Vulnerable Road Users Research Articles

Seeing through eyes of Drivers: Space consideration in investigating visibility of Vulnerable road users involved in crashes from Driver’s perspective

Vulnerable Road User’s (VRUs) invisibility by vehicle drivers hasn’t been well explored despite having a substantial influence on crash involvement and resulting severity level. Additionally, obtaining comparison crashes for analysis of the VRU invisibility has been a challenge. For that reason, this study used crashes that occurred between 2017 and 2022 in Ohio to understand VRU invisibility from the driver’s perspective. The study further proposes the comparison of crashes as those that occurred within 250 feet of the crashes involving drivers not seeing the VRU. Two logistic regression models, one for the entire dataset (full model) and the second for only crashes that occurred within 250 feet (space-constrained model), were developed. It was found that the results from the full model and space-constrained model differ significantly in terms of the magnitude and the direction of the effect. Using the space-constrained model, the topmost key factors associated with the highest likelihood of VRU invisibility are lighting conditions, pre-action of the driver, and senior VRU involvement. Further, text network analysis was performed to understand the key reasons for VRU invisibility. The text network revealed that the VRU invisibility related to left turning pre-action was due to the driver’s failure to yield at an intersection’s pedestrian crossing. Further, the most invisible VRUs in the dark conditions were on the side of the roadway. Additionally, drivers backing up were more likely to report that they did not see pedestrians walking behind them. Lastly, senior-related crashes were associated with crossing in front of turning vehicles. The findings can be utilized to enhance VRU visibility at various locations to improve safety.

The impact of target speed on pedestrian, bike, and speeding crash frequencies

This research aims to investigate the influence of adopting the target speed concept on different types of crashes including pedestrian, bike, and speeding-related crashes. The Target speed is the highest speed that vehicles should operate on a roadway segment in a specific context. Based on the reviewed literature, this is the first study to investigate the relationship between target speed and crash frequency. Hence, big data including probe-vehicle data, traffic characteristics, geometric features, and land use attributes were utilized to develop crash prediction models. The main contributions of this research are to quantify the impacts of target speed on traffic safety considering context categories and to conclude the potential recommendations to lower different types of crashes. The 85th percentile speed was calculated and utilized in the developed models. Three crash prediction models were developed for pedestrian, bike, and speeding-related crashes. They were used in the analysis to quantify the impact of adopting target speed on different crash types. The results showed a significant reduction in the three crash types when using the target speed. Most of the improvements took place in three context categories: C3C: Suburban Commercial Segments, C3R: Suburban Residential Segments, and C4: Urban General Segments. Hence, this research recommends adopting target speed specifically in urban and suburban areas. Further, it suggests considering some measures to lower vulnerable road users’ and speeding-related crashes. Following the recommendations of this research would help to reduce different types of crash frequency, hence, improving the mobility and safety for all users in different context classifications.

Identifying User Needs and Current Challenges of External Interface Design for AV-VRU Communications: Insights from an Expert Survey Data Analysis

The recent deployment of highly automated vehicles (AVs) in various cities highlights the potential improvement of transportation safety but also reveals the need for improved interaction between AVs and other road users, particularly vulnerable road users (VRUs). The reduction of driver involvement in this AV-VRU interaction results in the absence of human-to-human communication cues and thus calls for new system designs. Researchers have proposed designing external human–machine interfaces (e-HMIs) to replace these communication cues and improve the AV-VRU interaction. However, designing e-HMIs is complex and requires consideration of various factors, such as stakeholder responsibilities, context of use, user needs, implementation, design principles, and usability testing. This paper examines potential e-HMI design solutions from the perspective of experts to determine user needs and identify challenges. An online survey was conducted with 47 experts in the AV and VRU field to gather their opinions on anticipated issues in AV-VRU interaction, preferred e-HMI design options, and related concerns. Both quantitative and qualitative analyses of the data revealed that experts view “Awareness” and “Intent” as the most important information to be communicated through visual e-HMIs placed on AVs. Open-ended questions also showed that there were fewer experts concerned about “Awareness” and “Intent” compared with other information types, and more experts concerned about auditory cues compared with visual cues and about VRU wearable or handheld devices installation locations compared with other locations. These findings provide insights into the efficiency, feasibility, effectiveness, and challenges in developing e-HMIs for improved and safer communication between AVs and VRUs.

