What turns a bicycle street into a street for cyclists? A multimodal study on subjective safety of infrastructure measures on bicycle streets using an approach in virtual reality

Introduction

In a world increasingly shaped by urban mobility and sustainable transportation, cycling is gaining more and more importance as an environmentally friendly and health-promoting means of travel. According to the results of the Bicycle Monitor (Sinus Institut, 2021), cycling has the highest growth potential compared to other modes of transportation in Germany. In addition to the steadily increasing number of cyclists, the sense of safety among them while using bicycles increased by the year 2021 compared to the reports of 2017 and 2019 (Sinus Institut, 2021). Subjective safety is an important concept as it influences the decision to choose cycling as a mode of transportation (Sanders, 2013). In general, 63% of cyclists in Germany reported feeling very or somewhat safe (2021: 63%, 2019: 56%, 2017: 53%), still, around a third of cyclist in Germany fells unsafe or very unsafe (Sinus Institut, 2021). Speed limits for motorized vehicles and their presence on or directly next to bicycle infrastructure have the greatest influence on this perception of safety (Sørensen & Mosslemi, 2009; Vandebona & Kiyota, 2001). So, a lack of dedicated bicycle lanes might contribute to this feeling as well since they can make biking more comfortable, quicker, and safer (Monsere et al., 2014), but their effectiveness largely depends on their design quality and road environment, which is constrained by regulatory uniformity.

Analyzing the road environment and infrastructure characteristics is crucial for both the objective and subjective safety of cyclists. The ‘road environment’ encompasses all physical and traffic-related features, including traffic volume and composition, infrastructure characteristics and environmental conditions. Together, these elements determine how safe and comfortable cyclists and other road users can navigate the road (Meuleners et al., 2019). Infrastructure elements significantly impact perceived safety. Cyclists feel safer on routes with dedicated bike lanes, traffic calming measures, and residential streets, while high-risk perceptions are linked to major streets without bike infrastructure and areas with high vehicle interaction (Manton et al., 2016). Cyclist safety perceptions are influenced by road type, presence of bike lanes, and vehicle interactions. Cyclists report feeling unsafe on roads without bike lanes and near large vehicles, underscoring the need for safer road designs (Beck et al., 2021). In Germany this is mirrored by cyclists who wish more bikes lanes (56%), better separation from car traffic (50%), and more cyclist-friendly streets (39%) (Sinus Institut, 2021).

Bicycle lanes are also an opportunity to increase safety, as they are used exclusively or mainly by cyclists and offer significant safety benefits and are strongly preferred by users (Lusk et al., 2011). They promote the use of bicycles as a sustainable mode of transport and help to reduce traffic risks (Blitz et al., 2020; Tsubohara, 2015). However, conflicts with motorised traffic and inadequate design can reduce the acceptance and effectiveness of such measures (Blitz et al., 2020). For example, a study in the Rhine-Main region, Germany, found that the introduction of cycle lanes had a positive impact on the use of bicycles, while the reduction in car use was limited.

In Germany, a bicycle street is a designated road for bicycle and E-Scooter traffic. According to a survey by the German Federal Ministry of Transport and Digital Infrastructure (BMDV, 2021), 39% of respondents listed ‘more bicycle routes‘ as the most urgent demand for the government. Hence, bicycle streets need to become an important element in bicycle traffic planning to increase the safety and attractiveness of cycling in Germany (BMDV, 2021). To accomplish this, cars and other vehicles are allowed but with speed restrictions and special rules to increase the attractiveness and safety of cycling and create advantages over motor vehicle traffic. Other European countries have similar infrastructure measures, e.g. in the Netherlands, so-called ‘fietsstraaten’ (bicycle streets) are established, which are shared by bicycles and cars but give priority to bicycles, are a common practice. These roads have proven to be effective in reducing the speed of cars and increasing the safety of cyclists (Tsubohara, 2015).

