Pedestrian-vehicle Interactions Research Articles

Level of safety and crossing warrants for cross-walks using surrogate safety measures

Pedestrians are the most vulnerable road users and they are highly susceptible to conflicts at mid-block cross-walk locations. They are prone to a high risk of injury from vehicular impact incidents and their decision making behaviour at unmarked cross-walks (UCWs) is more critical than at marked cross-walks (MCWs). The objective of this study was to determine level of safety thresholds and warrants for both marked and unmarked pedestrian crossing facilities at urban mid-block sections under mixed traffic conditions. Data were collected at MCWs and UCWs in four different locations in Trivandrum city, India. The crossing behaviours of pedestrians at both types of cross-walks were compared by analysing characteristics such as gap acceptance, a surrogate safety measure (post-encroachment time (PET)), waiting time, crossing time, platoon size, type of critical vehicle and critical gap. The multiple linear regression technique was used to develop gap acceptance models for MCWs and UCWs. The pedestrian level of safety (PLS) criterion is proposed; the PLS describes the safety of pedestrians at cross-walks and warrants for crossings based on PETs. The developed PLS thresholds and the pedestrian–vehicle interaction values were correlated to the PETs to develop the pedestrian crossing warrants.

Pedestrian-vehicle interaction severity level assessment at uncontrolled intersections using machine learning algorithms

As a consequence of the rapid growth of vehicular traffic, there is an increase in interactions between vehicles and pedestrians. The severity of these interactions varies with pedestrian, vehicle and roadway geometric characteristics. In the absence of real crash data, Surrogate Safety Measures (SSMs) are used to analyse the pedestrian-vehicle (P-V) interactions. The present study is intended to propose threshold risk indicator (RI) values for severe P-V interactions using both pedestrian and vehicle characteristics. A multilinear regression (MLR) P-V interaction model was developed using SPSS (Statistical Package for the Social Sciences) software. Videography method was used to collect traffic data from two 4-legged uncontrolled intersections. Pedestrian and vehicular data were extracted from the video using DataFromSky viewer software and risk indicator was calculated using post encroachment time and approaching vehicular speed. The interactions between pedestrians and vehicles were classified as normal conflicts and severe conflicts based on visual observations during the data extraction process. Python interface with support vector machines (SVM) algorithm was used to get threshold RI values for various pedestrian (gender and speed) and vehicle (type) characteristics.From SVM results, it was observed that the threshold RI value for severe interactions decreases as the pedestrian crossing speed increases for the same vehicle and pedestrian characteristics. MLR results showed that pedestrian gender, age and speed, vehicle type and speed, interaction location and crossing position have a significant effect on RI. The results can be used to evaluate pedestrian-vehicle interaction severity level at an uncontrolled intersection.

Multiagent modeling of pedestrian-vehicle conflicts using Adversarial Inverse Reinforcement Learning

There is a need for a better understanding of the collision avoidance behavior of road users in near misses. Recently, several models of road user behavior in near misses have been proposed. However, despite the multiagent nature of road user interactions, most of these studies modeled their behavior using a single-agent approach. However, this approach is unrealistic and can limit the models’ accuracy. Therefore, this study proposes the Markov-Game (MG) framework for modeling pedestrian-vehicle interactions and their collision avoidance mechanisms. Pedestrian-vehicle conflicts in a mixed traffic environment in China are extracted using computer-vision algorithms. Pedestrian and vehicle reward functions are recovered via the Multiagent Adversarial Inverse-Reinforcement-Learning approach. Road user optimal policies and collision avoidance mechanisms are predicted using multiagent Actor-Critic deep-reinforcement-learning. The results demonstrate the superiority of the multiagent modeling approach in predicting road user behavior, their collision avoidance mechanisms, and the Post-Encroachment-Time (PET) compared to a baseline single-agent model.

