Conference PaperPDF Available

Empirical Survey on Bicycle Accidents to estimate the Potential Benefits of Braking Dynamics Assistance Systems

Authors:
  • Robert Bosch GmbH, Stuttgart, Germany
  • University of Applied Sciences Pforzheim

Abstract and Figures

The ease of movement while driving an Electric Bicycle (EB) inspires many. The total number of EBs in Germany has now grown to 2.1 million. In the medium term EBs are estimated to have a market share of 15 % in the overall bicycle market. In addition to assist the driver while accelerating the bicycle, the available electrical energy on EBs also offers the possibility to support the driver during deceleration. This is the initial point for the interdisciplinary research project BikeSafe aiming to develop an innovative Braking Dynamics Assistance system (BDA) for bicycles with hydraulic brakes. The essential function of the BDA is to prevent both major critical braking situations for single-track vehicles: front wheel lockup and nose-over (falling over the handlebars). In order to ensure a target group oriented and accident specific development of the BDA, a better understanding of EBs' accident situations being influenceable by active safety systems is necessary. A substantial problem is the well-known underreporting of bicycle accidents in official statistics based on police reports. Especially when investigating single-vehicle accidents in which both critical braking situations mainly occur this phenomenon needs to be addressed by an adequate approach. For this reason, the present paper evaluates a literature review of studies on underreporting, single-vehicle accidents, cyclist behaviour, and changed mobility patterns due to EBs. Based on these findings several research questions concerning the development of the BDA, which were additionally considered within an empirical survey (1,040 participants, representative for German adults above 18), are described and discussed. For example, a proportion of single-vehicle accidents on all accidents of 65 % was determined. In conclusion, the results of this paper provide both quantitative and qualitative information to estimate the potentials of and to define future requirements for the BDA.
Content may be subject to copyright.
International Cycling Safety Conference 2015
15-16 September 2015, Hannover, Germany
ABSTRACT
The ease of movement while driving an Electric Bicycle (EB) inspires many. The total number of
EBs in Germany has now grown to 2.1 million. In the medium term EBs are estimated to have a
market share of 15 % in the overall bicycle market. In addition to assist the driver while
accelerating the bicycle, the available electrical energy on EBs also offers the possibility to
support the driver during deceleration. This is the initial point for the interdisciplinary research
project BikeSafe aiming to develop an innovative Braking Dynamics Assistance system (BDA)
for bicycles with hydraulic brakes. The essential function of the BDA is to prevent both major
critical braking situations for single-track vehicles: front wheel lockup and nose-over (falling
over the handlebars).
In order to ensure a target group oriented and accident specific development of the BDA, a
better understanding of EBs’ accident situations being influenceable by active safety systems is
necessary. A substantial problem is the well-known underreporting of bicycle accidents in
official statistics based on police reports. Especially when investigating single-vehicle accidents
in which both critical braking situations mainly occur this phenomenon needs to be addressed
by an adequate approach.
For this reason, the present paper evaluates a literature review of studies on underreporting,
single-vehicle accidents, cyclist behaviour, and changed mobility patterns due to EBs. Based on
these findings several research questions concerning the development of the BDA, which were
additionally considered within an empirical survey (1,040 participants, representative for
German adults above 18), are described and discussed. For example, a proportion of single-
vehicle accidents on all accidents of 65 % was determined. In conclusion, the results of this
paper provide both quantitative and qualitative information to estimate the potentials of and
to define future requirements for the BDA.
Keywords: sustainable mobility, electric bicycles, accident research, single-vehicle crashes.
Empirical Survey on Bicycle Accidents to estimate
the Potential Benefits of Braking Dynamics Assistance Systems
Oliver Maier1, Martin Pfeiffer2, Christa Wehner3, Jürgen Wrede4
1
Institute of Applied Research
Pforzheim University
Tiefenbronner Str. 65, 75175 Pforzheim, Germany
e-mail: oliver.maier@hs-pforzheim.de
2
Department of Information Technology
Pforzheim University
Tiefenbronner Str. 66, 75175 Pforzheim, Germany
e-mail: martin.pfeiffer@hs-pforzheim.de
3 Department of Market Research
Pforzheim University
Tiefenbronner Str. 65, 75175 Pforzheim, Germany
e-mail: christa.wehner@hs-pforzheim.de
4 Department of Mechanical Engineering
Pforzheim University
Tiefenbronner Str. 66, 75175 Pforzheim, Germany
e-mail: juergen.wrede@hs-pforzheim.de
2
1 INTRODUCTION
1.1 Vulnerable Road Users
In its policy orientations on road safety, the European Commission is putting particular
emphasis on the need to improve the safety of Vulnerable Road Users (VRUs) [1]. VRUs are
defined as those participants in traffic that are not protected by any mechanical system:
pedestrians, bicyclists and motorcyclists [2].
Actually, VRUs account for about 45 % of European Union (EU) road fatalities [1]. Although the
total number of fatalities and severe injuries due to traffic accidents is decreasing, the number
of VRUs that are killed and wounded in traffic tends to decrease at a much slower pace [1-2,
4]. While pedestrians form the largest sub-group of all fatalities, Powered Two-Wheelers
(PTWs) are involved in a disproportionately high percentage of serious and fatal accidents [1].
Due to the high speed reachable and the weight of the vehicle, single-vehicle accidents
involving PTWs can also be fatal or lead to very severe injuries [2].
Even today, cyclists are the passenger group that comprises those individuals who are the
most seriously injured, nearly 2,000 of approximate 4,500 in Sweden in 2012 [3]. In the whole
EU cyclists account in official accident statistics for ca. 7 % of all road fatalities [5], and there
are concerns that the growing number of Electric Bicycles (EBs) as a new subgroup of PTWs
might cause additional safety risks [1].
Especially the ease of movement while driving an EB inspires many. The total number of EBs in
Germany has grown to 2.1 million [6]. In the medium term EBs are estimated to have a market
share of 15 % in the overall bicycle market [7]. Summarizing, the provision of power assistance
to the rider potentially will extend the role of bicycles in urban transport [8]. From a societal
point of view this modal shift from mainly cars to bicycles is a very desirable development. A
study in the Netherlands proves that, on average, for individuals shifting their mode of
transport the estimated health benefits of cycling are substantially larger than the risks
compared to car driving [9].
Considering the specific safety risks of all VRUs, there is an important need for the
development and introduction of advanced technologies and systems for VRU protection. First
priority should be on collision avoidance or mitigation [1].
