Development of German pedelec (and bicycle) crashes between 2013 and 2021.

INTRODUCTION: The continuous growth in the use of e-bikes (in Germany mostly pedelecs that support pedaling up to 25 km/h) raises questions about the use of historic crash data for the development of road safety measures. The aim of this study was to address this issue, by conducting a longitudinal analysis of pedelec and bicycle crash data over a period of nine years to identify trends and to clarify whether such trends are specific to pedelecs. METHOD: We analyzed 95,338 police reported pedelec and bicycle injury crashes from 2013 through 2021. The dataset consisted of crashes from three federal states of Germany: Brandenburg, Hesse and Saxony. Data were analyzed with respect to sex and age distribution, time, location and type of accident, conflict partner, cause of crash and injury severity. RESULTS: Many of the analyzed variables showed a considerable degree of temporal stability, with differences as well as similarities between the two bicycle types staying quite consistent over the years. One notable difference was the mean age of the involved riders, with crashed pedelec riders being significantly older than conventional cyclists. At the same time, however, the mean age of these pedelec riders has decreased by eight years over time. Single vehicle crashes were consistently more common for pedelec riders than for cyclists. Similarly, pedelec rider crashes went with a higher injury severity over all the years. CONCLUSIONS AND PRACTICAL APPLICATIONS: While, on a more detailed level, we found differences between the two bicycle types, overall crash characteristics were remarkably similar and consistent over time. Our findings provide no clear argument for road safety measures that are specifically designed to target pedelecs. Instead, the stable crash total, and the increases in ridership of both bicycles and pedelecs, highlight the demand for new, innovative solutions to improve cycling safety in general.

The interaction between perceived safety and perceived usefulness in automated parking as a result of safety distance.

Improved safety and traffic efficiency are among the proclaimed benefits of automated driving functions. In many scenarios, traffic safety and efficiency can be somewhat contradictory, especially in the perception of a user. In order for potential users to accept the automated system, it is necessary to find the optimal system configuration. Therefore, it is important to understand how the factors underlying acceptance develop and interact. In this study, seven safety distances of an automated parking system were implemented resulting in parking manoeuvres of varying efficiency (in terms of required moves). Participants experienced each configuration twice and rated their perceived safety and perceived usefulness. The results show that maximizing safety distances results in high perceived safety, yet also a diminished perceived usefulness due to reduced efficiency. On the other hand, maximum efficiency leads to a lower perceived safety and thus, a reduced rating of perceived usefulness. Furthermore, in some participants, perceived safety increased gradually, while for others, a threshold effect could be observed. The results demonstrate that the specification of a sole system characteristic can have multiple effects. These have to be considered to maximize acceptance.

Cross-cultural differences in the acceptance of decisions of automated vehicles.

Automated vehicles are expected to enhance driving safety and comfort. In order to fulfil these expectations, they have to be widely accepted and used. Implementing an acceptable driving style is therefore a must. Previous research on automated vehicle acceptance has largely concentrated on the effects of driving dynamics. This study takes a different approach and focuses on the effects of the driving decisions. To assess the effects of driving decisions on acceptance, an online experimental study was conducted in China, Germany, Japan and the US. Four overtaking scenarios, in which the automated vehicle took a decision, were presented as short texts. The situations differed with regard to the action (overtaking vs. stay in lane) and potential consequence (high or low hindrance of another driver). Participants then rated their acceptance. The results indicate that acceptance is dependent on the driving decisions and is further influenced by cultural background. Chinese drivers show high acceptance to the decisions and there were no significant differences between the presented scenarios. In the US and Germany, decisions leading to high hindrance of others are rejected, whereas in cases of low hindrance, overtaking is preferred. Japanese participants reject decisions, which lead to hindrance of others.

Impact of SARS-CoV-2 on the mobility behaviour in Germany.