Agent-Based Modeling for Predicting Pedestrian Trajectories Around an Autonomous Vehicle

This paper addresses modeling and simulating pedestrian trajectories when interacting with an autonomous vehicle in a shared space. Most pedestrian–vehicle interaction models are not suitable for predicting individual trajectories. Data-driven models yield accurate predictions but lack generalizability to new scenarios, usually do not run in real time and produce results that are poorly explainable. Current expert models do not deal with the diversity of possible pedestrian interactions with the vehicle in a shared space and lack microscopic validation. We propose an expert pedestrian model that combines the social force model and a new decision model for anticipating pedestrian–vehicle interactions. The proposed model integrates different observed pedestrian behaviors, as well as the behaviors of the social groups of pedestrians, in diverse interaction scenarios with a car. We calibrate the model by fitting the parameters values on a training set. We validate the model and evaluate its predictive potential through qualitative and quantitative comparisons with ground truth trajectories. The proposed model reproduces observed behaviors that have not been replicated by the social force model and outperforms the social force model at predicting pedestrian behavior around the vehicle on the used dataset. The model generates explainable and real-time trajectory predictions. Additional evaluation on a new dataset shows that the model generalizes well to new scenarios and can be applied to an autonomous vehicle embedded prediction.

A Bayesian correlated grouped random parameters duration model with heterogeneity in the means for understanding braking behaviour in a connected environment

Driver’s response to a pedestrian crossing requires braking, whereby both excess and inadequate braking is directly associated with crash risk. The highly anticipated connected environment aims to increase drivers’ situational awareness by providing advanced information and assisting them during critical driving tasks such as braking. Focussing on this crucial behaviour and combined with the promise of a connected environment, the objective of this study is to examine the braking behaviour of drivers in response to a pedestrian at a zebra crossing in a connected environment. Seventy-eight participants from diverse backgrounds performed this driving task in the CARRS-Q Advanced Driving Simulator in two randomised driving scenarios: a baseline scenario (without driving aids) and a connected environment (with driving aids) scenario. A Weibull accelerated failure time duration modelling approach is adopted to model the braking behaviour of drivers. In particular, this duration model is specified to capture the panel nature of the data and unobserved heterogeneity through correlated grouped random parameters with heterogeneity-in-the-means in the Bayesian framework. Results indicate that, for most drivers in the connected environment, it takes longer to reduce their speed with less speed variation and a larger safety margin. In addition, a decision tree analysis for the braking time suggests that for older drivers, when the distance to the zebra crossing is larger in the connected environment than that in the baseline scenario, braking time is likely to increase. The model also reveals that the braking time of female drivers is longer in the connected environment compared to that of male drivers. Overall, the connected environment is associated with increased braking time by providing advanced information, giving drivers additional time to smoothly reduce their speed in response to a pedestrian at a zebra crossing, and ultimately making the vehicle–pedestrian interaction safer.

LiDAR-Based Vehicle–Pedestrian Interaction Study on Midblock Crossing Using Trajectory-Based Modified Post-Encroachment Time

Understanding pedestrian road crossing behavior is crucial in the interests of pedestrian traffic safety. The interaction between pedestrians and approaching vehicles at midblock road crossings is a complex process. The high degree of vehicle heterogeneity and poor lane discipline in low- and middle-income countries further adds to its complexity. Moreover, because of the lack of proper pedestrian facilities, pedestrian crossing behavior is quite different in low- and middle-income countries. It is necessary to have high-resolution data to study pedestrian crossing behavior. At midblock, pedestrians do not always take the shortest route while crossing the road and cross diagonally, moving toward or away from the approaching vehicle. The risk indicator of crossing changes based on the pedestrian path chosen. This study was based on studying vehicle–pedestrian interaction at midblock locations using light detection and ranging (LiDAR) data. This study considered a trajectory-based pedestrian modified post-encroachment time (PET) to account for the various possible pedestrian crossing trajectories. Suitable algorithms were developed to identify and track vehicles and pedestrians from the LiDAR point cloud data. The results showed the trajectory-based pedestrian modified PET to perform better than the traditional PET. The risk-taking appetite of pedestrians was also observed to be high when motorbikes were the approaching vehicle. This study showed the motorbike–pedestrian conflicts to be riskier as compared with car–pedestrian conflicts. The results obtained from this study would be helpful to practitioners in investing higher priority to proper pedestrian-related road infrastructure.