1.2 Braking Dynamics Assistance System
In-vehicle systems (active safety systems), roadside systems (e.g. camera’s overlooking a
crossing and identifying road users with their future paths), car-to-infrastructure and
infrastructure-to-car communication, and an intelligent traffic management system are the
most promising technical ways for improving the safety of VRUs [2]. Regarding active safety
systems, the available electrical energy on EBs offers the possibility to support the cyclist
during deceleration. For this reason, Grimaldi considers an Anti-lock Braking System (ABS) on
EBs as a realistic and effective measure for improved safety [3].
The interdisciplinary research project BikeSafe links precisely to this idea. It aims at the
development of an active Braking Dynamics Assistance system (BDA) for bicycles equipped
with hydraulic brakes. The BDA to be designed will target to prevent both major critical braking
situations for single-track vehicles: front wheel lockup and nose-over (falling over the
handlebars). In order to ensure a target group oriented and accident specific development of
3
the BDA, a better understanding of EBs’ accident situations - being influenceable by active
safety systems - is necessary. A substantial problem is the well-known underreporting of
bicycle accidents in official statistics based on police reports.
1.3 Research Questions
The aims of this paper are to provide both quantitative and qualitative information for
estimating the potential benefit of such a BDA and to define future requirements for it. In
detail, the following issues are examined:
Validity and limitations of official accident statistics based on police, hospital and
expert (e.g. German In-Depth Accident Study, GIDAS) reports
Underreporting of bicycle accidents and resulting impacts on the proportion of single-
vehicle accidents
Effects on the current accident situation caused by changes in mobility patterns due to
electrical bicycles
Cyclist behaviour and accident cause factors with special focus on the critical braking
situations to be prevented by the BDA
The paper begins with a literature review of studies on underreporting of bicycle accidents in
official statistics. Next studies on single-vehicle accidents and cyclist behaviour investigated by
both health care data analysis with additional surveys and Naturalistic Cycling Studies (NCS)
are analysed. Subsequent sections examine changes in mobility patterns due to EBs and
explains the methodology of the conducted survey representative for German population
above 18. We then present the results of the study to provide both quantitative and
qualitative information to estimate the potentials of the BDA. Based on these findings several
research questions concerning the development of the BDA are described and discussed. The
final section concludes with the most important findings and their implications for future
works.
2 MATERIAL AND METHODS
2.1 Underreporting of Bicycle Accidents in Official Statistics
Comparing bicycle use among different countries reveals enormous differences from near
absence to widespread use. While there are countries where cycling is mostly practiced for
recreation, cycling is a substantial part of everyday life in other countries such as the
Netherlands. Although cycling activities also take place in rural areas, the majority of the
bicycle kilometres in the latter countries are travelled in towns and cities. In conclusion,
country-specific differences in bicycle use lead to significant differences in bicycle culture,
cyclists’ role in traffic, and technical and regulatory measures employed for safer cycling [4].
Despite the difficult comparability of individual countries, it can be noted that all worldwide
accidents statistics indicate that cyclists have a relatively high crash rate compared to other
road users [1-2]. When drawing conclusions from official statistics, three substantial issues are
challenging: underreporting of bicycle accidents, age distribution of injured cyclists, and basic
problems cause by bicycle design [4].
4
Especially when non-fatal bicycle injuries are concerned, statistics suffer from underreporting
[10]. Some countries have made initial attempts to link data sources, such as police and
hospital data, in order to overcome at least part of this problem [4]. A study in the Netherlands
[11] compared data for serious traffic injuries stemming from hospital data bases with police
data. It was concluded that the police records contained only 59 % of the seriously injured
cyclists that crashed with motor vehicles and only about 4 % of those being injured with no
motor vehicle involved. Similar findings were obtained in studies carried out in the German
bicycle-capital Muenster [12-13] and in the French County Rhône-Alpes [14].
In summary, confirmed by the results of a meta-analysis of studies made in 13 countries [15],
reporting rate tends to be highest for collisions between cyclists and cars and lowest for
cyclists on single-vehicle accidents (the latter less than 10 % in all countries studied).
2.2 Single-Vehicle Accidents and Cyclist Behaviour
Both critical braking situations (front wheel lockup and nose-over) mainly happen in single-
vehicle accidents. For this reason, as the studies investigating underreporting show, an
accident-specific development of the BDA based on official statistics from police reports is not
convincing. Although studies based on health care services data found that single-vehicle
accidents make up more than 60 % of injurious cyclist accidents, they lack in-depth
information on the circumstances of these accidents [16]. In this context, the integration of
additional surveys with crashed cyclists increases explanatory power in a very effective manner
[17]. Naturalistic Cycling Studies (NCSs) are another convincing method that was derived from
motor vehicle accident research and recently has been adopted for bicycles [18].
The basic idea of NCSs is to analyse road user behaviour and traffic safety by cameras and
sensors on instrumented vehicles. Main advantages are the availability of records for all events
leading to an accident and the quality of that data as there is no influence due to an artificial
experimental setup [19-20]. One of the first known NCSs with Conventional Bicycles (CBs) was
carried out by Chalmers University of Technology and took place in Gothenburg [19]. The study
found that cycling on poor road surfaces increased the likelihood of experiencing a critical
event defined as anything that made the cyclist uncomfortable about safety by the factor
of ten.
This result is supported by another study from Sweden [21] that paired hospital accident data
with specific interviews and found 70% of all severe bicycle injuries being due to single-vehicle
accidents. As a further result of this study, 27 % of those single-vehicle accidents were related
to maintenance and slippery surfaces. Although cycling is partly determined by a country’s
geography and its climate [4], a quarter of cyclists being involved in on-road single-vehicle
accidents in Australian Capital Territory also reported surface characteristics as a contributing
factor to the accident [16]. This includes irregular surface (e.g. cracked surface, potholes, etc.),
wet or slippery surface, loose gravel, sand and debris. In addition to other contributing factors
(carrying extra weight effecting balance, speeding, distraction, and bike failure multiple
cause could be given) almost three quarters reported losing control of the bicycle as a trigger
for their single-vehicle accident.
A study carried out by German Insurers Accident Research based on hospital accident data
paired with specific interviews confirms these findings [17]. According to this study, there are
two scenario types leading to single-vehicle accidents. The first one representing around
50 % of all single-vehicle accidents is due to poor road conditions (mainly at high speeds)
and/or the loss of rider’s balance (mainly at low speeds) and results in a lateral fall of the
5
bicycle. The second scenario is a critical braking situation leading to nose-over (ca. 14 %). Main
reasons are either a defective reaction of the cyclist resulting in an over-braked front wheel or
a front wheel lockup caused by an object such as a sprig. Furthermore, the study states that
the nose-over motion typically occurs in longitudinal direction with less lateral deviation and
mainly at high speed.