Currently, about 96% of German bicycle streets allow motorized traffic (ADFC, 2022). Moreover, the establishment of bicycle streets is governed by technical regulations based on objective safety analysis, often overlooking cyclists' subjective safety (Gössling & McRae, 2022; Schwedes et al., 2021). Bicycle streets in Germany are built based on recommendations (Alrutz, 2011), some of which are viewed as outdated, and the experiential knowledge of planners, who typically do not evaluate completed infrastructure projects for their perceived safety (Schwarzkopf et al., 2023). This can result in infrastructures that are safe in objective terms but do not feel safe to cyclists, deterring bike usage (Hull & O'holleran, 2014; Wang et al., 2014). One way to make infrastructure more user-centred is to introduce a method that enables urban planners to involve the public more, taking into account the (road) environment and make planning more transparent (Wolf et al., 2020).

In the last decade, an increase in the use of Virtual Reality (VR) in infrastructure design and urban planning has been observed. VR has been successfully used in the area of citizen participation for construction projects, also infrastructure, buildings and green spaces have been planned with the help of VR (Mahbubur & Kitson, 2020; Wolf et al., 2020; Zhang et al., 2020). VR can also be used early in the planning stage to evaluate the impact of infrastructure on safety, comfort, and (dis-)satisfaction (Schwarzkopf et al., 2023; Wolf et al., 2020). VR offers several advantages compared to previously used evaluation methods in urban planning (e.g. inspections, on-site surveys): Besides providing a more motivating and realistic evaluation of infrastructure, it can be employed anywhere due to its integrated computing units (Wolf et al., 2020). Furthermore, an evaluation of the perceived safety of cycling infrastructure is generally not carried out by planners, as on-site inspections and surveys are considered too time-consuming and therefore also resource-intensive (Schwarzkopf et al., 2023). Accordingly, the user's perspective is generally not taken into account when planning and evaluating infrastructure projects in Germany. The capabilities of VR technology could be used, among other things, to evaluate the perceived safety of (cycling) infrastructure in a resource and time-efficient manner, as it is suitable for replacing or supplementon-site surveys. Furthermore, multiple (cycling) infrastructure measures could be sequentially evaluated in a short time under controlled conditions, even if the person who evaluates the infrastructure does not know it in real life (Schwarzkopf et al., 2022; Wolf et al., 2020). This also opens up the possibility of simulating traffic situations that demonstrate safety-critical behavior and making them accessible to the test subjects without endangering them (e.g. in the case of critical overtaking maneuvers). In summary, VR can help to bring the user's perspective into infrastructure planning and give subjective factors (e.g. subjective safety) more influence in planning. In addition, critical traffic situations can be easily simulated without putting the test subjects in danger, which also opens up the possibility of interviewing the group of people who refrain from cycling due to safety concerns.

VR can also contribute to closing some research gaps in the field of bicycle streets: There is little research on how specific environmental factors of bicycle streets affect subjective safety, there is a lack of data on whether bicycle streets shared with cars under certain conditions improve perceptions of safety, and the focus is usually on active cyclists, ignoring those who do not cycle frequently for safety reasons. But, above all, the evaluation of subjective safety is rarely or not at all of interest to those who bear a significant portion of the responsibility for infrastructure design: the planners and authorities. In this paper, we have two goals: First, we aim to address these issues by adopting a multimodal approach. The second goal is to test a prototype for evaluating the perceived safety of cycling infrastructure with static, stereoscopic images in VR.

In study I, we employ an online questionnaire to assess the perceived safety of a selected bicycle street in Chemnitz (Reichenhainer Straße) that has through traffic. This study additionally investigates how the street environment, in this case: infrastructure elements, street condition and associated rules, are evaluated in terms of perceived safety. In study II, we use a VR approach to evaluate perceived safety on the same bicycle street and the rating of the subjective safety of its street environment as well. For this purpose, static, stereoscopic images are presented in a self-developed VR application.

Method

For both studies, it was initially essential to identify infrastructure elements and safety habits relevant to cyclists. To this end, relevant elements were first extracted from the literature (Lawson et al., 2013; Manton et al., 2016) and the ‘Empfehlungen für Radverkehrsanlagen’ (Recommendations for Cycling Infrastructure) ERA 10 (Alrutz, 2011) which is a widely used basis for bicycle infrastructure planning in Germany. The infrastructure elements listed in the ERA-10 are intended to contribute to the (objective) safety of cycling infrastructure. Subsequently, an onside walkthrough on the bicycle street was conducted with a group of N = 7 cyclists and N = 2 urban planning experts to identify important infrastructure characteristics for the perceived safety of cyclists. The total set of identified infrastructure characteristics was then discussed and categorized in a focus group of N = 4 experts from the fields of cycling safety and urban planning. Additionally, possible implementations of further elements (e. g. bumpers; if urban planning permits), rule-based (e.g. allowed maximum speed) and condition-based (e.g. traffic volume) influences were also considered. This set of street environment characteristics was either fully (study I) or partially (Study II) utilized in both studies (refer to Table 1). Content that could not be equally represented in both the online questionnaire and the VR study was not considered (e.g. the priority of cyclists over other road users).