Designing Wearable Augmented Reality Concepts to Support Scalability in Autonomous Vehicle-Pedestrian Interaction

Wearable augmented reality (AR) offers new ways for supporting the interaction between autonomous vehicles (AVs) and pedestrians due to its ability to integrate timely and contextually relevant data into the user's field of view. This article presents novel wearable AR concepts that assist crossing pedestrians in multi-vehicle scenarios where several AVs frequent the road from both directions. Three concepts with different communication approaches for signaling responses from multiple AVs to a crossing request, as well as a conventional pedestrian push button, were simulated and tested within a virtual reality environment. The results showed that wearable AR is a promising way to reduce crossing pedestrians' cognitive load when the design offers both individual AV responses and a clear signal to cross. The willingness of pedestrians to adopt a wearable AR solution, however, is subject to different factors, including costs, data privacy, technical defects, liability risks, maintenance duties, and form factors. We further found that all participants favored sending a crossing request to AVs rather than waiting for the vehicles to detect their intentions—pointing to an important gap and opportunity in the current AV-pedestrian interaction literature.

Impacts of enforcement cameras on pedestrians’ risk perception and drivers’ behaviors at non-signalized crosswalks

Introduction: Pedestrians are more vulnerable to traffic crashes than other road users, particularly at non-signalized crosswalks. Safety measures (such as law enforcement cameras) can be implemented to regulate road users’ behavior and traffic safety. This study evaluates the effects of such cameras on pedestrian-vehicle conflicts by investigating different interaction patterns of pedestrian risk perception and driving style. Data: Field investigations were conducted at four non-signalized crosswalks. Video data were collected using unmanned aerial vehicles and roadside cameras. Method: Two-step cluster analysis and k-means cluster analysis were employed to classify the pedestrian’s behavior and driving style, respectively. Surrogate safety measures were adopted to measure the pedestrian-vehicle conflicts. Results and conclusions: The results suggest that the implementation of cameras would decrease both the actual and perceived risks of pedestrians, while the heterogeneity between the actual and perceived risk is more obvious at camera sites. They also indicate that the cameras have a positive influence on reducing drivers’ aggressiveness and conflict severity. In terms of pedestrian-vehicle interaction patterns, the most severe conflicts occur when the pedestrian perceived risk level is low and the driving style is aggressive. Such dangerous interactions are observed more frequently at camera sites. In contrast, a safer interaction pattern is associated with a moderate driving style and cautious crossing behavior, which is more frequently observed at comparison sites. However, regardless of which interaction pattern is observed, the conflict severity is found to be lower at camera sites, indicating the effectiveness of the cameras. Practical Applications: Supplementary facilities, such as warning signs, flash lights, and speed control measures, should be implemented to maintain the effectiveness of the law enforcement cameras.

Vehicle Pedestrian Interaction Analysis at Unsignalized Intersection using Trajectory Data

As the pedestrian’s rapid progress, pedestrian safety has recently assumed greater importance in the research population activities. To assess road user interactions at an unsignalized junctions with heterogeneous traffic complexity, innovative trajectory-based data was used and to make urban intersections safer for road users, the proposed severity levels will be used to test and evaluate numerous infrastructures and control upgrades. Study authors suggests an advanced pattern-based technique to characterize pedestrian-vehicle interactions based on road user behaviors. Surrogate safety measures (SSM) can be estimated more accurately with this study than with the regular grid-based analysis. In order to evaluate SSM (Speed, Time to Collision (TTC), Gap Time of the pedestrian and vehicle interactions) at an unsignalized crossing, Safe Distance and Stopping sight distance with trajectory data are to be estimated. Support Vector Machine (SVM) is used to classify severity grades based on specified indicators generated at an unsignalized junction in India. From the analysis, Severity at the intersection found as 8.12, 8.91sec respectively, Average gap time, stopping sight distance also calculated. Plots between Time to Collision and Gap Time for pedestrian passing first (PPF), vehicle passing first (VPF) are compared by developing a linear regression model and R2=0.684, 0.656 developed with some independent parameters respectively. Concluded that TTC for PPF is higher than VPF and by considering all the evaluated values this research had improved the methods for analysing and improving the safety of uncontrolled intersections.