2.3 Changes in Mobility Patterns by Electric Bicycles
EBs represent a new dimension for cycling and a rapidly growing market. In the near future,
EBs will cover almost all bicycle segments [3]. When examining EBs’ accident situation, it is
important to begin by distinguishing between e-bikes and pedelecs [20]. On e-bikes, the
auxiliary electric motor operates with a switch or throttle and provides power assistance
without any pedal action. In contrast, pedelecs only give support when the rider is pedalling
[8]. For the latter category of vehicles the Comité Européen de Normalisation (CEN)
implemented the EPAC (Electrically Power Assisted Cycles) standard [22].
One of the largest EB studies to date was undertaken in 2000 by the Centre for Electric Vehicle
Experimentation in Quebec (CEVEQ) [23]. The results suggest that participants perceived the
benefits provided by EBs for utilitarian travel, particularly commuting [8]. Both Parker [24] and
Cherry & Cervero [25] come to similar conclusions. In addition, EBs present a mobility option
for the elderly and those with a physical impairment being unable or finding it uncomfortable
to ride a CB [26-27]. Summarizing, EBs appeal to individuals who otherwise would not ride and
increase frequency of usage and/or range of riding for those currently using a CB [8].
Despite increasing acceptance and usage of EPACs, very little research has been conducted to
assess road safety implications for this emerging vehicle type with possible higher travel speed.
The significant health burden among cyclists due to road accidents, as shown in the
introduction of this paper, raises the question of how safe EBs are compared to CBs. [28-29]
In Germany, EPACs only recently (in 2014) have been included as a separate vehicle category in
the official accident statistics. Reliable accident statistics will therefore not be available before
2017 [30]. For bridging this period, a study of Robert Bosch Company [31-32] selected an
obvious approach. Assuming general behaviour of EPAC users is similar to CB usersjustified
with the fact that a normal cyclist is able to reach nearly the same driving performance like a
user of an EPAC if some more muscle power is considered an extrapolation of the traffic
accident situation involving CBs with comprehensible assumptions respect to EPACs is possible.
The study was based on bicycle accidents in 2009 available in the GIDAS data base. Results
show that in 50 % of the bicycle accidents registered in GIDAS, travel speed will not increase
with EPACs. Main causes were the driving situation (e.g. travel speed higher than supported by
electric auxiliary drive) and the fact that a higher speed with respect to the situation (e.g.
turning) is not realistic. Taking also into account the accident situations in which the bicycle
was hit laterally by a motor vehicle - travel speed was assumed to have no influence the
study concludes that in most real situations (68 %) an electrical support of the bicycle has no
influence on the accident.
In Switzerland, EBs are explicitly included in official accident statistics since 2011. Initial results
[33] show that cyclists on EBs involved in accidents were mostly 45 - 65 years old. The
evaluation of the injury severity of cyclists on EBs compared to those on CBs led to diverging
results and no clear statement could be drawn.
6
Both studies presented [31-33] are based on official accident statistics and suffer from
underreporting. There are several NCSs investigating vehicle usage, speed and road safety of
EB users. A large-scale NCS carried out by German Insurers Accident Research [30, 34] and
Chemnitz University of Technology [20, 35] gave no evidence that EPAC users are not less safe
or associated with other risks than CB users. Furthermore, results of the study show that riders
on EPACs use the electric auxiliary motor for convenience primarily, which validates the
assumptions in [31-32]. Just like the findings in [33], the study concludes that EPACs are
currently mainly used by elder people, but there are no differences in the number of critical
incidents between EPACs and CBs. Notably, the study points out that EPAC users vary cycling
speed more often than CB users and thus more often find themselves in braking situations.
In addition to this enhanced transient driving behaviour, Chalmers University of Technology
found in their NCS [36, 37] that riders of EBs may become more prevalent in critical areas (e.g.
small hills) and critical periods (e.g. while raining) with limited visibility. Possibly as a
consequence of EBs’ higher travel speed distraction of the cyclists may be particularly
dangerous.
2.4 Representative Survey
Official accident statistics are clearly marked by the disadvantage of an underreporting of
bicycle accidents of unknown magnitude and, apparently, a lack of recent studies that could fill
this gap, which is why we have initiated our own survey.
Thanks to the generous cooperation with the market research and consulting institute YouGov
Germany we have been able to conduct an OMNIBUS survey. The results qualify to be
representative for the German resident population (18+ years old). Founded in London in
2000, YouGov is considered a pioneer in online market research and according to the American
Marketing Association ranks among the top 25 market research institutes in the world.
During summer of 2014 (field work August, 8-11) a representative sample of 1,040 persons had
to answer one or two questions. The first question was:
If you cycle at least once in a while: Do you remember any crash or bicycle accident within the
last 10 years?”
Following answers were possible - multiple choices were explicitly welcome:
1. Yes, I had a crash/an accident with a normal bicycle that had no electric motor.
2. Yes, I had a crash/an accident with an e-bike.
3. The crash/accident has been reported to the police.
4. No, I cycle but I did not have a bike accident.
5. No, I practically did not cycle within the last 10 years.
The first question was followed by an open question that addresses only those persons, which
remembered a crash/accident with a CB and/or EB (choices 1 or 2 being checked). Accident
cyclists were asked to describe the crash/accident as detailed as possible with no word
restriction. The exact wording of the question reads as follows:
7
Please describe as detailed as possible, how the accident happened. We are interested in
every detail, specifically the kind of crash/accident (e.g. nose-over, falling over the handlebars,
front wheel skid, etc.), road surface and condition, speed and everything that possibly effected
the braking manoeuvre. Is your bicycle equipped with hydraulic brakes? Did you need to
initiate an emergency braking?”
3 RESULTS
3.1 Representative Outcomes for Germany
The representativeness of the conducted survey allows a discussion of bicycle culture in
Germany. Figure 1 shows 78 % of Germans have driven their bicycle in the last decade. Cycling
was done without an accident by half of the people (49 %). An amount of 93% of those being
involved in an accident (27% of the overall amount) were using a CB and 7 % an EB. Based on
the bicycle stock in Germany (an estimated amount of 72 m), it could be concluded that riders
of EBs have an accident risk higher than riders of CBs (7 % accident risk at 2.9 % EB share in the
total bicycle stock). However, due to the small amount of EB riders involved (N=20), these
findings at best signify a slight tendency, but not a statistically significant finding. It also can be
assumed that EBs are used more often than CB as their number also includes very old bicycles.