Background: The COVID-19 pandemic and the measures taken to combat it led to severe constraints for various areas of life, including mobility. To study the effects of this disruptive situation on the mobility behaviour of entire subgroups, and how they shape their mobility in reaction to the special circumstances, can help to better understand, how people react to external changes. Methodology: Aim of the study presented in this article was to investigate to what extent, how and in what areas mobility behaviour has changed during the outbreak of SARS-CoV-2 in Germany. In addition, a focus was put on the comparison of federal states with and without lockdown in order to investigate a possible contribution of this measure to changes in mobility. We asked respondents via an online survey about their trip purposes and trip frequency, their choice of transport mode and the reasons for choosing it in the context of the COVID-19 crisis. For the analyses presented in this paper, we used the data of 4157survey participants (2512 without lockdown, 1645 with lockdown). Results: The data confirmed a profound impact on the mobility behaviour with a shift away from public transport and increases in car usage, walking and cycling. Comparisons of federal states with and without lockdown revealed only isolated differences. It seems that, even if the lockdown had some minor effects, its role in the observed behavioural changes was minimal.

Using European naturalistic driving data to assess secondary task engagement when stopped at a red light.

PROBLEM: Some evidence exists that drivers choose to engage in secondary tasks when the driving demand is low (e.g., when the car is stopped). While such a behavior might generally be considered as rather safe, it could be argued that the associated diversion of attention away from the road still leads to a reduction of situational awareness, which might increase collision risk once the car regains motion. This is especially relevant for texting, which is associated with considerable eyes-off-the-road-time. Nonetheless, it seems that previous research has barely addressed the actual engagement in secondary tasks while waiting at a red light (as compared to just addressing the tasks' mere prevalence). OBJECTIVE: The present study investigated secondary task engagement while stopped at a red light using European naturalistic driving data collected through the UDRIVE project. Attention was given to the whole engagement process, including simple prevalence and the tasks' relation (in terms of start/end) to the red light period. Moreover, given that texting is one of the most problematic forms of distraction, it was characterized in more detail regarding glance behavior. METHOD: Videos of 804 red light episodes from 159 drivers were annotated. Glance behavior was also coded for a sub-set of 75 texting events and their matched baselines. Results, conclusions and practical applications: Drivers engaged in at least one secondary task across almost half of the annotated red light episodes. Drivers who texted while stopped spent most of the time looking at their cell phone. Consequently, drivers might not have been prepared for potentially unexpected events once the light turned green. Further, drivers concluded texting a considerable number of times well after the red light period, which has potential implications for traffic safety.

Can a unique appearance of e-bikes, coupled with information on their characteristics, influence drivers' gap acceptance?

Objective: Car drivers tend to underestimate the speed of e-bikes and accept smaller gaps for crossing in front of them compared to conventional bicycles. As an explanation, it has been suggested that car drivers rely on their previous experience with conventional bicycles, which tells them that those mostly travel at low speeds. E-bikes, which look just like regular bicycles, do not conform to this expectation, resulting in potentially dangerous interactions. Based on this assumption, researchers have suggested to increase other road users' awareness of e-bikes' higher speeds by giving them a distinct appearance. The goal of our experiment was to investigate the effects of such a unique appearance, aided by clear instructions about the higher speeds of e-bikes, on gap acceptance.Method: In order to investigate the effect of appearance independent of the effect of bicycle type, we used video sequences of conventional bicycles and e-bikes approaching at different levels of speed. The riders (regardless of what type of bike they were actually riding) either wore an orange helmet as an indicator for an e-bike, or a gray helmet indicating a conventional bicycle. Fifty participants were asked to indicate the smallest acceptable gap for a left turn in front of the cyclist or e-bike rider.Results: The results showed significantly smaller acceptable gaps when confronted with the gray helmet (signal for bicycle) compared to the orange helmet (signal for e-bike), whereas there was no difference between the actual bicycle types.Conclusions: Overall, the results indicate that informing about e-bikes characteristics in combination with a unique appearance can lead to a more cautious behavior among car drivers.

(E-)Cyclists running the red light - The influence of bicycle type and infrastructure characteristics on red light violations.