A hybrid approach of random forest and random parameters logit model of injury severity modeling of vulnerable road users involved crashes

Vulnerable road users (VRUs) involved crashes are a major road safety concern due to the high likelihood of fatal and severe injury. The use of data-driven methods and heterogeneity models separately have limitations in crash data analysis. This study develops a hybrid approach of Random Forest based SHAP algorithm (RF-SHAP) and random parameters logit modeling framework to explore significant factors and identify the underlying interaction effects on injury severity of VRUs-involved crashes in Shenyang (China) from 2015 to 2017. The results show that the hybrid approach can uncover more underlying causality, which not only quantifies the impact of individual factors on injury severity, but also finds the interaction effects between the factors with random parameters and fixed parameters. Seven factors are found to have significant effect on crash injury severity. Two factors, including primary roads and rural areas produce random parameters. The interaction effects reveal interesting combination features. For example, even though rural areas and primary roads increase the likelihood of fatal crash occurrence individually, the interaction effect of the two factors decreases the likelihood of being fatal. The findings form the foundation for developing safety countermeasures targeted at specific crash groups for reducing fatalities in future crashes.

Multi-attention network for pedestrian intention prediction based on spatio-temporal feature fusion

An essential prerequisite for autonomous vehicles deploying in urban scenarios is the ability to accurately recognize the behavioral intentions of pedestrians and other vulnerable road users and take measures to ensure their safety. In this paper, a spatial-temporal feature fusion-based multi-attention network (STFF-MANet) is designed to predict pedestrian crossing intention. Pedestrian information, vehicle information, scene context, and optical flow are extracted from continuous image sequences as feature sources. A lightweight 3D convolutional network is designed to extract temporal features from optical flow. Construct a spatial encoding module to extract the spatial features from the context. Pedestrian motion information are re-encoded using a collection of gated recurrent units. The final network structure is created through ablation research, which introduces attention mechanisms into the network to merge pedestrian motion features and spatio-temporal features. The efficiency of the suggested strategy is demonstrated by comparison experiments on the datasets JAAD and PIE. On the JAAD dataset, the intent recognition accuracy is 9% more accurate than the existing techniques.

Investigating two-wheelers risk factors for severe crashes using an interpretable machine learning approach and SHAP analysis

The use of two-wheelers (TWs) has gained popularity as an alternative to personal vehicles due to their flexibility, fuel economy, ease of parking, and size, especially in congested cities. However, TWs are considered vulnerable road users due to their higher riding risk compared to other modes. This study proposes a novel framework to extract latent and dependent heterogeneous risk factors that affect the crash severity of TWs. By combining eXtreme Gradient Boosting (XGBoost) and SHapley Additive exPlanation (SHAP) analysis, this study investigates the factors affecting TW crash severity, providing both local and global interpretability. The XGBoost method is employed to model crash severity, while SHAP analysis facilitates the derivation of explanations from the model, enhancing our understanding of the contributing factors. The French crash dataset for TWs between 2014 and 2017 is utilized for this analysis. The findings highlight that the department of the crash, road category, urbanization level, TW category, and age of the user significantly influence TW crash severity. Furthermore, severe injuries are more likely to occur in TW crashes associated with rural areas, older riders, riders not wearing helmets, run-off-road crashes, and crossing roads. The insights derived from this study can be leveraged to develop targeted interventions that address the identified risk factors and promote the safety of TW riders. By focusing on these key factors, policymakers and stakeholders can implement effective measures to reduce the severity of TW crashes and enhance the overall safety of TW users.

