Road user hazard perception tests: A systematic review of current methodologies

The efficacy of hazard perception training and education: A systematic review and meta-analysis

BACKGROUND

Hazard perception (HP) has been argued to improve with experience, with numerous training programs having been developed in an attempt to fast track the development of this critical safety skill. To date, there has been little synthesis of these methods.

OBJECTIVE

The present study sought to synthesise the literature for all road users to capture the breadth of methodologies and intervention types, and quantify their efficacy.

DATA SOURCES

A systematic review of both peer reviewed and non-peer-reviewed literature was completed. A total of 57 papers were found to have met inclusion criteria.

RESULTS

Research into hazard perception has focused primarily on drivers (with 42 studies), with a limited number of studies focusing on vulnerable road users, including motorcyclists (3 studies), cyclists (7 studies) and pedestrians (5 studies). Training was found to have a large significant effect on improving hazard perception skills for drivers (g = 0.78) and cyclists (g = 0.97), a moderate effect for pedestrians (g = 0.64) and small effect for motorcyclists (g = 0.42). There was considerable heterogeneity in the findings, with the efficacy of training varying as a function of the hazard perception skill being measured, the type of training enacted (active, passive or combined) and the number of sessions of training (single or multiple). Active training and single sessions were found to yield more consistent significant improvements in hazard perception.

CONCLUSIONS

This study found that HP training improved HP skill across all road user groups with generally moderate to large effects identified. HP training should employ a training method that actively engages the participants in the training task. Preliminary results suggest that a single session of training may be sufficient to improve HP skill however more research is needed into the delivery of these single sessions and long-term retention. Further research is also required to determine whether improvements in early-stage skills translate to improvements in responses on the road, and the long-term retention of the skills developed through training.

The effect of rear bicycle light configurations on drivers' perception of cyclists' presence and proximity

The optimal cycle light configuration for maximizing cyclists' conspicuity to drivers is not clear. Advances in sensor technology has led to the development of 'reactive' cycle lights that detect changes in the environment and consequently increase their flashing speed and brightness in risky situations - for example, when a rearward car is approaching - but no research has examined the effect of such lights on driver perception. The aim of the present study is to compare different cycle light configurations, including 'reactive' light technology, on drivers' ability to detect cyclists and estimate their proximity. We recruited 32 drivers to participate in two experiments, in which they viewed life-size real-world stimuli filmed from a driver's perspective in daytime and at dusk. The footage showed a cyclist on a bicycle with a rear light mounted on the seat post, in various configurations: static light, steady flashing, reactive flashing and no light. In Experiment 1, the drivers were required to detect the presence or absence of a cyclist on the road ahead as quickly as possible. In Experiment 2, they were required to estimate the distance of the cyclist from their vehicle, and to rate their confidence in their estimates. Experiment 1 revealed that drivers were quicker to detect the cyclist's presence in all rear cycle light conditions relative to the no light condition, but there were no differences in speed or accuracy across rear light conditions. Experiment 2 showed that drivers were more accurate in estimating the cyclist's proximity in the steady flashing and reactive flashing conditions, compared to static and no light conditions. Drivers were also more confident in their judgements in all rear light conditions compared to the no light condition. In conclusion, flashing rear cycle lights, regardless of reactive technology, enhanced drivers' perception of a cyclist ahead, notably in terms of their judgements of distance to that cyclist. Further investigation is needed to fully understand the impact of cycle light technology on driver perception, as well as the use of drivers' distance-to-cyclist estimates as an index of cyclists' cognitive conspicuity.

The relationship between central and mid-peripheral motion perception and the hazard perception test in younger and older adults

INTRODUCTION

Vision standards for driving are typically based on visual acuity, despite evidence that it is a poor predictor of driving safety and performance. However, visual motion perception is potentially relevant for driving, as the vehicle and surroundings are in motion. This study explored whether tests of central and mid-peripheral motion perception better predict performance on a hazard perception test (HPT), which is related to driving performance and crash risk, than visual acuity. Additionally, we explored whether age influences these associations, as healthy ageing impairs performance on some motion sensitivity tests.