Figure 1: Bicycle use and general accident situation
Those respondents with a bicycle accident were also asked if they have reported the accident
to the police. Only 28 of total 304 involved in an accident stated that they did a report. This
results in an underreporting rate of 91 %, which means only one in ten bicycle accidents was
reported by the police (cf. figure 2).
8
In addition to the police reports, all crashed cyclists were asked the second question. 270 of
the 304 answered this question in sufficient detail for further in-depth accident analysis. As a
first result of the content analysis, it was found that 65 % of respondents had a single-vehicle
accident (cf. figure 3).
Figure 2: Underreporting rate of bicycle accidents
Figure 3: Proportion of single-vehicle accidents contributing to all bicycle accidents
9
3.2 In-Depth Accident Analysis
The primary objective of the following in-depth accident analysis is to estimate the possible
benefits of the BDA by quantitative and qualitative statements. For this reason, types, triggers,
and fall events of bicycle accidents are evaluated.
The following figure 4 shows the distribution of accident types for CBs and EBs based on the
exploitable answers of 270 respondents to the second question (equates to 100 %). The type
of accident refers to the transport process which led to the accident. Collision type and
accident cause are not decisive for the determination of the accident type. In accident
research a total of seven groups exists to distinguish accident types [38]. As previously
indicated, the most common group of accident types with a share of around two-thirds (65 %)
are single-vehicle accidents.
Figure 4: Types of accidents (N=270 exploitable bicycle accidents)
Figure 5 illustrates the distribution of accident triggers. Accident triggers can also be objects
such as streetlights or bollards. The accident trigger not necessarily describes the collision
opponent or the main cause but rather the event that actually triggered the accident. For
example, an accident can be triggered by a crossing pedestrian, but the main reason for the
accident is inappropriate speed of the cyclist. In almost half of all accidents (49 %) the main
trigger was a driving error by the cyclist. Remarkably, there is a high proportion (18 %) of other
triggers. Responses open up a broad field from baggage slipped into the spokes to technical
defects such as broken components.
10
Figure 5: Triggers of accidents (N=270 exploitable bicycle accidents)
For the accident specific development of the BDA, the evaluation of fall events is highly
interesting. Figure 6 shows a 39 %-share of both front and rear wheel lockup regarding all fall
events. Loss of balance (25 %), mainly caused by an interference in lateral dynamics,
represents the second largest fall event. Nose-over motion is quantified to 13 %.
The main reasons cited for front wheel lockup were tram rails, curbs and excessive speed in
curves (often linked with strong braking). In the latter case, most riders noted an unfavourable
combination with slippery road surface. Rear wheel lockup mainly happened after over-
braking the rear wheel with subsequent loss of bicycle control. Considering all fall events of
front and rear wheel lockup, it can be stated that the accident severity and frequency of front
wheel lockup is higher. Reasons for nose-over motion can be divided in three groups. In the
first group, the front wheel is blocked by a penetrating object (e.g. sprig). In the second group,
nose-over happens mainly due to braking with low to medium force while riding downhill. The
last group includes events where speeding or distraction of the cyclists followed by an
emergency braking on surfaces with a high coefficient of friction (e.g. dry asphalt) are main
reasons for the accident.
When drawing conclusions for the potential of the BDA, it is important to consider the
proportion of braking. Unfortunately, the answers of the affected riders were not very
meaningful. For this reason, an estimation of braking probability was permitted in the analysis
of the answers. As figure 7 shows, 25 % of the casualties explicitly stated a brake action. In a
further 16 % of all accidents a brake action can be assumed.
11
Figure 6: Fall events (N=270 exploitable bicycle accidents)
Figure 7: Proportion of braking situations (N=270 exploitable bicycle accidents)
12
4 DISCUSSION
4.1 Resulting Impacts of Underreporting
The first research question in this paper focuses on validity and limitations of official accident
statistics based on police, hospital and expert (e.g. German In-Depth Accident Study, GIDAS)
reports. As shown in the beginning of this paper, known studies [4, 10] come to the conclusion
that especially bicycle accidents suffer from underreporting in official statistics. Although
underreporting is a well-known phenomenon, it has not been researched very well. Specifically
when non-fatal injuries are concerned, even to-date cyclists’ injury data should be used and
interpreted with caution.
The study presented in this paper in the following called “BikeSafe study” due to its
representativeness for the German population provides a solid base for verifying common
assumptions. As determined in previous studies [11-15], an underreporting rate of 91 %
synonymous to less than 10 % reporting rate of bicycle accidents in official statistics can be
confirmed. Furthermore, a proportion of 65 % single-vehicle accidents on all accidents is
found. This result corresponds to studies presented in [16].
One major advantage of the BikeSafe study is the possibility of an in-depth analysis on types,
triggers and fall events of bicycle accidents. Comparing these results with those from official
accident statistics, the second research question of this paper (“Underreporting of bicycle
accidents and resulting impacts on the proportion of single-vehicle accidents”) can be
discussed. The upper chart in figure 8 shows the impact of underreporting on the distribution
of accident types. It is remarkable that accident types, in which motor vehicles are typically
involved, have a much smaller proportion in the BikeSafe study than in GIDAS. On the other
hand, the significantly higher number of single-vehicle accidents in the BikeSafe study is visible.
This obvious impact of underreporting leads to a completely distorted image when
investigating fall events of bicycle accidents (see lower chart in figure 8). Based on the statistics
derived from GIDAS, an active safety system like the BDA does not seem to be a very effective
measure for improved bicycle safety (share of wheel lockup 3 % and nose-over 1 %). The only
significant advantage of the BDA would be that cyclists lose their fear to use the front brake,
because there is no risk of front wheel lockup or nose-over. The resulting stronger deceleration
of the bicycle could avoid more collisions.
Regarding the findings of BikeSafe study, the potential for a system such as the BDA is much
higher (share of wheel lockup 39 % and nose-over 13 %). Nevertheless, as shown in figure 7, it
is important to remember that only a share of 41 % of all accidents occur while braking.
The following comparison of accident triggers allows an evaluation of different methods in
bicycle accident research. In addition to the statistics based on GIDAS and BikeSafe, the
findings of a study based on hospital data supplemented by additional surveys with crashed
cyclists [17] are plotted in figure 9. Assuming accuracy of the BikeSafe study, investigating
healthcare data with additional surveys seems to be a precise method as well. Despite some
discrepancies (e.g. stationary objects), the results of both studies coincide largely.