Naturalistic driving study on the impact of an aftermarket blind spot monitoring system on the driver’s behaviour of heavy goods vehicles and buses on reducing conflicts with pedestrians and cyclists

The risk being severely or fatally injured in crashes with heavy goods vehicle and buses is much higher compared to other vehicles. Especially vulnerable road users such as pedestrian and cyclists are at high risk. In the European Union 4600 pedestrian and more than 2000 cyclists were killed in 2019. 18% of the fatalities were counted in crashes with heavy goods vehicles or buses. Blind spot situations and driver inattention or distraction are causing and contributing crash factors. Driver assistance systems are intended to support drivers and might have a positive effect on the crash avoidance. The objective of the study is the analysis of the change of driver behaviour in heavy goods vehicles and buses due to a aftermarket blind spot monitoring system. In a naturalistic driving study 15 heavy goods vehicles and five buses were equipped with a blind spot monitoring system and data were collected over a period of two years. The results revealed that the system would reduce the number of warnings with vulnerable road users by one third for heavy goods vehicles and 10% for buses. Up to 200 lives annually could potentially be saved with the analysed system, on the assumption that the collision warnings correlate directly with crashes.

Accurate detection of vehicle, pedestrian, cyclist and wheelchair from roadside light detection and ranging sensors

Accurate detection plays a critical role in improving the safety situation of vulnerable road users. This study extends infrastructure-based LiDAR application to all three major vulnerable road user groups including pedestrians, cyclists, and wheelchair users. Two critical problems for accurate detection of small-sized road users are scanning angle variability and feature fluctuation. To address these issues, a feature-based classification method combined with prior LiDAR trajectory information is developed. Effective dimension-related features are proposed and five classifiers including artificial neural network (ANN), random forest (RF), adaptive boosting (AdaBoost), random under-sampling boosting (RUSBoost), and long short-term memory (LSTM) are tested with a novel feature engineering process. A total of seven features are selected from the point cloud of clusters for vehicle/pedestrian/cyclist/wheelchair classification. By updating these significant features based on prior information of the entire trajectory, the performance of road user classification (imbalanced datasets) has been significantly improved. Experimental study is conducted to examine the recall rate, F1-score, and AUC of vehicles, pedestrians, cyclists, and wheelchairs before and after integration with prior trajectory information. The result shows the trained AdaBoost, RUSBoost, and LSTM classifiers with prior trajectory information can achieve recall/F1-score/AUC: (1) Low traffic volumes – vehicles (100%/99.96%/99.96%), pedestrians (99.96%/99.96%/99.97%), cyclists (99.74%/99.45%/99.67%), and wheelchairs (99.22%/99.68%/99.01%) and (2) Moderate traffic volumes – vehicles (99.39%/99.44%/99.69%), pedestrians (98.33%/97.99%/98.64%), and cyclists (95.41%/94.29%/94.40%), using 32-laser LiDAR sensors (10 Hz).

Comparative analysis of pedestrian crash severity at United Kingdom rural road intersections and Non-Intersections using latent class clustering and ordered probit model

Pedestrian safety is a critical issue in the United Kingdom (UK) as pedestrians are the most vulnerable road users. Despite numerous studies on pedestrian-vehicle crashes globally, limited research has been conducted to explore the factors contributing to such incidents in the UK, especially on rural roads. Therefore, this study aimed to investigate the severity of pedestrian injuries sustained on rural roads in the UK, including crashes at intersections and non-intersections. We utilized the STATS19 dataset, which provided comprehensive road safety data from 2015 to 2019. To overcome the challenges posed by heterogeneity in the data, we employed a Latent Class Analysis to identify homogeneous clusters of crashes. Additionally, we utilized the Ordered Probit model to identify contributing factors within each cluster. Our findings revealed that various factors had distinct effects on the severity of pedestrian injuries at intersections and non-intersections. Several parameters like the pedestrian location in footway and one-way roads are only statistically significant in the intersection section. Certain factors such as the day of the week, the pedestrian's location in a refuge, and minor roads (class B roads) were found to be significant only in the non-intersection section.Parameters includingpedestrians aged over 65 years and under 15 years, drivers under 25 years, male drivers and pedestrians, darkness, heavy vehicles, speed limits exceeding 96 km/h (60 mph), major roads (class A roads), and single carriageway roadsare significant in both sections. The study proposes various measures to mitigate the severity of pedestrian-vehicle crashes, such as improving lighting conditions, enhancing pedestrian infrastructure, reducing speed limits in crash-prone areas, and promoting education and awareness among pedestrians and drivers. The findings and suggested measures could help policymakers and practitioners develop effective strategies and interventions to reduce the severity of these incidents and enhance pedestrian safety.