Application of the Theory of Planned Behavior in Autonomous Vehicle-Pedestrian Interaction

Automobile manufacturers, alongside technology providers, researchers, and public agencies, are conducting extensive testing to design autonomous vehicles (AVs) algorithms that will provide a complete understanding of road users, specifically pedestrians. Pedestrian behavior and actions determination are highly unpredictable depending on behavioral beliefs, context, and socio-demographic variables. Context includes everything that potentially affects one’s behavior; in AVs–pedestrian interaction, context may consist of weather conditions, road structure, social factors norms, and traffic volume. These influencing elements, therefore, need to be focused on during the development of pedestrian interaction algorithms. For this purpose, the pedestrian behavior questionnaire for FAVs (PBQF) is designed based on the theory of planned behavior (TPB). A total of almost 1000 voluntary participants completed this multilingual survey. As socio-demographic values and physiological perception varies with local norms, regions, and ethnicity, participants from 27 countries were therefore chosen to account for this variation. One of the key findings of this study is the influence of pedestrian attributes and the context on pedestrian behavior. Pedestrian action cannot be understood without visual observation of the pedestrian themselves and their context. The findings showed that pedestrians build communication with vehicles based on their driving styles. The vehicle’s driving style leads pedestrians to think that the vehicle is human-driven or autonomous. The results also revealed that pedestrians use several cues to show their intention. The general perception of AVs was also analyzed, and the communication between AVs and pedestrians with different displaying options was investigated.

Real-Time Pedestrian Conflict Prediction Model at the Signal Cycle Level Using Machine Learning Models

Compared with traditional traffic studies, real-time safety analyses can be better incorporated into proactive traffic management strategies to improve traffic safety. However, few studies have investigated the real-time pedestrian safety model. Intersections usually have mixed traffic conditions with more pedestrian-vehicle interactions. This paper uses conflict indicators, PET (Post Encroachment Time) and TTC (Time to Collision) to identify pedestrians’ conflicts from CCTV (closed-circuit television) videos. The high-resolution traffic data from the Automated Traffic Signal Performance Measures (ATSPM) system are used to derive traffic flow-related variables. The pedestrian exposure is also estimated. Pedestrians’ conflicts are predicted using multiple machine learning models and Logistic Regression. The resampling methods, random over-sampling, and random under-sampling are compared. The best model, Extreme Gradient Boosting (XGBT) with random over-sampling method can achieve AUC (area under the ROC curve) value of 0.841 and recall value of 0.739 on the test data set. The proposed model can predict pedestrians’ conflicts one cycle ahead, which can be 2–3 min. The proposed model has the potential to be implemented in the Connected and Automated Vehicles (CAV) environment to adjust signal timing accordingly and enhance traffic safety.

Adaptive Road Configurations for Improved Autonomous Vehicle-Pedestrian Interactions Using Reinforcement Learning

The deployment of Autonomous Vehicles (AVs) poses considerable challenges and unique opportunities for the design and management of future urban road infrastructure. In light of this disruptive transformation, the Right-Of-Way (ROW) composition of road space has the potential to be renewed. Design approaches and intelligent control models have been proposed to address this problem, but we lack an operational framework that can dynamically generate ROW plans for AVs and pedestrians in response to real-time demand. Based on microscopic traffic simulation, this study explores Reinforcement Learning (RL) methods for evolving ROW compositions. We implement a centralised paradigm and a distributive learning paradigm to separately perform the dynamic control on several road network configurations. Experimental results indicate that the algorithms have the potential to improve traffic flow efficiency and allocate more space for pedestrians. Furthermore, the distributive learning algorithm outperforms its centralised counterpart regarding computational cost (49.55\%), benchmark rewards (25.35\%), best cumulative rewards (24.58\%), optimal actions (13.49\%) and rate of convergence. This novel road management technique could potentially contribute to the flow-adaptive and active mobility-friendly streets in the AVs era.