Red light running is one of the most common traffic violations among cyclists. From different surveys, we know that about 40% of all cyclists run a red light at least occasionally. However, specific data on red light running of e-bike riders (pedelec and S-pedelec riders), a population of cyclists that has been growing steadily in the past few years in Germany and elsewhere, is largely missing. Similarly unclear is the role of the used infrastructure (e.g., carriageway or bike path) or the intersection type on the riders' propensity to run the red light. The goal of this study was to investigate the red light running behaviour of three different bicycle types (bicycle, pedelec, S-pedelec) in Germany, with specific focus on various infrastructure characteristics. We reanalysed data obtained in a naturalistic cycling study, in which we observed 90 participants riding their own bicycles (conventional bicycles, pedelecs, S-pedelecs) on their daily trips over four weeks each. The video material of these trips was annotated and analysed with regard to red light running. Overall, our participants experienced nearly 8000 red light situations. In 16.3% of these situations, they ran the red light, with nearly identical rates for cyclists, pedelec and S-pedelec riders. Red light running rates were lowest when cyclists rode on the carriageway, while the complexity of the intersection appeared to play a role as well. In general, red light running was more common when riders were about to turn right instead of turning left or riding straight through the intersection. Interestingly, we also observed a considerable number of cases in which the riders changed their used infrastructure (e.g., from the carriageway onto the pavement) to avoid a red light.

Are you an ambitious cyclist? Results of the cyclist profile questionnaire in Germany.

Objective: A number of studies have already grouped cyclists according to different aspects of their mobility behavior. This could be used e.g., to improve the bicycle infrastructure planning, to detect critical spots and, to reduce obstacles for cycling. This wide, preexisting, range of cyclist typologies usually concentrates on one or two influence factors and differs, content-wise, in both factors used, as well as, methodically. Based on existing cyclist typologies we extracted all possible influence factors to integrate them in one single questionnaire. The objective of this study, using an empirical, based approach, is to compare this typology of cyclists with existing ones, integrating all known influence factors of recent studies.Methods: To address these issues, we conducted a Germany-wide online survey on cycling behavior, covering all relevant aspects we derived from both literature and especially, former cyclist typology studies including: social factors; the impact of environmental, individual; and route factors; as well as motives. The main goal was to identify distinct types of cyclists, and describe them as detailed as possible. The heterogeneous sample included a total of 10,294 responses.Results: Using factor and cluster analyses, a multidimensional typology with four groups of cyclists was derived which were interpreted as: ambitious, functional, pragmatic, and passionate cyclists. In addition, socio-economic factors, cyclist's motivation, and crash history were analyzed.Conclusion: The results produced by grouping different characteristics of cyclists can lead to policy recommendations or communal bicycle traffic planning. Policy planners can estimate reactions of the different types on interventions and adjust their decisions which can serve to support already passionate cyclists or, encourage normally under-represented infrequent cyclists to cycle more. The extent of perceived safety plays here an important role in the classification, e.g., the handling of high-risk areas for crashes.

Risk compensation? The relationship between helmet use and cycling speed under naturalistic conditions.

INTRODUCTION: An argument against mandatory helmet use is based on the idea of risk compensation, which means that cyclists might ride faster when wearing a helmet (Lardelli-Claret et al., 2003). However, questionnaire and experimental studies were unable to find evidence for this assumption (Fyhri et al., 2012; Fyhri & Philipps, 2013). Simultaneously, other factors with a potential role in helmet use and cycling speed, such as trip length and rider characteristics have been neglected in such considerations. The goal of the analysis presented in this paper was therefore to investigate the relationship between helmet use and cycling speed under naturalistic conditions while taking characteristics of cyclists and bicycles into account. METHOD: As part of a naturalistic cycling study, we equipped the bicycles of conventional and e-bike riders with data acquisition systems to record speed and trip distance. It included two cameras (one for the face of the participant, another one for the forward scenery). For the analysis presented in this paper, we used the data of 76 participants (28 conventional bicycles, 48 e-bikes). RESULTS: In total, participants used their helmet for 56% of all trips. Helmets were used more frequently for longer trips. A linear mixed model, in which trip length, helmet use, bicycle type, age, and gender were used as predictors showed that helmet use did not play a significant role for cycling speed. Instead, all other factors that were analyzed, with the exception of gender, had a significant relationship to cycling speed. DISCUSSION: The assumption of risk compensation as a result of the use of a helmet could not be confirmed. Instead, the findings seem to support the suggestion that cyclists who undertake trips at potentially higher speed levels are aware of their increased risk, and actively try to reduce it through the use of a helmet.