A mobile mapping solution for VRU Infrastructure monitoring via low-cost LiDAR-sensors

ABSTRACT LiDAR scanning is a widely-used method for accurately assessing roads and vehicular infrastructure. However, it can be costly and often overlooks infrastructure important for Vulnerable Road Users (VRU) like pedestrians and cyclists. Advances in electronics now permit high-performance LiDAR sensors made of standard semiconductor materials to be integrated into personal devices like smartphones and tablets. This study introduces a new classification category of handheld LiDAR systems for the purpose of VRU infrastructure mapping on a smaller scale. Qualitative experiments using a handheld LiDAR system were conducted to achieve low-cost mobile mapping of VRU infrastructure. The 3D Scanner App facilitated detailed and qualitative point cloud generation at low speeds. Moreover, it demonstrated precise detection of perpendicular potholes and bumps on roads. A quantitative analysis based on several infrastructure damages was performed, revealing mean average percentage errors of 3.53%, 4.43%, and 10.57% in the x-, y-, and z-directions, respectively. While the accuracy in the x- and y-directions is acceptable, the higher z-direction error indicates overestimation of infrastructural damage depth. In conclusion, the sensor's capabilities enable cost-effective mobile mapping solutions for VRU infrastructure monitoring, such as crowd-sourcing for urban data mining and urban digital twin interfaces, significantly reducing expenses compared to traditional sensors.

Interaction Patterns of Motorists and Cyclists at Intersections: Insight from a Vehicle–Bicycle Simulator Study

At intersections, road users need to comprehend the intentions of others while also implicitly expressing their own intentions using dynamic information. Identifying patterns of this implicit communication between human drivers and vulnerable road users (VRUs) at intersections could enhance automated driving functions (ADFs), enabling more human-like communication with VRUs. To this end, we conducted a coupled vehicle–bicycle simulator study to investigate interactions between right-turning motorists and crossing cyclists. This involved 34 participants (17 pairs of motorists and cyclists) encountering each other in a virtual intersection. The analysis focused on identifying interaction patterns between motorists and cyclists, specifically aiming to discern which patterns were more likely to be accepted by both parties. We found that in CM (vehicles overtaking), the post-encroachment time (PET) and the average speed of vehicles were higher than in the other two interaction patterns: C (bicycles always in front) and CMC (bicycles overtake). However, subjective ratings indicated that CM was viewed as more critical and less cooperative. Furthermore, this study unveiled the influence of crossing order and overtaking position on subjective ratings through ordered logistic regressions, suggesting that earlier overtaking could improve cyclists’ acceptance of the interaction. These findings may contribute to the optimization of communication strategies for ADF, thereby ensuring safety in interactions with VRUs.

Analyzing the Risk Factors of Traffic Accident Severity Using a Combination of Random Forest and Association Rules

This study explores risk factors influencing the at-fault party in traffic accidents and analyzes their impact on traffic accident severity. Based on the traffic accident data of Shenyang City, Liaoning Province, China, from 2018 to 2020, 19 attribute variables including road attributes, time attributes, environmental attributes, and characteristics of the at-fault parties with either full responsibility, primary responsibility, or equal responsibility of the traffic accidents were extracted and analyzed in conjunction with the built environment attributes, such as road network density and POI (points of interest) density at the sites of traffic accidents. Using the RF-SHAP method to determine the relative importance of risk factors influencing the severity of traffic accidents with either motor vehicles or vulnerable groups at-fault, the top ten risk factors influencing the severity of traffic accidents with vulnerable road users as the at-fault parties are: functional zone, density of shopping POI, density of services POI, cause of accident, travel mode, collision type, season, road type, age of driver, and physical isolation. Travel mode, season, and road speed limit are more important risk factors for traffic accidents, with motor vehicle drivers as the at-fault parties. The density of service POI and cause of the accident are less critical for traffic accidents with motor vehicle drivers than traffic accidents with vulnerable road users who are at-fault. Subsequently, the Apriori algorithm based on association rules is used to analyze the important causal factors of traffic accidents, so as to explore the influence mechanism of multiple causal factors and their implied strong association rules. Our results show that most combined factors are associated with the matched Service and Shopping POI features. This study provides valuable information on the perceived risk of fatal accidents and highlights the built environment’s significant influence on fatal traffic accidents. Management strategies targeting the most typical combinations of accident risk factors are proposed for preventing fatalities and injuries in serious traffic accidents.