Analyzing vehicle–pedestrian interactions: Combining data cube structure and predictive collision risk estimation model

Road traffic accidents are a severe threat to human lives, particularly to vulnerable road users (VRUs) such as pedestrians causing premature deaths. Therefore, it is necessary to devise systems to prevent accidents in advance and respond proactively, using potential risky situations as one of the surrogate safety measurements. This study introduces a new concept of a pedestrian safety system that combines the field and the centralized processes. The system can warn of upcoming risks immediately in the field and improve the safety of risk-frequent areas by assessing the safety levels of roads without actual collisions. In particular, this study focuses on the latter by introducing a new analytical framework for a crosswalk safety assessment with various behaviors of vehicles/pedestrians and environmental features. We obtain these behavioral features from actual traffic video footages in the city with complete automatic processing. The proposed framework mainly analyzes these behaviors in multi-dimensional perspectives by constructing a data cube structure, which combines the Long Short-Term Memory (LSTM)-based predictive collision risk (PCR) estimation model and the on-line analytical processing (OLAP) operations. From the PCR estimation model, we categorize the severity of risks as four levels; “relatively safe,” “caution,” “warning,” and “danger,” and apply the proposed framework to assess the crosswalk safety with behavioral features. With the proposed framework, the various descriptive results are harvested, but we aim at conducting analysis based on two scenarios in our analytic experiments; the movement patterns of vehicles and pedestrians by road environment and the relationships between risk levels and car speeds. Consequently, the proposed framework can support decision-makers (e.g., urban planners, safety administrators) by providing the valuable information to improve pedestrian safety for future accidents, and it can help us better understand cars’ and pedestrians’ proactive behavior near the crosswalks. In order to confirm the feasibility and applicability of the proposed framework, we implement and apply it to actual operating CCTVs in Osan City, Republic of Korea.

A Review of Virtual Reality Studies on Autonomous Vehicle–Pedestrian Interaction

An increasing number of studies employ virtual reality (VR) to evaluate interactions between autonomous vehicles (AVs) and pedestrians. VR simulators are valued for their cost-effectiveness, flexibility in developing various traffic scenarios, safe conduct of user studies, and acceptable ecological validity. Reviewing the literature between 2010 and 2020, we found 31 empirical studies using VR as a testing apparatus for both implicit and explicit communication. By performing a systematic analysis, we identified current coverage of critical use cases, obtained a comprehensive account of factors influencing pedestrian behavior in simulated traffic scenarios, and assessed evaluation measures. Based on the findings, we present a set of recommendations for implementing VR pedestrian simulators and propose directions for future research.

Experimental analysis of pedestrian behavior at different configurations of crosswalks at roundabout legs

Walking has recently been one of the most popular forms of mobility proposed in many urban plans and in the new concept of sustainable mobility. This study investigated pedestrian behavior when crossing the legs of urban roundabouts. Pedestrian crossing behavior was analyzed using video recordings made at seven crosswalks located on urban roundabouts. To describe pedestrian behavior, crossing conditions (legal or illegal) and Kerb Delay were chosen. The data obtained from the video analysis were treated with CHAID analysis to determine the distribution of the parameters chosen as a function of features considered more conditioning. A path analysis was then performed to assess the influence of these features on the parameters defining pedestrian behavior. In order to encourage legal pedestrian behavior, the results of the present study suggest implementation of refuge islands at the legs of roundabouts. The research results can help refine the algorithms that define autonomous vehicle-pedestrian interactions.