Using SHRP 2 naturalistic driving data to assess drivers' speed choice while being engaged in different secondary tasks.

INTRODUCTION: The engagement in secondary tasks while driving has been found to result in considerable impairments of driving performance. Texting has especially been suspected to be associated with an increased crash risk. At the same time, there is evidence that drivers use various self-regulating strategies to compensate for the increased demands caused by secondary task engagement. One of the findings reported from multiple studies is a reduction in driving speed. However, most of these studies are of experimental nature and do not let the drivers decide for themselves to (not) engage in the secondary task, and therefore, eliminate other strategies of self-regulation (e.g., postponing the task). The goal of the present analysis was to investigate if secondary task engagement results in speed adjustment also under naturalistic conditions. METHOD: Our analysis relied on data of the SHRP 2 naturalistic driving study. To minimize the influence of potentially confounding factors on drivers' speed choice, we focused on episodes of free flow driving on interstates/highways. Driving speed was analyzed before, during, and after texting, smoking, eating, and adjusting/monitoring radio or climate control; in a total of 403 episodes. RESULTS: Data show some indication for speed adjustment for texting, especially when driving with high speed. However, the effect sizes were small and behavioral patterns varied considerably between drivers. The engagement in the other tasks did not influence drivers' speed behavior significantly. CONCLUSIONS AND PRACTICAL APPLICATIONS: While drivers might indeed reduce speed slightly to accommodate for secondary task engagement, other forms of adaptation (e.g., strategic decisions) might play a more important role in a natural driving environment. The use of naturalistic driving data to study drivers' self-regulatory behavior at an operational level has proven to be promising. Still, in order to obtain a comprehensive understanding about drivers' self-regulatory behavior, a mixed-method approach is required.

Size speed bias or size arrival effect-How judgments of vehicles' approach speed and time to arrival are influenced by the vehicles' size.

Crashes at railway level crossings are a key problem for railway operations. It has been suggested that a potential explanation for such crashes might lie in a so-called size speed bias, which describes the phenomenon that observers underestimate the speed of larger objects, such as aircraft or trains. While there is some evidence that this size speed bias indeed exists, it is somewhat at odds with another well researched phenomenon, the size arrival effect. When asked to judge the time it takes an approaching object to arrive at a predefined position (time to arrival, TTA), observers tend to provide lower estimates for larger objects. In that case, road users' crossing decisions when confronted with larger vehicles should be rather conservative, which has been confirmed in multiple studies on gap acceptance. The aim of the experiment reported in this paper was to clarify the relationship between size speed bias and size arrival effect. Employing a relative judgment task, both speed and TTA estimates were assessed for virtual depictions of a train and a truck, using a car as a reference to compare against. The results confirmed the size speed bias for the speed judgments, with both train and truck being perceived as travelling slower than the car. A comparable bias was also present in the TTA estimates for the truck. In contrast, no size arrival effect could be found for the train or the truck, neither in the speed nor the TTA judgments. This finding is inconsistent with the fact that crossing behaviour when confronted with larger vehicles appears to be consistently more conservative. This discrepancy might be interpreted as an indication that factors other than perceived speed or TTA play an important role for the differences in gap acceptance between different types of vehicles.

The influence of speed, cyclists' age, pedaling frequency, and observer age on observers' time to arrival judgments of approaching bicycles and e-bikes.