Replacing the human driver: An objective benchmark for occluded pedestrian detection

Early detection of vulnerable road users is a crucial requirement for autonomous vehicles to meet and exceed the object detection capabilities of human drivers. One of the most complex outstanding challenges is that of partial occlusion where a target object is only partially available to the sensor due to obstruction by another foreground object. A number of leading pedestrian detection benchmarks provide annotation for partial occlusion, however each benchmark varies greatly in their definition of the occurrence and severity of occlusion. Research demonstrates that a high degree of subjectivity is used to classify occlusion level in these cases and occlusion is typically categorized into 2–3 broad categories such as “partially” and “heavily” occluded. In addition, many pedestrian instances are impacted by multiple inhibiting factors which contribute to non-detection such as object scale, distance from camera, lighting variations and adverse weather. This can lead to inaccurate or inconsistent reporting of detection performance for partially occluded pedestrians depending on which benchmark is used. This research introduces a novel, objective benchmark for partially occluded pedestrian detection to facilitate the objective characterization of pedestrian detection models. Characterization is carried out on seven popular pedestrian detection models for a range of occlusion levels from 0%–99% to demonstrate the impact of progressive levels of partial occlusion on pedestrian detectability. Results show that the proposed benchmark provides more objective, fine grained analysis of pedestrian detection algorithms than the current state of the art.

Atypical intersection of a bicycle path and a carriageway with marked car priority: Which car drivers give way anyway?

The issue of the safety of vulnerable road users is particularly important at points of collision with other roads and road users. The most common solution at the intersection of a carriageway and a cycle path is a cycle crossing, with specific rules of priority for cyclists and drivers. However, the constantly increasing number of accidents involving cyclists prompts the creation of new and innovative solutions in infrastructure and traffic organisation. One novel solution in this respect takes the form of intersections of cycle paths with carriageways with traffic organisation that always gives priority to drivers, introduced in the city of Rybnik in Poland. The aim of the study is to analyse the behaviour of drivers at the intersection of the cycle path and the carriageway with signage indicating car priority in a situation when simultaneously a driver was coming along the carriageway and a cyclist was preparing to cross and crossed the intersection. The study took the form of a trial in which two cyclists—a man and a woman—negotiated the designated intersection in two directions. Drivers’ decisions to give or not to give priority were analysed in the context of the cyclist’s gender and direction of travel, as well as observable driver characteristics such as gender and residence based on the vehicle registration number. The study confirms that a significant number of drivers give priority to cyclists despite the rules at the intersection under study. However, no significant differences were found between the behaviour of female and male drivers. The study found that only one variable—place of vehicle registration—had a statistically significant relationship with giving priority to cyclists, contrary to the variables defining the gender of drivers and cyclists.

Pedestrian Safety in Road Traffic in the Era of the SARS-CoV-2 Pandemic in the Example of Szczecin