An exploratory study of pedestrian crossing speeds at midblock crossing in India using LiDAR

Objective Understanding pedestrian road crossing behavior is essential from the perspectives of traffic flow and pedestrian safety. Limited research is available on pedestrian behavior in low- and middle-income countries. The main objective of this study is to understand pedestrian–vehicle interactions during midblock crossings in heterogeneous traffic conditions. Specifically, this study aims to understand whether pedestrians alter their crossing behavior depending on the type of approaching vehicles. Methods To better understand pedestrian road crossing behavior at midblock crossings, an instrumented vehicle collected data from Kanpur, a large city in Uttar Pradesh, India. Because light detection and ranging provides point clouds at high frequency, an algorithm was developed to identify and track vehicles and pedestrians. Specifically, 2 types of interactions at midblock crossings were studied: car–pedestrian and motorized bike–pedestrian. The walking speed profiles and trajectories of the pedestrians were analyzed. Results The results show that pedestrians are more willing to engage in risky road crossing behavior in front of motorized bikes than in front of cars. Pedestrian walking speed profiles were unaffected by motorized bikes, but for cars, pedestrians tended to increase their speed in the first half of road crossing and then decrease in the second half. Conclusions Pedestrian crossing speed profiles play an essential role in understanding pedestrian midblock crossing behavior. The speed data for pedestrians at various points of crossing are challenging to capture, but this study shows that LiDAR can be used to capture detailed pedestrian movements. The findings from this study demonstrate the importance of considering vehicle heterogeneity when analyzing pedestrian risk exposure and designing pedestrian crossing facilities.

Kinetic and Kinematic Features of Pedestrian Avoidance Behavior in Motor Vehicle Conflicts.

The active behaviors of pedestrians, such as avoidance motions, affect the resultant injury risk in vehicle–pedestrian collisions. However, the biomechanical features of these behaviors remain unquantified, leading to a gap in the development of biofidelic research tools and tailored protection for pedestrians in real-world traffic scenarios. In this study, we prompted subjects (“pedestrians”) to exhibit natural avoidance behaviors in well-controlled near-real traffic conflict scenarios using a previously developed virtual reality (VR)-based experimental platform. We quantified the pedestrian–vehicle interaction processes in the pre-crash phase and extracted the pedestrian postures immediately before collision with the vehicle; these were termed the “pre-crash postures.” We recorded the kinetic and kinematic features of the pedestrian avoidance responses—including the relative locations of the vehicle and pedestrian, pedestrian movement velocity and acceleration, pedestrian posture parameters (joint positions and angles), and pedestrian muscle activation levels—using a motion capture system and physiological signal system. The velocities in the avoidance behaviors were significantly different from those in a normal gait (p < 0.01). Based on the extracted natural reaction features of the pedestrians, this study provides data to support the analysis of pedestrian injury risk, development of biofidelic human body models (HBM), and design of advanced on-vehicle active safety systems.

Interactions between autonomous vehicles and pedestrians at unsignalized mid-block crosswalks considering occlusions by opposing vehicles

Visibility can be identified as one of the critical determinants for the safety performance of autonomous vehicles (AVs) on unsignalized mid-block crosswalks (UMC), which may be significantly influenced by build-up environment and surrounding vehicles. This study investigates the safety performance when AVs interact with pedestrians approaching from far-side sidewalks to UMCs considering the visual occlusion of opposing vehicles. A mathematical model is proposed for judging the visibilities of objects from observers’ location under the impact of visual obstacles and is embedded into an agent-based pedestrian-vehicle interaction framework. Two yielding decision modules is assumed for AVs: The normal decision module implements the pedestrian priority rule simply based on the current detectable information, whereas the memory aid decision module extends AVs’ detection abilities by incorporating the memory data. Through simulation experiments, it is found that the percentages of short post encroachment time (%SPET) between AVs and far-side pedestrians reach peaks when the pedestrian flow rate is 300–400 ped/h. When opposing vehicles are in stationary queue conditions, %SPETs are only sensitive to the net distance between the last opposing vehicles in the queue and crosswalks (Dqueue). As the Dqueue decreases to lower than 15 m, %SPETs start to increase drastically. However, when opposing vehicles are in free flow conditions, %SPETs are influenced by multiple factors such as pedestrians’ crossing decisions, sizes and flow rates of opposing vehicles. Furthermore, only when opposing vehicles are in free flow conditions, memory aid AVs can significantly eliminate the impacts of opposite vehicles. Finally, several countermeasures are developed to enhance the visibility and safety at UMCs based on the findings of this study.