METHODS

Sixty-five visually healthy drivers (35 younger, mean age: 25.5; SD 4.3 years; 30 older adults, mean age: 71.0; SD 5.4 years) underwent a computer-based HPT, plus four different motion sensitivity tests both centrally and at 15° eccentricity. Motion tests included minimum displacement to identify motion direction (Dmin ), contrast detection threshold for a drifting Gabor (motion contrast), coherence threshold for a translational global motion stimulus and direction discrimination for a biological motion stimulus in the presence of noise.

RESULTS

Overall, HPT reaction times were not significantly different between age groups (p = 0.40) nor were maximum HPT reaction times (p = 0.34). HPT response time was associated with motion contrast and Dmin centrally (r = 0.30, p = 0.02 and r = 0.28, p = 0.02, respectively) and with Dmin peripherally (r = 0.34, p = 0.005); these associations were not affected by age group. There was no significant association between binocular visual acuity and HPT response times (r = 0.02, p = 0.29).

CONCLUSIONS

Some measures of motion sensitivity in central and mid-peripheral vision were associated with HPT response times, whereas binocular visual acuity was not. Peripheral testing did not show an advantage over central testing for visually healthy older drivers. Our findings add to the growing body of evidence that the ability to detect small motion changes may have potential to identify unsafe road users.

The Effectiveness of a Digital Game-Based Intervention on Hazard Perception and Visual Skills in Novice Drivers: A Single Blind, Randomized Controlled Trial

Novice drivers show poorer performance than experienced drivers in terms of visual skills and hazard perception. This study aimed at evaluating the effectiveness of a digital game-based intervention on hazard perception and visual skills in novice drivers. Forty-six novice drivers (6 men, 40 women) were randomized to the intervention group (n = 23; 20.79 ± 0.81 years) or control (n = 23; 20.65 ± 0.93 years) group. The intervention group received a game-based intervention in addition to a hazard perception training, whereas the control group received only the hazard perception training. Hazard perception and visual skills were assessed in both groups before and after the 14-day interventions. Between-group comparisons revealed significantly greater improvements in visual short time memory, visual closure, visual discrimination, figure-ground and total scores in the game-based group than in the control group (p < 0.05 for all). Our results showed that 14 days of game-based intervention enhanced hazard perception and visual skills in novice drivers. Using game-based interventions in driving rehabilitation is recommended to improve hazard perception and visual skills of novice drivers.

Road crashes in Adelaide metropolitan region, the consequences of COVID-19

Background

Many countries instituted lockdown rules as the COVID-19 pandemic progressed, however, the effects of COVID-19 on transportation safety vary widely across countries and regions. In several situations, it has been shown that although the COVID-19 closure has decreased average traffic flow, it has also led to an increase in speeding, which will indeed increase the severity of crashes and the number of fatalities and serious injuries.

Methods

At the local level, Generalized linear Mixed (GLM) modelling is used to look at how often road crashes changed in the Adelaide metropolitan area before and after the COVID-19 pandemic. The Geographically Weighted Generalized Linear Model (GWGLM) is also used to explore how the association between the number of crashes and the factors that explain them varies across census blocks. Using both no-spatial and spatial models, the effects of urban structure elements like land use mix, road network design, distance to CBD, and proximity to public transit on the frequency of crashes at the local level were studied.

Results

This research showed that lockdown orders led to a mild reduction (approximately 7%) in crash frequency. However, this decrease, which has occurred mostly during the first three months of the lockdown, has not systematically alleviated traffic safety risks in the Greater Adelaide Metropolitan Area. Crash hotspots shifted from areas adjacent to workplaces and education centres to green spaces and city fringes, while crash incidence periods switched from weekdays to weekends and winter to summer.

Implications

The outcomes of this research provided insights into the impact of shifting driving behaviour on safety during disorderly catastrophes such as COVID-19.