13
Figure 8: Comparison of self-determined accident types and fall events
with GIDAS [31-32]
Figure 9: Comparison of self-determined accident triggers
with GIDAS [32] and hospital data [17]
14
4.2 Requirements for the BDA
The available electrical energy and the equipment rate with hydraulic brakes on EBs combined
with the possibility to support the cyclist during deceleration are main ideas of the BDA.
Furthermore, EBs have a more favourable mass and cost ratio than CBs, which makes the
initial introduction of an active safety system for bicycles promising. In addition to clarifying
technical topics such as the influence of the dominant cyclist’s mass on vehicle dynamics,
suspension kinematics of bicycles and largely unknown wheel and tire properties, it is very
important to investigate effects on the current accident situation caused by changes in
mobility patterns due to EBs (third research question) for a target group oriented development
of the BDA.
Results of previous studies on EB users’ behaviour [23-27] point out two main groups:
commuters and the elderly. Furthermore, results show that cyclists use the electric auxiliary
motor primarily for convenience. Despite the anticipated increased accident risk of EBs caused
by e.g. greater travel speed, no uniform picture regarding accident risk can be derived from
accident research studies up to now [30-37]. So far, the only consensus is that higher travel
speed results in a higher injury severity as Nilsson’s power model proves [39]. For this reason,
it was considered at the BikeSafe study that there are no significant changes in mobility
patterns or accident risks due to EBs. Hence, a joint analysis of CB and EB users was possible.
One main objective of the second question in the BikeSafe study was to investigate research
question four, namely cyclist behaviour and accident causation factors with a special focus on
both critical braking situations that should be prevented by the BDA. Unfortunately, it was
found that most of the responses lack in-depth technical details. This is due to
representativeness of the study for the German population. More conclusions than the
qualitative statements listed above could not be derived.
4.3 Future Qualitative Research: Explorations based on qualitative Interviews
Further important tasks primarily concern technical assessment of the effectiveness and
usefulness of active safety systems on bicycles such as the BDA as well as obtaining the main
requirements for the design of such a system.
Detailed narrative interviews with carefully selected respondents will serve as an appropriate
survey method. On the basis of a pre-screening questionnaire we will attempt to find about 20
respondents, who in detail remember a crash/an accident that relates to a brake action.
In comprehensive guideline based interviews (approximately 45 - 60 minutes) the accident
cyclists should describe the course of events during the accident as detailed as possible. We
intend to understand in more detail how the front wheel lockup or nose-over accidents are
caused and how they can be prevented.
In addition, we aim to gain first assessments on the general acceptance of bicycles equipped
with active safety systems such as the BDA. First indications on the willingness to buy and pay
are of particular interest and should be quantified in further research.
15
5 CONCLUSIONS
The main objectives of this paper are to provide both quantitative and qualitative information
to estimate the potential benefits of a BDA and to define future requirements for it. For this
reason, a better understanding of both critical braking situations preventable by the BDA (front
wheel lockup and nose-over) has to be gained. The well-known underreporting of bicycle
accidents in official statistics based on police reports poses a particular challenge. Especially
when investigating single-vehicle accidents, in which both critical braking situations mainly
occur, this phenomenon of underreporting needs to be counteracted by effective and
innovative approaches.
Findings in the BikeSafe study include new valuable representative statements for Germany.
Regarding accidents in accordance with type of bicycle, injured riders were traveling to 27 %-
share with a CB and to 2 %-share with an EB. An accident report to the police was made only in
every tenth case. This results in an underreporting rate of 91 %. The conducted in-depth
accident analysis identified single-vehicle accidents with a share of around two-thirds (65 %) as
the most common group of accident types. A driving error by the cyclist was in almost half of
all accidents (49 %) the main trigger leading to the accident. Regarding the most common fall
event, there was a 39 %-share of both front and rear wheel lockup. Nose-over motion was
quantified to 13 %. It remains remarkable that only a share of 41 % of all accidents occur with
braking.
As a major spin-off of the in-depth accident analysisassuming accuracy of the BikeSafe study
it turns out that investigating healthcare data with additional surveys seems to be a very
precise method as well.
6 ACKNOWLEDGEMENT
The authors would like to thank the ministry of education and research in Germany (BMBF) for
sponsoring the project (FKZ 03FH063PX3). Furthermore, the authors thank all academic and
industrial partners for supporting BikeSafe. Finally, the authors would like to thank YouGov
Germany GmbH for supporting the presented empirical study.
REFERENCES
[1] P. Urban and O. O. den Camp, Vulnerable road user safety, EARPA Position Paper,
Brussels, 2013.
[2] ERTRAC, “Safety of vulnerable road users”, in ERTRAC, European Roadmap: Safe Road
Transport, Brussels, 2011, pp. 7-10.
[3] T. Grimaldi, “New urban qualities: Cycling’s (& e-bikes) role in the growing city region”,
International Cycling Safety Conference 2014, Gothenburg, Sweden, 18-19 November
2014.
[4] F. Wegman, F. Zhang and A. Dijkstra, “How to make more cycling good for road safety”,
Accident Analysis and Prevention Vol. 44 (2012), pp. 19-29.
[5] European Commission, Fatalities at 30 days by traffic unit type in EU countries, reference
to CARE database, Brussels, 2013.
[6] D. Eisenberger, Zahlen-Daten-Fakten zum Deutschen E-Bike-Markt 2014, Press Release of
the German Bicycle Industry Association (ZIV), Berlin, 2015.
16
[7] S. Neuberger, Pedelecs Neue Chancen für den Radverkehr, 50. German Council on
Jurisdiction in Traffic, Cologne, 2012.
[8] G. Rose, “E-bikes and urban transportation: emerging issues and unresolved questions”,
Transportation Vol. 39 No. 1 (2012), pp. 81-96.
[9] J. J. de Hartog, H. Boogaard, H. Nijland and G. Hoek, “Do the Health Benefits of Cycling
Outweigh the Risks?”, Environmental Health Perspectives Vol. 118 No. 8 (2010), pp. 1109-
1116.
[10] H. Hautzinger, H. Dürholt, E. Hörnstein and B. Tassaux-Becker, Dunkelziffer bei Unfällen
mit Personenschaden, M13 Reports of Federal Highway Research Institute, Bergisch
Gladbach, 1993.