Given their potential to reach higher speed levels than conventional bicycles, the growing market share of e-bikes has been the reason for increased concerns regarding road safety. Previous studies have shown a clear relationship between object approach speed and an observers' judgment of when the object would reach a predefined position (i.e., time to arrival, TTA), with higher speed resulting in longer TTA estimates. Since TTA estimates have been linked to road users' decisions of whether or not to cross or turn in front of approaching vehicles, the higher potential speeds of e-bikes might result in an increased risk for traffic conflicts. The goal of the two experiments presented in this paper was to examine the influence of speed and a variety of other factors on TTA estimation for conventional bicycles and for e-bikes. In both experiments, participants from two age groups (20-45 years old and 65 years or older) watched video sequences of bicycles approaching at different speeds (15-25km/h) and were asked to judge the TTA at the moment the video was stopped. The results of both experiments showed that an increase in bicycle approach speed resulted in longer TTA estimates (measured as the proportion of estimated TTA relative to actual TTA) for both bicycle types (ηp(2)Exp.1=.489, ηp(2)Exp.2=.705). Compared to younger observers, older observers provided shorter estimates throughout (Exp. I: MDiff=0.35, CI [0.197, 0.509], ηp(2)=.332, Exp. II: MDiff=0.50, CI [.317, 0.682], ηp(2)=.420). In Experiment I, TTA estimates for the conventional bicycle were significantly shorter than for the e-bike (MDiff=0.03, CI [.007, 0.044], ηp(2)=.154), as were the estimates for the elder cyclist compared to the younger one (MDiff=0.05, CI [.025, 0.066], ηp(2)=.323). We hypothesized that the cause for this effect might lie in the seemingly reduced pedaling effort for the e-bike as a result of the motor assistance it provides. Experiment II was able to show that a high pedaling frequency indeed resulted in shorter TTA estimates compared to a low one (MDiff=0.07, CI [0.044, 0.092], ηp(2)=.438). Our findings suggest that both the e-bikes' potential to reach higher speeds and the fact that they reduce the perceived cycling effort increase the risk of TTA misjudgments by other road users.

Use of adaptive cruise control functions on motorways and urban roads: Changes over time in an on-road study.

The study aimed at investigating how drivers use Adaptive Cruise Control and its functions in distinct road environments and to verify if changes occur over time. Fifteen participants were invited to drive a vehicle equipped with a Stop & Go Adaptive Cruise Control system on nine occasions. The course remained the same for each test run and included roads on urban and motorway environments. Results showed significant effect of experience for ACC usage percentage, and selection of the shortest time headway value in the urban road environment. This indicates that getting to know a system is not a homogenous process, as mastering the use of all the system's functions can take differing lengths of time in distinct road environments. Results can be used not only for the development of the new generation of systems that integrate ACC functionalities but also for determining the length of training required to operate an ACC system.

The critical tracking task: a potentially useful method to assess driver distraction?

OBJECTIVE: We report on four experiments that investigated the critical tracking task's (CTT) potential as a tool to measure distraction. BACKGROUND: Assessment of the potential of new in-vehicle information systems to be distracting has become an important issue. An easy-to-use method, which might be a candidate to assess this distraction, is the CTT. The CTT requires an operator to stabilize a bar, which is displayed on a computer screen, such that it does not depart from a predefined target position. As the CTT reflects various basic aspects of the operational level of the driving task, we used it as a simple surrogate for driving to assess the CTT's capabilities. METHOD: We employed secondary tasks of varying demand, artificial tasks as well as tasks representative of secondary tasks while driving, and asked participants to perform them together with the CTT in parallel. CTT performance, secondary task performance, and subjective ratings of load were recorded and analyzed. RESULTS: Overall, the CTT was able to differentiate between different levels of demand elicited by the secondary tasks. The results obtained corresponded with our a priori assumptions about the respective secondary tasks' potential to distract. CONCLUSION: It appears that the CTT can be used to assess in-vehicle information systems with regard to their potential to distract drivers. Additional experiments are necessary to further clarify the relationship between driving and CTT performance. APPLICATION: The CTT can provide a cost-effective solution as part of a battery of tests for early testing of new in-vehicle devices.

On the relationship between pedestrian gap acceptance and time to arrival estimates.