Safety of road users is one of the priority tasks of state institutions and local government authorities. Compared to other EU member states, Poland is characterized by a low level of road safety. According to the World Health Organization (WHO), more than half of road fatalities affect vulnerable road users (VRU). Most interactions in the vehicle–VRU relationship occur in built-up areas. Limiting the public’s mobility related to the spread of the COVID-19 virus led to changes in road traffic, which was reflected in the level of safety of individual road users. As part of a clear emphasis on the issue of VRU protection, the aim of this article is to examine the level of pedestrian safety in urban space at the time of restrictions on the public’s mobility in Poland during the SARS-CoV-2 pandemic and after lifting the limitations, based on the example of Szczecin. The authors put forward two research hypotheses: (1) the level of road safety is influenced by the intensity and saturation of the transport network; (2) the intensity of road traffic determines the occurrence of a greater number of road accident consequences. This study analyses traffic accident data over three periods—2017–2018, 2019–2020, and 2021—to identify pedestrian safety trends. The research shows that the intensity of road traffic has a significant impact on the occurrence of road accidents. The reduced road traffic due to the movement restrictions has kept the severity of road accidents moderate. In the post-pandemic period and the gradual lifting of the mobility restrictions, the number of road accidents increased while their consequences decreased.

Identification of Solutions for Vulnerable Road Users Safety in Urban Transport Systems: Grounded Theory Research

The share of road vehicles in urban transport systems is a challenge for modern and dynamically developing urban areas in accordance with the concept of sustainable and Smart Cities. Increasingly, there is a need to promote and adapt urban space to the movement of vulnerable road users (VRU). As part of a clear emphasis on the issue of pedestrians and other vulnerable road users, the aim of the article is to define the typology and hierarchy of solutions contributing to the increase in VRU safety in cities. The research process was based on the use of grounded theory. In the adopted research methodology, the use of the Delphi method made it possible to identify the approach of various European cities to the use and implementation of technical, technological and organizational solutions affecting safety. The research made it possible to evaluate individual solutions in VRU safety management and to indicate a list of recommendations for improving security, taking into account the views of international experts. In addition, the results of this study may enrich the current literature, helping to understand the perception of solutions implemented in urban transport systems as a holistic set of interrelated elements supporting pedestrian safety and increasing their role in cities.

How can front crash prevention systems address more police-reported crashes in the United States?

Government and consumer-information organizations can motivate automakers to address additional crash types through front crash prevention (FCP) testing programs. This study examined the current state of crashes potentially relevant to current and future FCP systems to provide a roadmap for the next crash types that vehicle testing programs in the United States should evaluate.Crash records from 2016 to 2020 were extracted from the Crash Report Sampling System (CRSS) and the Fatality Analysis Reporting System (FARS). Crashes were restricted to ones involving no more than two vehicles where the striking or path-intruding vehicle was a passenger vehicle and a vehicle defect was not coded. Percentages of police-reported crashes, nonfatal-injury crashes, and fatal crashes were computed for different crash types and circumstances.Rear-end and pedestrian crashes evaluated in existing FCP testing programs accounted for 27% of all police-reported crashes, 19% of nonfatal-injury crashes, and 18% of fatal crashes. The remaining crash types relevant to FCP accounted for 25% of police-reported crashes, 31% of nonfatal-injury crashes, and 23% of fatal crashes. A turning passenger vehicle crossing the path of an oncoming vehicle accounted for the largest proportion of the remaining police-reported (8%) and nonfatal-injury crashes (13%). Head-on crashes accounted for the largest proportion of remaining fatal crashes (9%). Most FCP-relevant police-reported crashes occurred on roads with a posted speed limit between 30 and 50 mph. Medium/heavy trucks were the crash partner in a disproportionate number of fatal head-on and rear-end crashes and motorcycles in a disproportionate number of fatal rear-end and turning crossing-path crashes. Fatal bicyclist and pedestrian crashes were overrepresented at night.The findings from this study indicate that testing organizations should evaluate FCP performance at higher speeds; with non-passenger vehicles and vulnerable road users; during the night; and in more complex head-on and turning crash scenarios to reduce crashes of all severities. Some of these conditions are currently assessed by other testing organizations and can be readily adopted by U.S. programs or possibly addressed with new approaches like virtual testing.