[11] M. C. B. Reurings and N. M. Bos, Ernstig gewonde verkeersslachtoffers in Nederland in
1993-2008: Het werkelijke aantal in ziekenhuizen opgenomen verkeersslachtoffers met
een MAIS van ten minste 2, R-2009-12 SWOV, Leidschendam, 2006.
[12] C. Juhra, B. Wieskötter, K. Chu, L. Trost, U. Weiss, M. Messerschmidt, A. Malczyk and M.
Raschke, “Bicycle accidents Do we only see the tip of the iceberg? A prospective multi-
centre study in a large German city combining medical and police data”, Injury Vol. 43
(2012), pp. 2026-2034.
[13] U. Weiss, M. Messerschmidt, A. Malczyk, C. Juhra, B. Wieskötter, K. Chu, L. Trost and M.
Raschke, “Fahrradunfallstudie Münster Eine interdisziplinäre Studie des
Universitätsklinikum Münster, der Unfallforschung der Versicherer und der Polizei
Münster”, Die Polizei Vol. 7 (2011), pp. 210-215.
[14] E. Amoros, J.-L. Martin and B. Laumon, “Under-reporting of road crash casualties in
France”, Accident Analysis and Prevention Vol. 38 (2006), pp. 627-635.
[15] R. Elvik and A. Mysen, “Incomplete Accident Reporting: Meta-Analysis of Studies Made in
13 Countries”, Transportation Research Record: Journal of the Transportation Research
Board Vol. 1665 (1999), pp. 133-140.
[16] S. Boufous, T. Senserrick, R. Ivers, D. Richardson and L. De Rome, “Circumstances of on-
road single-vehicle cyclist crashes in the Australian Capital Territory”, International Cycling
Safety Conference 2014, Gothenburg, Sweden, 18-19 November 2014.
[17] K. Bauer, S. Schick, A. Wagner, K. Zhou, S. Peldschus and A. Malczyk, “Untersuchungen zur
Schutzwirkung des Fahrradhelms”, Research Report No. 32, German Insurance
Association, Berlin, 2015.
[18] M. Johnson, J. Charlton, J. Oxley and S. Newstead, “Naturalistic cycling study: Identifying
risk factors for on-road commuter cyclists”, in Annals of Advances in Automotive
Medicine. Annual Scientific Conference Vol. 54 (2010), pp. 275-283.
[19] M. Dozza and J. Werneke, “Introducing naturalistic cycling data: What factors influence
bicyclists’ safety in the real world?”, Transportation Research Record Part F Vol. 24 (2014),
pp. 83-91.
[20] K. Schleinitz, T. Petzoldt, L. Franke-Bartholdt, J. Krems and T. Gehlert, “The German
Naturalistic Cycling Study- Comparing cycling speed of riders of different e-bikes and
conventional bicycles”, International Cycling Safety Conference 2014, Gothenburg,
Sweden, 18-19 November 2014.
17
[21] A. Niska, “Cycling Safety and Infrastructure Maintenance”, International Cycling Safety
Conference 2013, Helmond, The Netherlands, 20-21 November 2013.
[22] EN 15194, Cycles Electrically power assisted cycles EPAC bicycles, German and English
version prEN 15194:2015.
[23] V. Lamy, Electric bike 2000 project, report TP 13732E, Transport Canada, Montreal, 2001.
[24] A. A. Parker, “The electric power assisted bicycle: a clean vehicle to reduce oil
dependence and enhance the mobility of the elderly”, in International Conference on
Sustainability Engineering and Science, Auckland, New Zealand, 2004.
[25] C. R. Cherry and R. Cervero, “Use characteristics and mode choice behaviour of electric
bike users in China”, Transport Policy Vol. 14 (2007), pp. 247-257.
[26] A. A. Parker, “The electric powered assisted bicycle: a clean vehicle to enhance the
mobility of the able, the elderly and people with disabilities”, in New Zealand Cycling
Conference, Auckland, New Zealand, 2003.
[27] A. A. Parker, “Electric power-assisted bicycles reduce oil dependence and enhance the
mobility of the elderly”, in Australasian Transport Research Forum (ATRF), Gold Coast,
Queensland, Australia, 2006.
[28] J. P. Schepers, E. Fishman, P. den Hertog, K. Klein Wolt and A. L. Schwab, “The safety of
electrically assisted bicycles compared to classic bicycles”, Accident Analysis and
Prevention Vol. 73 (2014), pp. 174-180.
[29] M. Kuehn, Safety Aspects of High-Speed Pedelecs, Compact accident research No. 30,
German Insurance Association, Berlin, 2012.
[30] T. Gehlert, “Traffic safety of electric bicycles”, Compact accident research No. 46, German
Insurance Association, Berlin, 2014.
[31] J. Moennich, T. Lich, A. Georgi and N. Reiter, “Did a higher distribution of pedelecs results
in more severe accidents in Germany?”, 6th ESAR International Conference, Hannover,
Germany, 20-21 June 2014.
[32] D. Schmukal, Untersuchung der möglichen Veränderungen des Verkehrsunfallgeschehens
durch Elektrofahrräder im Straßenverkehr, Robert Bosch GmbH, Internal Study, Diploma
Thesis, Schwieberdingen, 2012.
[33] T. Weber, G. Scaramuzza and K.-U. Schmitt, “Evaluation of e-bike accidents in
Switzerland”, Accident Analysis and Prevention Vol. 73 (2014), pp. 47-52.
[34] K. Schleinitz, L. Franke-Bartholdt, T. Petzoldt, S. Schwanitz, T. Gehlert and M. Kuehn,
Pedelec Naturalistic Cycling Study”, Research Report No. 27, German Insurance
Association, Berlin, 2014.
[35] K. Schleinitz, T. Petzoldt, L. Franke-Bartholdt, J. Krems and T. Gehlert, “The German
Pedelec Naturalistic Cycling Study Study Design and First Experiences”, International
Cycling Safety Conference 2012, Helmond, The Netherlands, 7-8 November 2012.
[36] M. Dozza, J. Werneke and M. Mackenzie “e-BikeSAFE: A Naturalistic Cycling Study to
Understand how Electrical Bicycles Change Cycling Behaviour and Influence Safety.”,
International Cycling Safety Conference 2013, Helmond, The Netherlands, 20-21
November 2013.
18
[37] M. Dozza and G. F. Bianchi Piccinini, “Do cyclists on e-bikes behave differently than cyclists
on traditional bicycles?”, International Cycling Safety Conference 2014, Gothenburg,
Sweden, 18-19 November 2014.