Is safety in the eye of the beholder? Discrepancies between self-reported and proxied data on road safety behaviors—A systematic review

Recent studies have problematized on the lack of agreement between self-reported and proxied data in the field of road safety-related behaviors. Overall, and although these studies are still scarce, most of them suggest that the way we perceive our own road behavior is systematically different from the perspective from which we perceive others' behavior, and vice versa. The aim of this review paper was to target the number and type of studies that have researched the behavioral perceptions of different groups of road users, contrasting self-reported behavioral data with those reported by other users (proxied), and their outcomes. This systematic review followed the PRISMA methodology, which allows for the identification of relevant articles based on the research term. A total number of 222 indexed articles were filtered, and a final selection of 19 articles directly addressing the issue was obtained. Search strategies were developed and conducted in MEDLINE, WOS, Scopus and APA databases. It is remarkable how road users perceive themselves as behaviorally “safer” than the rest of road users in what concerns the knowledge of traffic norms and their on-road performance. In addition, and regardless of the type of user used as a source, self-reported data suggest their perceived likelihood to suffer a traffic crash is lesser if compared to any other user. On the other hand, proxied reports tend to undervalue third users' performance, and to perceive riskier behaviors and crash-related risks among them. The outputs of this systematic review support the idea that the perception of road users' behavior and its related risks substantially differ according to the source. It is also necessary to increase the number, coverage and rigor of studies on this matter, perhaps through complementary and mixed measures, in order to properly understand and face the bias on road users' risk-related behaviors.

Concept Mapping to Identify Content for a Performance-Based Measure of Low Luminance Vision-Related Activities of Daily Living

Purpose

The purpose of this study was to identify low luminance activities of daily living (ADL) relevant to adults with vision impairment using a concept-mapping approach.

Methods

“Group concept mapping” was utilized to identify specific ADLs that persons with vision impairment find challenging under low light conditions. In the first “brainstorming” phase, 24 adults with vision impairment from a range of eye conditions (mean age = 73 years, SD = 14 years) and 26 international low vision experts (mean experience = 22, SD = 11 years) generated statements to the focus prompt, “Thinking as broadly as possible, generate a list of statements detailing specific day-to-day activities a person with vision impairment might find challenging under low light conditions, such as in a poorly lit room or outside at dusk.” In the second phase, participants sorted activities by similarity and rated the importance of each activity. Multidimensional scaling and hierarchical cluster analysis were applied to produce concept maps showing clusters of prioritized activities.

Results

One hundred thirteen unique ideas/activities were generated, rated and sorted. Eight clusters were identified (from highest to lowest importance): hazard detection and safety outside; social interactions; navigation; near reading; selfcare and safety at home; distance spotting; searching around the home; and cooking and cleaning.

Conclusions

The conceptual framework and low luminance ADLs identified (the most important being hazard detection and safety outside, and social interactions) provide a basis for developing a performance-based measure of low luminance visual function.

Translational Relevance

A performance-based measure of low luminance vision-related ADLs is required for comprehensively and objectively assessing efficacy of eye treatments and low vision rehabilitation outcomes in adults with vision impairment.

Does music affect performance on a hazard perception test for cyclists?

Whereas it has been shown that listening to music impairs the detection of auditory and visual signals, it is unclear to what extent music affects a cyclist's ability to detect and interpret hazardous traffic situations. In the current experiment, thirty-seven participants carried out a hazard perception test for cyclists. Participants were divided into three groups: control, passive, or active. The control group did the test without hearing music. The passive and active group did hear music, yet the passive group was asked to ignore the music, while the active group was asked to pay attention to the lyrics. Results showed no differences in reaction rate, reaction time, or gaze behaviour between any of the groups. These findings temper the existing safety concerns about the negative effect of music on traffic safety. Nevertheless, music might still have consequences under certain conditions or in certain risk-groups such as children. Practitioner summary: It is unclear how music affects traffic safety. The current experiment tested to what extent hazard perception was affected by listening actively or passively to music. Under the current experimental conditions, listening to music was found to have no effect on hazard perception. Abbreviation: NHTSA: national highway traffic safety administration; AOI: area of interest; TTFS: time to first saccade; FFD: first fixation duration.