Vulnerable Road User Trajectory Prediction for Autonomous Driving Using a Data-Driven Integrated Approach

In this paper, Vulnerable Road User (VRU) trajectory prediction for autonomous driving based on the Intention-Attention-Gate Recurrent Unit (IA-GRU), Improved Social Force Model (ISFM) and Adaptive Boosting (AdaBoost) is systematically investigated. Firstly, a novel IA-GRU is proposed for VRU (pedestrian, cyclist, and electric cyclist) trajectory prediction. VRU intention (waiting/crossing), VRU heterogeneity (age and gender), VRU-VRU interactions and VRU-dynamic vehicle interactions are taken into account. Attention is used to obtain the influence weights of the above factors used for VRU trajectory prediction. Secondly, a micro-dynamic ISFM is developed for VRU trajectory prediction. The impact of zebra crossing, collision avoidance with vehicles and VRUs, and VRU heterogeneity are considered. Moreover, traffic data collected by an unmanned aerial vehicle (UAV) is obtained and analyzed, and the parameters of the ISFM are calibrated by the Maximum Likelihood Estimation (MLE). Finally, a data-driven integrated approach based on the IA-GRU and ISFM is proposed, and AdaBoost is used to prevent the model from overfitting and improve the prediction accuracy. The results indicate that the integrated model outperforms the existing methods, and the prediction accuracy is improved by more than 11% based on the collected traffic data, which can give us great confidence to use the integrated model in the autonomous driving domain to improve the safety of VRUs.

DECENT: Decentralized and Efficient Key Management to Secure Communication in Dense and Dynamic Environments

Intelligent Transportation Systems (ITS), one aspect of the Smart City paradigm, aim to improve the efficiency, convenience, and safety of travelers. The integration of (vehicular) communication technologies allows communication between the on-board communication units (OBUs) of vehicles, roadside units (RSUs), and vulnerable road users (VRUs), and contribute to the efficacy of ITS applications. However, these additional sources of information must be reliable and accurate. Security primitives such as confidentiality, integrity, and authenticity are required, but only achievable when supported with a suitable cryptographic key management scheme. This paper presents the design of a decentralized and efficient key management scheme, abbreviated as the DECENT scheme. This scheme provides secure multi-hop communication in dense and dynamic network environments while functioning in a self-organized manner. Through threshold secret sharing techniques, network nodes act as a distributed trusted third party (TTP) such that a threshold number of nodes can collaborate to execute key management functions. These functions include decentralized node admission and key updating. Novelties include (i) the unique self-healing characteristic, meaning that DECENT is capable of independently recovering from network compromise, and (ii) guidelines for choosing an appropriate security threshold in any deployment scenario which maximizes the level of security while simultaneously guaranteeing that decentralized key management services can be provided.

Deep RNN Based Prediction of Driver’s Intended Movements at Intersection Using Cooperative Awareness Messages

This paper presents an early prediction framework to classify drivers’ intended intersection movements in a connected vehicle environment. Intersections are considered accident blackspots with major traffic violations that cause property damage, injuries and fatalities. An accurate perception of drivers’ intended movements at intersections is required for advanced red-light (ARLW) or turning warnings for vulnerable road users (TWVR). Early prediction of intersection movement and adequate warning assistance will ensure road users’ safety at the intersection. In this study, we adopted recurrent neural networks (RNN): long short-term memory (LSTM) and gated recurrent units (GRU) networks to predict driver intended movements at intersections using the vehicle kinematics extracted from the Cooperative Awareness Messages (CAMs). We used naturalistic driving data of the Ipswich Connected Vehicle Pilot (ICVP) project, Queensland, which was collected from 351 participants who drove their connected vehicles during the pilot period. The pilot study installed roadside equipment at 29 signalised intersections to enable the Cooperative Intelligent Transportation System (C-ITS) use cases. Vehicle speed, speed limit, longitudinal acceleration, lateral acceleration, and yaw rate are used as predictors and monitored in 100-millisecond intervals for 1s to 4s at different warning distances from the stop line. Separate prediction models are trained based on different monitoring windows. Furthermore, drivers’ intended intersection movements are predicted at two individual intersections to evaluate intersection-specific prediction performance and are found with improved prediction accuracy than overall prediction models trained with all 29 intersections data. Overall prediction models are useful for some intersections which lack available data for individual intersection-based prediction.