[38] GIDAS, “German In-Depth Accident Study”, http://www.vufo.de/forschung-und-
entwicklung/gidas/, Dresden, Germany, 2015, seen 08/26/2015.
[39] L. Aarts and I. van Schagen, “Driving speed and the risk of road crashes: A review”,
Accident Analysis and Prevention Vol. 38 (2006), pp. 215-224.
Conference Paper
Full-text available
Safety is a concern jeopardizing cycling sustainability. In Europe, more than 2.000 cyclists die every year in traffic accidents. Electrical bicycles amplify this concern because of their high speed and increasing prevalence. Previous studies, such as the naturalistic cycling study BikeSAFER, showed that interaction among cyclists and other road users (e.g., drivers, pedes‐ trians or other bicyclists) is crucial to cycling safety. As electrical bicycles become increasingly popular, other road users may need to recalibrate their expectations to maintain a safe inter‐ action with this new type of bicycle. However, the extent to which electrical bicycles fail to meet other road users' expectations because of their higher speed is yet to be confirmed. Fur‐ thermore, whether the growing number of electrical bicycles influences cycling behaviour and results in new safety challenges is still unknown. The e‐BikeSAFE project is currently collecting naturalistic cycling data in Gothenburg. Equipped bicycles in e‐BikeSAFE presently record cyclist behaviour in real traffic from a camera GPS, and kinematics sensors. This data will be used to 1) understand how bicyclists with electrical bicy‐ cles behave in traffic and 2) the extent to which safety critical situations (crash and near‐ crashes) are different for electrical bicycles compared to traditional ones. For this later analy‐ sis, data from BikeSAFE will be used as a reference. This paper explains how the naturalistic methodology was adapted to the electrical bicycles and shows how e‐BikeSAFE data will be combined with existing data from BikeSAFE for safety analyses including cycling behaviour and accident causation.
Conference Paper
Full-text available
Cycling is a healthy, environmentally-friendly and enjoyable activity, which unfortunately also claims more than 2000 lives every year in Europe. Many municipalities across Europe are waging successful campaigns to increase cycling and, as a consequence, reduce pollution and congestion. However, at least in the short term, a surge in cycling is also challenging existing infrastructure, regulations, and the interaction among different road users. Further, the nature of cycling is chang-ing as new electrified bicycles (e-bikes) become more prevalent, since they are able to maintain a constant 25km/h speed independent of road gradient or wind. The extent to which e-bikes preva-lence impacts safety is currently unknown and very hard to simulate with statistical models. In 2012, the BikeSAFE project collected 1474 km of naturalistic cycling data from traditional bicy-cles. Similarly, in 2013, the e-BikeSAFE project collected 1549 km of naturalistic data from e-bikes. All studies took place in the urban area of Göteborg in the same period of the year, and involved the same participants as much as possible. While these naturalistic data sets are limited and pos-sibly not representative of the cycling situation in all of Europe, they are also the most advanced data available today for comparing how traditional and electrical bicycles behave in traffic, thus offering a promising test bed for developing data analysis methodologies. Five random video clips of 30 seconds duration were extracted for each participant from the data collected in BikeSAFE and e-BikeSAFE, forming an overall analysis database of 140 full HD video clips. Video reduction identified which road users were involved in interactions with the bikes (tra-ditional or electric). During the analysis, potential influencing factors (e.g. width, gradient, and cur-vature of the cycle path) were also taken into account. Information from the video reduction of e-bikes and traditional-bikes was compared by means of odds ratios and combined with subjective data from questionnaires, to determine the extent to which safety concerns about e-bikes are le-gitimate. Results show that e-bikes and traditional bicycles are ridden differently: cyclists riding e-bikes ex-perience different, more frequent interactions with other road users, and prefer different riding conditions, possibly because of their higher speed. Further, infrastructure (such as crossings) and secondary tasks (such as using a phone) may be particularly dangerous for e-bikers. The results presented in this paper provide new ideas for the design of safer cycle paths and more conspicu-ous e-bikes.
Conference Paper
Full-text available
This paper presents the study design and first experiences of a Pedelec Naturalistic Cycling study. There are 90 participants: 30 bicylists, 50 Pedelec cyclists and 10 E-bike cyclists. The bicycles are equipped with a data acquisition system that records among others speed data and videos on the traffic situation over a period of four weeks. Questionnaires assessing current travel and traffic behaviour and changes thereof, motives and experiences with Pedelecs / E-Bikes are used when recruiting participants, before and after the observation period. A one-week time use travel diary was used to collect qualitative information on the cycle trips and related activities. Despite a low modal share of bicycling in the study area there were no problems recruiting participants. Recruiting E-bike user proved to be a challenge as their market share in Germany is quite low. Participants are very cooperative even though the study procedure puts quite considerable demand on them. The data acquisition system provides reliable trip and video data, even though there are problems with the GPS data. Thus we expect an exceptional dataset that will improve our understanding of travel and traffic behaviour of E-Bike users.