The identification of safe gaps between passing cars when crossing a street is a task most of us accomplish successfully on a daily basis. Objectively, how safe a specific gap is, is mainly dependent on how long it would take the approaching vehicle to arrive (time to arrival; TTA). Common sense might suggest that TTA is the basis for pedestrians' gap selection. However, it has been shown repeatedly that vehicle approach speed has a substantial influence on the size of chosen gaps. At higher speeds, pedestrians tend to accept smaller time gaps, i.e. they initiate riskier crossings. Some researchers have gone so far as to suggest that pedestrians rely more on physical distance of a vehicle in their crossing decisions than TTA. Yet, at the same time, there is evidence that TTA estimates themselves are influenced by object approach speed. It is suspected that pedestrians are more apt to base their decisions on systematically distorted TTA estimates, rather than physical distance. The goal of the two experiments described in this article was to explore the relationship between gap acceptance and TTA estimation. Participants were presented with video clips of approaching vehicles, and were either required to indicate a crossing decision, or to estimate TTA. Results show the typical effects of speed (smaller gaps at higher speed, lower TTA estimate at lower speed) and age (larger gaps for older participants). However, when using subjective time gap size (the TTA estimate) instead of objective time gap size to predict gap acceptance, the effect of speed either disappeared (Experiment I) or decreased substantially (Experiment II). The results indicate that systematic differences in TTA estimates can be a reasonable explanation for the effect of speed on gap acceptance.

How does a lower predictability of lane changes affect performance in the Lane Change Task?

The Lane Change Task (LCT) is an established method to assess driver distraction caused by secondary tasks. In the LCT ISO standard, "course following and maneuvering" and "event detection" are mentioned as central task properties. Especially event detection seems to be a reasonable feature, as research suggests that distraction has profound effects on drivers' reactions to sudden, unexpected events. However, closer inspection of the LCT reveals that the events to be detected (lane change signs) and the required response are highly predictable. To investigate how the LCT's distraction assessment of secondary tasks might change if lane change events and responses were less predictable, we implemented three different versions of the LCT - an "original" one, a second one with lowered predictability of event position, and a third one with lowered predictability of event position and response. We tested each of these implementations with the same set of visual and cognitive secondary tasks of varying demand. The results showed that a decrease in predictability resulted in overall degraded performance in the LCT when using the basic lane change model for analysis. However, all secondary task conditions suffered equally. No differential effects were found. We conclude that although an ISO conforming implementation of the LCT might not be excessively valid regarding its depiction of safety relevant events, the results obtained are nevertheless comparable to what would be found in settings of higher validity.

Learning effects in the lane change task (LCT)--realistic secondary tasks and transfer of learning.

Driver distraction is a factor that is heavily involved in traffic crashes. With in-vehicle devices like navigation systems or mobile phones on the rise, the assessment of their potential to distract the driver has become a pressing issue. Several easy-to-use methods have been developed in recent years to allow for such an assessment in the early stages of product development. One of these methods is the lane change task (LCT), a simple driving simulation in which the driver has to change lanes as indicated by different signs along the road. Although the LCT is an ISO sanctioned procedure, there are still open questions. One issue are learning effects which have been found in previous studies and which have the potential to compromise the comparability of test results. In this paper, we present results on two experiments that further explored the effect of previous experience on LCT and secondary task performance. The results confirm that learning effects occur when combining the LCT with a realistic secondary task. Also, we found evidence for the transfer of learning from one secondary task to another to some degree, provided that the two tasks are sufficiently similar.

Can driver education be improved by computer based training of cognitive skills?

Deficits in cognitive skills, such as hazard perception, appear to have a tremendous influence on accident involvement of younger drivers. However, conventional forms of driver training have largely failed to build skills that extend beyond the provision of a descriptive knowledge of how to drive. Computer based training (CBT) has the potential to provide new ways to deal with this problem. In this study, a CBT module was developed to complement existing driver training programs by addressing critical cognitive skills. The CBT made use of video sequences of potentially hazardous driving situations, including multiple-choice questions with adaptive feedback, to increase levels of elaboration and understanding. To test effects, a sample of learner drivers completed either CBT, paper based training with similar content, or no training at all. A simulator experiment confirmed that CBT participants exhibited earlier glances towards critical cues and relevant areas in the visual field than participants of the other two groups. It is concluded that CBT can potentially assist instruction of cognitive skills necessary for save driving.