Deep learning for autonomous vehicle and pedestrian interaction safety

The present work aims to study how deep learning approaches solve the safety problems in the interaction between autonomous vehicles and pedestrians. A Vehicle-Pedestrian Detection (VPD) algorithm based on Convolutional Neural Network (CNN) is proposed regarding the massive amounts of parameters during feature extraction of traditional vehicle–pedestrian interaction algorithms. Furthermore, the Squeezenet algorithm is applied to extract traffic characteristics with fewer parameters. The performance of the proposed algorithm is analyzed through simulation experiments. Results demonstrate that when the successful transmission probability reaches 100% and the λ value is 0.01–0.05, the proposed algorithm can provide the result closest to the actual value, with the smallest data delay and the highest data transmission security. In different categories, the proposed algorithm can provide the highest accuracy as the number of iterations increases compared with other algorithms (AlexNet, DenseNet, VGGNet, IGCNet, and ResNet), which can accurately forecast traffic safety accidents. The proposed algorithm can provide an accuracy of 81.98%, an improvement of at least 1.94% over other advanced CNNs, and at least 3.3% over algorithms included in comparative simulations. Hence, it can recognize and identify safe interactions between autonomous vehicles and pedestrians. Through experiments, the constructed algorithm can significantly reduce the data transmission delay, improve the prediction accuracy of the safe interaction between autonomous vehicles and pedestrians, and increase the recognition accuracy remarkably, which can provide experimental references for the intelligent development of the transportation industry in the future.

Investigating yielding behavior of heterogeneous vehicles at a semi-controlled crosswalk

It is well known that pedestrians are vulnerable road users. Their risk of being injured or killed in road traffic crashes is even higher as vehicle drivers often violate traffic rules and do not slow down or yield in front of crosswalks. In order to reduce this risk, many countries have issued strict regulations requiring vehicles to yield to pedestrians in front of crosswalks. While extensive literature exists on the interaction between vehicles and pedestrians, the consideration of heterogeneity in the behavior of vehicles is vastly overlooked. Accordingly, this study analyzes the yielding behavior of three types of vehicles under the “pedestrian priority” policy by processing drone footage collected in Xi'an City (China) with a Machine Vision Intelligent Algorithm. Moreover, this study proposes four additional indicators to the widely used yielding rate and yielding delay with the aim of evaluating yielding behavior of three types of vehicles. The results show that buses have the best yielding behavior from the perspective of yielding rate, yielding delay, waiting time, yielding angle and waiting site. Buses perform well in observing pedestrian dynamics near crosswalk, and perform exceptionally well in considering the “blind area” of vision. The location of the waiting site in front of the stop line and the length of the waiting time contribute to the safe crossing of pedestrians. In contrast, private cars perform badly in yielding to pedestrians. However, serious polarization can be observed across private cars, as the performance varies across the board. The relaxation of the homogenization assumption of the behavior of vehicles in pedestrian-vehicle interaction, alongside the improvements in the analysis via Machine Vision Intelligent Algorithm of videos acquired via drone, shows the possibility of having a deeper understanding of the yielding behavior of vehicles at crosswalk. The extension of the use of artificial intelligence methods to analyze drone footage has immense potential in understanding road user behavior and hence providing knowledge for crash prevention.