicle-vehicle interactions since the new approach angle affords drivers a better view of the traffic stream they are to merge with. The evaluation is conducted using a video-based automated traffic conflict analysis. There are several advantages that support the adoption of traffic conflict techniques in BA safety studies. Traffic conflicts are more frequent than road collisions and are of marginal social cost, they provide insight into the failure mechanism that leads to road collisions, and BA studies based on traffic conflicts can be conducted over shorter periods. As well, the use of automated conflict analysis overcomes the reliability and repeatability problems usually associated with manual conflict observations. Data for three treatment intersections and one control intersection in Penticton, British Columbia, are used in this study. The results of the evaluation show that the implementation of the right-turn treatment has resulted in a considerable reduction in the severity and frequency of merging, rear-end, and total conflicts. The total average hourly conflict was reduced by about 51% while the average conflict severity was reduced by 41%.
This study uses continuously logged driving data from 166 private cars to derive the level of jerks caused by the drivers during everyday driving. The number of critical jerks found in the data is analysed and compared with the self-reported accident involvement of the drivers. The results show that the expected number of accidents for a driver increases with the number of critical jerks caused by the driver. Jerk analyses make it possible to identify safety critical driving behaviour or "accident prone" drivers. They also facilitate the development of safety measures such as active safety systems or advanced driver assistance systems, ADAS, which could be adapted for specific groups of drivers or specific risky driving behaviour.
Annual electric bike (e-bike) sales in China grew from 40,000 in 1998 to 10 million in 2005. This rapid transition from human-powered bicycles and gasoline-powered scooters to an all-electric vehicle/fuel technology system is special in the evolution of transportation technology and, thus far, unique to China. We examine how and why e-bikes developed so quickly in China with particular focus on the key technical, economic, and political factors involved. This case study provides important insights to policy makers in China and abroad on how timely regulatory policy can change the purchase choice of millions and create a new mode of transportation. These lessons are especially important to China as it embarks on a large-scale transition to personal vehicles, but also to other countries seeking more sustainable forms of transportation.
A traffic encounter between individual road users is a process of continuous interplay over time and space and may be seen as an elementary event with the potential to develop into an accident. This paper proposes a framework for organising all traffic encounters into a severity hierarchy based on some operational severity measure. A severity hierarchy provides a description of the safety situation and trade-off between safety and efficiency in the traffic system. As a first approach to study the encounter process, a set of indicators is proposed to describe an encounter. These indicators allow for a continuous description even if the relationship between the road users changes during the process (e.g., when they are on a collision course or leave it). Automated video analysis is suggested as a tool that will allow data collection for validation of the proposed theories.
This paper presents a survey investigating the effects of age, gender and conformity tendency on Chinese pedestrians' intention to cross the road in potentially dangerous situations. A sample of 426 respondents completed a demographic questionnaire, a scale measuring their tendency towards social conformity, and a questionnaire based on the theory of planned behavior (TPB). This questionnaire measured people's intentions to cross the road in two different road crossing situations, their attitude towards the behavior, subjective norms, perceived behavioral control, anticipated affect, moral norms, and perceived risk. The two scenarios depicted (i) a situation where the crossing was consistent with other pedestrians' behavior (Conformity scenario) and (ii) a situation where the road crossing was inconsistent with other pedestrians (Non-Conformity scenario). Pedestrians reported greater likelihood in crossing the road when other pedestrians were crossing the road. People who showed greater tendencies towards social conformity also had stronger road crossing intentions than low conformity people for both scenarios. The predictive model explained 36% and 48% of the variance in the Non-Conformity and Conformity scenarios, respectively. Attitude, subjective norm, perceived behavioral control, and perceived risk emerged as the common predictors for both situations. The results have a number of theoretical and practical implications. In particular, interventions should focus on perceptions of risk that inform road users that crossing with other pedestrians against the signal is also unsafe and prohibited, and may lead to negative outcomes.
This paper describes a traffic conflicts computer simulation model and graphic display for both T and 4-leg unsignalized intersections. The goal of the model is to study traffic conflicts as critical-event traffic situations and the effect of driver and traffic parameters on the occurrence of conflicts. The analysis extends conventional gap acceptance criteria to describe driver's behaviour at unsignalized intersections by combining some aspects of gap acceptance criteria and the effect of several parameters including driver's characteristics such as age, sex, and waiting time. The effect of different traffic parameters such as volume and speed on the number and severity of traffic conflicts is also investigated. The model is unique insofar as it uses a technique of importance sampling and stores the traffic conflicts that occur during the simulation for later study. A graphical animation display is used to show how these conflicts occurred and the values of critical variables at the time. Model results were evaluated against previous work in the literature and validated by using field observations from four unsignalized intersections. The simulation results correlated reasonably well with actual conflict observations and should prove useful for assessing safety performance and feasible solutions for other unsignalized intersections.
We report a study in which we examined the gait modulations performed by a group of healthy adults, naïve to the purpose of the study, while walking across a curb under natural environmental conditions out of doors. Data were collected for step time and step length using a videocamera and a 20m walkway. Location of foot position was achieved using a novel method to account for parallax and out of plane motion. The results obtained indicated that during both curb ascent and descent, subjects were able to predict accurately the need to make an adjustment to step length in order to avoid placing the foot in a potentially hazardous position on or near the curb. These step adjustments involved both lengthening and shortening the step as well as keeping it unchanged. In any case, the selection of strategy was not apparently influenced by the walking speed but was a function of the predicted foot position with respect to the curb. Also, the time at which the step adjustment was made appeared independent of the time to contact with the curb. These results suggest that, under natural circumstances, accurate planning for obstacle negotiation is made some distance before the obstacle.
This paper reports on the development of a method for automatic monitoring of safety at Pelican crossings. Historically, safety monitoring has typically been carried out using accident data, though given the rarity of such events it is difficult to quickly detect change in accident risk at a particular site. An alternative indicator sometimes used is traffic conflicts, though this data can be time consuming and expensive to collect. The method developed in this paper uses vehicle speeds and decelerations collected using standard in situ loops and tubes, to determine conflicts using vehicle decelerations and to assess the possibility of automatic safety monitoring at Pelican crossings. Information on signal settings, driver crossing behaviour, pedestrian crossing behaviour and delays, and pedestrian-vehicle conflicts was collected synchronously through a combination of direct observation, video analysis, and analysis of output from tube and loop detectors. Models were developed to predict safety, i.e. pedestrian-vehicle conflicts using vehicle speeds and decelerations.
The aim of this work is to be a starting point for a more thorough description and analysis of safety related road user behaviour in order to better understand the different parts forming the traffic safety processes. The background is that it is problematic to use analysis of crash data and conflict data in the everyday traffic safety work due to low occurrence rates and the focus on rather exceptional and unsuccessful events. A new framework must consider the following aspects: (1) The importance of feedback to the road users. (2) Inclusion of more frequent events, "normal" road user behaviours and the possibility to link them to a severity dimension. (3) Prediction of safety/unsafety based on the more frequent events. By constructing severity hierarchies based on a uniform severity dimension (Time to Accident/Conflicting Speed value) it is possible to both describe the closeness to a crash and to get a comprehensive understanding of the connection between behaviour and safety by both considering unsuccessful and successful interactive situations. These severity hierarchies would make it possible to consider road users' expectations due to feedback and estimate its safety relevance.