Book
Full-text available
The popularity of electric bicycles, so called pedelecs, has increased rapidly over the last few years. Their numbers are expected to grow even further in the years to come. Given this increase in prevalence, new challenges for road traffic arise. The effects of potentially higher speeds that might be achieved with electric bikes on road safety are unclear. Likewise, the fact that older riders currently are the main user group for this type of bike leads to questions about physical fitness and increased safety risks. Changes in overall mobility behaviour can be a consequence as well. Within this project, aspects of mobility and safety behaviour of cyclists were investigated in a so called naturalistic cycling study. Core of this methodological approach is the instrumentation of bicycles with cameras and other sensors to observe riders over a specified period of time, to capture riders’ “normal” cycling and mobility behaviour. Overall, 90 cyclists were recruited for this study. Forty-nine of them were owners of a pedelec25. This type of electric bike, which supports pedalling up to a speed of 25 km/h, is treated legally like a regular bicycle. Ten participants were riding pedelec45, which support pedalling up to 45 km/h. When using this kind of electric bike, riders are required to use a helmet, and the bike needs to carry a license plate. An additional 31 participants were recruited that drove regular bicycles. To account for potential age effects, roughly one third of the participants was 40 years or younger, between 40 and 65, and 65 or older, respectively. Each of the 90 participants was observed with the instrumentation over a period of four weeks on his/her trips. In addition, for one of the four weeks, participants were required to fill in a mobility activity diary to track all of their trips (not just bike) and trip purposes. Furthermore, they completed questionnaires regarding aspects such as user behaviour and accident history before and after data acquisition. Overall, more than 4,000 trips with a total length of roughly 17,000 km were recorded. To uncover potentially safety relevant traffic situations, an extensive video annotation was conducted. To accomplish this, every single trip was screened for potentially critical situations, which were identified and annotated using a predefined coding scheme. In addition, trip length, trip duration and especially speed were analysed using wheel sensor data. The results of the study show that the use of electric bicycles did not result in an increased number of critical situations compared to regular bicycles. Age did not affect the relative frequency of occurrence of critical situations as well. When taking into account the different conflict partners, it appeared that most frequently, cyclists and ebike riders experienced conflict situations with motorised vehicles such as cars or lorries. However, also a substantial number of conflicts between (electric) bicycles and pedestrians could be found. The analysis of the sensor data revealed the expected differences regarding average riding speed. On average, pedelec25 riders travelled somewhat (ca. 2 km/h), pedelec45 riders substantially (ca. 8 km/h) faster than riders of regular bicycles. There were also profound age effects in the expected direction. In contrast, relevant differences in mobility behaviour between bike and e-bike users could not be found. Given the results of this project, the current regulations, which treat pedelec25 as a regular bicycle, and put up stronger requirements for the use of pedelec45, appear to be 9 reasonable. The comparable number of critical situations for all three vehicle types suggests that the risk to be involved in an accident is not increased for electric bicycles. At the same time, however, the fact that pedelec45 are travelling, on average, much faster than the other two vehicle types, suggests that the risk of severe injury in case of a fall or a collision is considerably higher for the faster e-bikes. Overall, however, it appears that a continuing assessment of the issue is required. Currently, user groups and use cases for the different types of bikes and e-bikes differ substantially. Whether there will be transformations within the user groups, and whether this will result changes in road safety relevant measures, needs to be closely observed.
Article
In recent years, the number of electric bicycles on European, American and especially Chinese roads has increased substantially. Today, 11% of all bicycles sold in Germany are e-bikes. Given their potential to reach higher maximum speeds, concerns have been raised about a possible increase in crash risk associated with e-bike use. However, as of now, it is unclear if and how often the potentially higher speed is actually reached in everyday cycling. As part of the German Naturalistic Cycling Study we measured and compared the speed of three bicycle types (conventional bicycles, pedelecs (pedalling supported up to 25. km/h), S-pedelecs (pedalling supported up to 45. km/h)) under naturalistic conditions. Ninety participants, divided in three age groups, took part in our study. Participants used their own bikes or e-bikes. The bicycles were equipped with a data acquisition system, which included sensors to record speed and distance, as well as two cameras. Data was collected over a period of four weeks for each participant. Nearly 17,000. km of cycling were recorded in total. The statistical analysis revealed significant differences in mean speed between all three bicycle types. Pedelec riders were, on average, 2. km/h faster than cyclists. S-pedelec speed was even 9. km/h higher. A similar pattern was also found when analysing free flow conditions and uphill or downhill cycling separately. The highest speed was measured on carriageways and bicycle infrastructure, regardless of bicycle type. Participants aged over 65. years rode significantly slower than younger participants. Data on acceleration from standstill largely confirm the differences between bicycle types and age groups. The results show that electric bicycles indeed reach higher speeds than conventional bicycles regularly. Although it is unclear if this also leads to an increase in crash risk, it can be assumed that the consequences of a crash might be, on average, more severe.
Article
Use of electrically assisted bicycles with a maximum speed of 25 km/h is rapidly increasing. This growth has been particularly rapid in the Netherlands, yet very little research has been conducted to assess the road safety implications. This case–control study compares the likelihood of crashes for which treatment at an emergency department is needed and injury consequences for electric bicycles to classic bicycles in the Netherlands among users of 16 years and older. Data were gathered through a survey of victims treated at emergency departments. Additionally, a survey of cyclists without any known crash experience, drawn from a panel of the Dutch population acted as a control sample. Logistic regression analysis is used to compare the risk of crashes with electric and classical bicycles requiring treatment at an emergency department. Among the victims treated at an emergency department we compared those being hospitalized to those being send home after the treatment at the emergency department to compare the injury consequences between electric and classical bicycle victims. The results suggest that, after controlling for age, gender and amount of bicycle use, electric bicycle users are more likely to be involved in a crash that requires treatment at an emergency department due to a crash. Crashes with electric bicycles are about equally severe as crashes with classic bicycles. We advise further research to develop policies to minimize the risk and maximize the health benefits for users of electric bicycles.
Article
Background: The acceptance and usage of electric bicycles has rapidly increased in Switzerland in the last years. Hence this topic has been addressed by policy makers with the aim to facilitate new transport modes and, moreover, to improve their safety. Methods: Police-recorded accidents of the years 2011 and 2012 involving a total of 504 e-bikers and 871 bicyclists were analysed. National figures were compared with those of a rural and an urban environment. Results: Most e-bikers who were involved in accidents were 40-65 years old. It was found that most e-bikers sustained single accidents and that helmet usage was higher in the investigated rural environment than in the investigated urban area. The evaluation of the injury severity of e-bikers, particularly compared to bicyclists, lead to diverging results. Conclusions: The findings presented in this study are intended to serve as a benchmark since basic information on characteristics of e-bike accidents is provided. With respect to differences between the injury severity of e-bikers and bicyclists to-date no clear statement can be drawn. It is suggested to regularly evaluate e-bike accidents to show trends and/or identify changes.
Article
Presently, the collection and analysis of naturalistic data is the most credited method for understanding road user behavior and improving traffic safety. Such methodology was developed for motorized vehicles, such as cars and trucks, and is still largely applied to those vehicles. However, a reasonable question is whether bicycle safety can also benefit from the naturalistic methodology, once collection and analyses are properly ported from motorized vehicles to bicycles. This paper answers this question by showing that instrumented bicycles can also collect analogous naturalistic data. In addition, this paper shows how naturalistic cycling data from 16 bicyclists can be used to estimate risk while cycling. The results show that cycling near an intersection increased the risk of experiencing a critical event by four times, and by twelve times when the intersection presented some form of visual occlusion (e.g., buildings and hedges). Poor maintenance of the road increased the risk tenfold. Furthermore, the risk of experiencing a critical event was twice as large when at least one pedestrian or another bicyclist crossed the bicyclist’s trajectory. Finally, this study suggests the two most common scenarios for bicycle accidents, which result from different situations and thus require different countermeasures. The findings presented in this paper show that bicycle safety can benefit from the naturalistic methodology, which provides data able to guide development and evaluation of (intelligent) countermeasures to increase cycling safety.