The ADFC (2022) Bicycle Climate Test for Chemnitz was used as the measurement tool for general subjective safety while cycling in both studies. It originally includes seven items related to subjective safety. For the survey on the bicycle street, these items were slightly modified to maintain comparability with the 2020 Bicycle Climate Test for the Chemnitz area (ADFC, 2022), we ignored an item that revolves around the topic of bicycle theft. The answers were assessed using a six-point Likert-scale (1—totally agree, 6—totally disagree). The items used are displayed in Table 2.

We opted for two studies to achieve both a general overview with as many participants as possible (Study I) and to take a more detailed look at individual infrastructure characteristics, as well as to test a new method in VR (Study II).

Analysis

In this section, we briefly summarize the analyses we are planning for the collected data. For both studies (online questionnaire and virtual reality), the first step is to determine how the respondents rate the perceived safety of the cycle lane and compare the result with the general assessment of cycling infrastructure in Chemnitz by the Allgemeiner Deutscher Fahrradclub (ADFC, 2022). Individual infrastructure elements and rules are then assessed in terms of their perceived safety, as we want to identify which elements or rules increase the perceived safety of the infrastructure. In addition, it is also analyzed how well respondents rate the form of presentation of street environment elements as pictures in an online questionnaire and in VR. Hierarchical regression is then used for Study I to analyze whether the evaluated street environment elements have a significant influence on the perceived safety of the bicycle street, e.g. the entire infrastructure. In Study II, we also want to get a first impression of how VR technology can be used as an assessment tool for the perceived safety of cyclists.

Study I—online questionnaire

Study I consisted of a questionnaire which was provided in an online format and comprised items on the following topics: awareness of the bicycle street, mobility behavior, cycling experience, knowledge of traffic rules on bicycle streets, usefulness, acceptance, subjective safety, and sociodemographic data. The infrastructure characteristics (see Table 1) were presented with two pictures from the bicycle street Reichenhainer Straße depicting typical scenarios (if possible). Examples are displayed in Figure 1.

For each element of road environment characteristics, participants evaluated whether the respective element increased or decreased perceived safety using a 6-point Likert scale (1—very unsafe; 6—very safe). This data was also used to calculate a hierarchical regression. For this purpose, a G-Power analysis was conducted, selecting a significance level of α = .05 with a test power (1-β) of .80 and a small effect size (f² = .02) for the expected influence of infrastructure characteristics on the subjective safety. This resulted in a minimum sample size requirement of 175 participants. The questionnaire was completed by N = 182 individuals (n = 68 females, M_age = 30.33, SD_age = 9.34). Only individuals who cycled on the bicycle street at least once a month were eligible to participate. A total of n = 114 participants defined themselves as experienced cyclists. The questionnaire was available online during the month of June 2022. Participants did not receive any compensation for their participation. Test subjects were recruited via a flyer with a QR code and mailing lists.