Correlations among self-report, static image, and video-based hazard perception assessments: The validity of a new Lithuanian hazard prediction test

Scholars suggest that improving hazard perception (HP) skills among different road users can lead to safer driving. To this end, a small number of countries have developed a national hazard perception test for their licensing procedures. The purpose of the current research is to develop and validate a new video-based hazard prediction test for the Lithuanian driving context. Eighty-eight drivers participated in the study. Initially, 25 naturalistic clips were included and edited into a hazard prediction test following the "what happens next?" paradigm. The overall score of the test was higher for experienced drivers when compared with novices. On further inspection the overall group difference was driven by 12 clips. The Lithuanian Hazard Perception test HPT-LIT featuring static images (Tūskė et al., 2019) and Hazard Perception Questionnaire developed by White et al. (2011) were used to validate the new test. This new version of the test (12 clips, LHP₁₂) showed acceptable psychometric properties. Drivers with less driving experience demonstrated poorer hazard prediction skills than experienced drivers. The results revealed a negative correlation between hazard prediction scores on LHP₁₂ and crash involvement in the novice driver group, as well as a positive correlation between prediction scores and self-reported hazard prediction skills. It was concluded that the new Lithuanian hazard test shows potential for future research.

Detecting driver stress and hazard anticipation using real-time cardiac measurement: A simulator study

OBJECTIVES

In the context of growing interest in real-time driver stress detection systems, we question the value of using heart rate change over short time periods to detect driver stress and hazard anticipation.

METHODS

To this end, we explored changes in heart rate and speed as well as perceived stress in 27 drivers in a driving simulator. Driver stress was triggered by using hazardous road events, while hazard anticipation was manipulated using three levels of hazard predictability: unpredictable (U), predictable (P), and predictable and familiar (PF).

RESULTS

The main results indicate that using heart rate change (1) is a good indicator for detecting driver stress in real time, (2) provides a cardiac signature of hazard anticipation, and (3) was affected by perceived stress groups. Further investigation is needed to validate the lack of relationship between increased anticipation/predictability and strengthened cardiac signature.

CONCLUSIONS

These results support the use of heart rate change as an indicator of real-time driver stress and hazard anticipation.

A comparison of cybersickness symptoms across 360-degree hazard perception and hazard prediction tests for drivers

Hazard perception assessment may benefit from VR-presentation by removing field-of-view restrictions imposed by single-screen tests. One concern is whether VR-induced 'cybersickness' will offset any benefits. Self-reported cybersickness ratings were recorded from 77 participants viewing two variants of a 360-degree hazard test: hazard perception and hazard prediction. The latter was hypothesised to be particularly susceptible as clips abruptly cut to a probe question at hazard onset. Such sudden occlusions are thought to increase cybersickness. Overall cybersickness levels were low, with only four participants excluded for above-threshold sickness ratings. The remaining participants showed unexpectedly lower symptoms for the hazard prediction test and rated this test format as more comfortable and engaging. These findings mitigate concerns over the use of 360-degree videos in formative hazard assessments, even when clips involve sudden occlusions. Nonetheless, removal of any participants due to cybersickness raises problems for using VR for formal assessments of hazard perception skill.

Analyzing drivers' hazard recognition: Precursors to single-vehicle collisions

Extensive driver behavior and performance information provided by real-world video surveillance and sensor data in the SHRP2 Naturalistic Driving Study has enabled the examination of new layers and pathways leading to crash outcomes. We note that the prominence of hazards and the importance of recognizing them vary systematically across single vs. multi-vehicle crashes, and address a fundamental question about safety: why do around three-quarters of drivers involved in single-vehicle crashes not recognize, perceive, or react to the precipitating event (PE)? Using a path-analytic framework through marginal effects, this study investigates factors correlated to recognition of the PE in single-vehicle events, and how these correlations may act as crash precursors. Logit models, accounting for heterogeneity among events and drivers by estimating both fixed and random parameters, quantified correlations among key variables, given a crash or near-crash event (N = 543). The type of PE, roadway environment factors, and driving maneuvers heavily influenced recognition chances. Drivers had a harder time recognizing less conspicuous hazards (e.g. departing the travel way, decreased recognition chances by 48.29%), but seemed better at recognizing prominent hazards (e.g. vehicle losing control, increased recognition chances by 46.71%). In addition, drivers are less likely to recognize PEs when executing less involved driving maneuvers in more relaxed environments, such as daylight (decreased recognition chances by 16.00%), but are more adept in environments that already demand more attention. Recognition reduced the chances of a crash by 12.23%, so we found similar correlations with crash outcome. Future intelligent transportation systems may focus on increasing driver recognition of potential hazards by bringing attention to less conspicuous hazards and less involved driving environments and actions.

Hazard Perception, Presence, and Simulation Sickness-A Comparison of Desktop and Head-Mounted Display for Driving Simulation

Driving simulators are becoming increasingly common in driver training and assessment. Since virtual reality is generally regarded as an appropriate environment for measuring risk behavior, simulators are also used to assess hazard perception, which is considered to be one of the most important skills for safe driving. Simulators, which offer challenges that are indeed comparable to driving in real traffic, but at a very low risk of physical injury, have the potential to complement theoretical and practical driver trainings and tests. Although configurations and fidelity differ considerably between driving simulators, studies comparing the impact of their distinct features on driving performance and test validity remain rare. In this context, prior research demonstrated that a wider field of view (three monitors compared to a single monitor) led to earlier speed adjustments in response to potential hazards-especially for experienced drivers. The wider field of view was assumed to cause the drivers to be more present in the virtual world, which in turn provoked more natural scanning of the road and therefore, earlier hazard detection in experienced drivers. Research on spatial presence in other contexts support this assumption. The present experiment investigated whether this effect could be enhanced by an even more immersive presentation technique for driving simulation: a head-mounted display (HMD). Moreover, we studied the interplay between display mode, sense of presence and simulation sickness. Eighty experienced and less experienced drivers completed six simulation-based hazard perception scenarios, which were displayed either via a triple-monitor set-up or an HMD. Results indicate that the experienced drivers showed very similar driving and risk behavior as the inexperienced drivers in both experimental conditions. However, there were significant differences between the two display conditions. The use of an HMD resulted in a clearer and more abrupt speed reduction, more virtual presence, and a higher degree of simulation sickness. However, the interrelation between these three variables could not be conclusively clarified in the present study and thus represents a research aim that could be addressed in future studies.

A novel driving assessment combining hazard perception, hazard prediction and theory questions

A new hazard test was created using high-fidelity computer animation containing ten hazards. Sixty learner drivers and sixty experienced drivers sat either a hazard-perception version of this test (requiring timed responses to materialized hazards) or a hazard-prediction variant of the test (where the screen is occluded as the hazard begins to appear and drivers are asked 'What happens next?'). Recent studies have demonstrated that the prediction test format outperforms the hazard perception format using naturalistic video, but there has not yet been a study replicating this effect with computer-animated materials similar to the quality of those used in the official UK hazard perception test. The new test also included eleven theory questions designed to probe drivers' knowledge of the rules of the road. The results demonstrated that both test variants differentiated between driver groups with considerable effect sizes. Theory-question scores were comparable across learner and experienced driver groups, reflecting learners' preparation for the test and possible issues with memory decay and overwriting in the experienced group. As an interesting aside, driving-related video game play negatively correlated with hazard perception performance, but not with hazard prediction scores. Some individual hazards better suited the prediction or perception test format, raising the possibility of a future hybrid test that combines the two approaches.