Study II—VR experiment

The Study II, a VR study, is based on our previous work (Schwarzkopf et al., 2022). From the infrastructure characteristics listed in Table 1, those from the ‘Existing Elements’ and ‘Condition’ categories were implemented in VR. Stereoscopic 360° images of these elements were captured, showing the same elements and comparable contexts as the pictures presented in the questionnaire. These images were then incorporated into a self-developed app and presented randomly. Some of the photos showing vehicles on the road reveal irregular behavior by motorized traffic. Among other things, the pictures show an overtaking situation, which is illegal due to the total width of the road (minimum distance of 1.5 meters from cyclists cannot be maintained during an overtaking maneuver). However, this behavior is very common on the bicycle road and is therefore part of our observations. During the use of the VR headsets, participants were asked to imagine they were cycling. For each of the twelve images, participants evaluated the subjective safety of each of the visible infrastructure characteristics on a 6-point Likert scale (1—very unsafe; 6—very safe) also using thinking aloud. Additionally, participants were asked to explain the reasoning behind their assessments. After viewing all the images, participants rated the subjective safety for the entire bicycle street on a 6-point Likert scale (1—very unsafe; 6—very safe). In contrast to Study I, participants evaluated the current state and assessed how safe they perceived it to be. The procedure is presented in Figure 2. After completing the VR part, a questionnaire was administered to collect demographic data as well as information on bicycle use and experience (identical questions as in Study I) as well as an inventory for affinity for technology (Franke et al., 2018) and the intuitiveness of technology use. In this study, N = 32 (n = 13 females, M_age = 26.22, SD_age = 3.04) cyclists participated, of whom n = 18 identified themselves as experienced cyclists, n = 11 participants do not use the bicycle street and n = 23 had no prior experience with VR headsets. The experiment took 15–20 minutes to complete. The study took place directly in front of or in the student canteen of Chemnitz university of technology. The test subjects were selected at random and approached before or during lunch. Participation was not remunerated.

Results

Discussion

Conclusions

The presented studies demonstrate consistent evaluation results regarding the subjective safety assessment of the showcased bicycle street. The evaluated bicycle street is perceived to be safer than the average bike-accessible infrastructure in the city area of Chemnitz. Key factors influencing this assessment were identified as the lane width and the restrictions on motorized traffic. These results are not surprising and align with the findings of existing literature. However, to the authors' knowledge, Study II is the first to evaluate the perceived safety of cycling infrastructure using VR and stereoscopic 360° images. The results regarding the general subjective safety obtained were not only consistent with those from Study I, a traditional online questionnaire study, but also provided a focused view of road environment characteristics and their perceived safety. It was shown that depending on the presentation (cars in the immediate vicinity) and cycling experience, there are significant differences in the evaluation of individual infrastructure characteristics in VR. These results are interesting because even during a site visit or on-site survey, it would be difficult (detailed evaluation of a single road environment element) or too risky (close but everyday overtaking maneuvers of motorized traffic) to precisely examine different contexts or environmental influences. Furthermore, the presented method offers a structured and controlled way to evaluate cycling infrastructure realistically under laboratory conditions in a resource-efficient manner. Within 15 minutes, participants could evaluate twelve images or scenarios and complete a follow-up questionnaire. Preparing the study was time-efficient due to a self-programmed app; only the scenarios needed to be identified and the images taken. Moreover, the method is location-independent, easily transportable, and can be intuitively used without technical knowledge thanks to VR headsets with integrated computing units. Assuming further evaluation of the method and its potential, it could be used for citizen participation measures or future infrastructure planning projects to evaluate their subjective safety in a resource-efficient manner after completion and to use the findings for future planning projects. This is a process that is rarely, if ever, carried out in planning offices in Germany and can contribute to increased safety of cycling infrastructure in general.

Limitations

Both studies were based on a list of infrastructure characteristics that may not have captured all relevant elements, meaning there could be relevant infrastructure characteristics that the studies did not include. Moreover, the images in the questionnaire and especially in VR may have shown influences (e.g. weather) that were not quantified or collected but could affect the evaluation of perceived safety, even though efforts were made to keep the images comparable across both studies and to exclude disturbing influences. Additionally, the rating scheme in both studies was different, which makes the results only partially comparable. Furthermore, the sample compositions in both studies were different: Study I involved only users of the bicycle street, while Study II also included non-users. Another point is the small sample size in Study II, which could impact the validity and reliability of the data. Moreover, the VR method applied in Study II is not yet fully mature, which could affect the participants' evaluations.

Funding

This research work was supported by the German Federal Ministry of Education and Research (BMBF) (project ‘NUMIC 2.0—Neues urbanes Mobilitätsbewusstsein in Chemnitz 2.0’, FKZ 01UR2204B) as part of the funding program ‘Teilprojekt B: Erforschung und Implementierung partizipativer Umsetzungsstrategien und Unterstützungs-formate für Bürgerbeteiligungsprozesse in der Stadt- und Verkehrsplanung’. The funder had no influence on the study design, the collection, analysis and interpretation of the data, the writing of the report or the submission of the article.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Ethics statement

This study was approved by the ethics committee of Chemnitz University of Technology (reference number 101579653).

CRediT contribution statement