A cross-cultural comparison of visual search strategies and response times in road hazard perception testing

Road hazard perception is considered the most prominent higher-order cognitive skill related to traffic-accident involvement. Regional cultures and social rules that govern acceptable behavior may influence drivers' interpretation of a traffic situation and, consequently, the correct identification of potentially hazardous situations. Here, we aimed to compare hazard perception skills among four European countries that differ in their traffic culture, policies to reduce traffic risks, and fatal crashes: Ukraine, Italy, Spain, and Sweden. We developed a static hazard perception test in which driving scenes with different levels of braking affordance were presented while drivers' gaze was recorded. The test required drivers to indicate the action they would undertake: to brake vs. to keep driving. We assessed 218 young adult drivers. Multilevel models revealed that the scenes' levels of braking affordance (i.e., road hazard) modulated drivers' behavior. As the levels of braking affordance increased, drivers' responses became faster and their gaze entropy decreased (i.e., visual search strategy became less erratic). The country of origin influenced these effects. Ukrainian drivers were the fastest and Swedish drivers were the slowest to respond. For all countries, the decrement in response times was less marked in the case of experienced drivers. Also, Spanish drivers showed the most structured (least erratic) visual search strategy, whereas the Italians had the most rigid (most constant) one. These results suggest that road hazard perception can be defined cross-culturally, with cultural factors (e.g., traffic climate, legislation) modulating response times and visual search strategies. Our results also support the idea that a multimodal assessment methodology is possible for mass testing of road hazard perception and its outcomes would be relevant to understand how different traffic cultures shape driving behavior.

Fifty Years of Accident Analysis & Prevention: A Bibliometric and Scientometric Overview

Accident Analysis & Prevention (AA&P) is a leading academic journal established in 1969 that serves as an important scientific communication platform for road safety studies. To celebrate its 50^th anniversary of publishing outstanding and insightful studies, a multi-dimensional statistical and visualized analysis of the AA&P publications between 1969 and 2018 was performed using the Web of Science (WoS) Core Collection database, bibliometrics and mapping-knowledge-domain (MKD) analytical methods, and scientometric tools. It was shown that the annual number of AA&P's publications has grown exponentially and that over the course of its development, AA&P has been a leader in the field of road safety, both in terms of innovation and dissemination. By determining its key source countries and organizations, core authors, highly co-cited published documents, and high burst-strength publications, we showed that AA&P's areas of focus include the "effects of hazard and risk perception on driving behavior", "crash frequency modeling analysis", "intentional driving violations and aberrant driving behavior", "epidemiology, assessment and prevention of road traffic injuries", and "crash-injury severity modeling analysis". Furthermore, the key burst papers that have played an important role in advancing research and guiding AA&P in new directions - particularly those in the fields of crash frequency and crash-injury severity modeling analyses were identified. Finally, a modified Haddon matrix in the era of intelligent, connected and autonomous transportation systems is proposed to provide new insights into the emerging generation of road safety studies.

Developing and testing a hazard prediction task for novice drivers: A novel application of naturalistic driving videos

INTRODUCTION

Novice drivers' inability to appropriately anticipate and respond to hazards has been implicated in their elevated crash risk. Our goal was to develop a driving hazard prediction task using naturalistic videos from the U.S. context that could distinguish between novice and experienced drivers.

METHOD

Using the query builder from the SHRP 2 InSight Data Access Website, we identified a sample of 1034 videos for further review. Task criteria reduced these to 30 videos of near-crash events that were split into event and non-event segments and were used to develop the driving hazard prediction task (task). Participants, aged 16-20 years-old (22 novice and 19 experienced drivers) completed the task during which they watched event and non-event videos and were asked, "How likely was the driver of this car to get into a crash?" after each video. Overall ratings for hazardousness were calculated for experienced and novice drivers as well as a group difference score for hazardousness.

RESULTS

All participants rated event videos as more hazardous than non-event videos, but there was no main effect of group. Rather, there was a significant EventbyGroup interaction in which there were no group differences in hazard ratings for non-event videos, but experienced drivers rated event videos as more hazardous than novice drivers. Specific characteristics of the event videos, such as the hazard development period, were related to differences between novice and experienced drivers' hazardousness ratings.

CONCLUSION

To the best of our knowledge, this is the first use of naturalistic driving videos from an existing database as experimental stimuli. We found that the task discriminated between novice and experienced drivers' ratings of hazardousness. This distinction suggests naturalistic driving videos may be viable stimuli for experimental studies. Practical Applications: The application of naturalistic driving video database for experimental research may hold promise.