In Germany, electric bicycles (pedelecs) have become highly popular over the past few years. Reasons for that are their potential to reach higher speeds and the reduction of
cycling effort. While these are desirable effects, safety concerns have been raised.
Pedelecs are, with regard to their design, hardly distinguishable from conventional
bicycles. It has been argued that this could result in other road users misjudging the
approach of an oncoming pedelec (e.g. at intersections) and subsequent unsafe
behaviour, e.g. choosing rather small time gaps for crossing in front of pedelec riders.
Therefore, the goal of this study was to conduct a series of experiments investigating
road users’ time-to-arrival estimations (TTA; the estimation of time gaps between the
road user and other vehicles) and their gap acceptance behaviour (the gap a road user
selects in front of or between other vehicles for turning or crossing) in relation to
approaching two-wheelers, especially pedelecs and conventional bicycles.
Three experiments were conducted to investigate the influence of two-wheelers’
approach speed, bicycle/vehicle type (conventional bicycle, S-pedelec, scooter), road
gradient, observer perspective, observer age, cyclist age, and pedalling frequency on
TTA estimation and/or gap acceptance. A fourth experiment investigated the impact of
measures to enhance the visibility of cyclists on TTA estimation.
An intersection scenario was implemented either in a realistic setting on a test track
(Experiment 1) or in a laboratory using video material (Experiments 2 to 4). On the test
track, participants were seated in a real car. They observed an approaching cyclist and
were instructed to depress a foot pedal to indicate the smallest acceptable gap to turn in
front of the bicycle rider (gap acceptance). In the laboratory studies, participants were
asked to indicate the smallest acceptable gap, too, but provided their judgements on the
basis of videos of approaching cyclists. For TTA estimations, participants watched short
videos of approaching two-wheelers. Before the rider reached the position of the
participant, the videos were masked, and the participants were required to press a button
to indicate the moment they believed the rider would have reached a predefined position
(TTA estimation, Experiment 2 to 4 only).
In all four experiments, a higher approach speed of the two wheeler lead to higher TTA
estimates and smaller accepted gaps in comparison to a lower approach speed. That
means participants’ turning decisions tended to be riskier for higher speeds. In addition,
there were differences in gap acceptance and TTA estimation between the two bicycle
types. Participants selected smaller gaps for the S-pedelec compared to the bicycle.
Likewise, TTA estimations for the S-pedelec were higher, i.e. participants judged the time
remaining until the S-pedelec reached the observers’ position as longer compared to the
bicycle, resulting in riskier turning behaviour. For the scooter, we found larger accepted
gaps and smaller TTA estimations in comparison to the two bicycle types. This suggests
a somewhat safer turning/crossing behaviour around scooters compared to bicycles.
Furthermore, cyclists’ age influenced TTA estimations, with an older cyclist being judged
as arriving earlier at the observers’ position than a younger cyclist. In addition, the
participants chose smaller gaps and provided higher TTA estimations for a lower
pedalling frequency in comparison to the higher one. The age of the participants affected
only TTA estimation, not gap acceptance behaviour. Throughout the experiments, older
participants judged TTA as being smaller compared to the younger participants. The
measures for enhancing the visibility of the cyclists had no influence on TTA estimation.
The results demonstrate that the higher speed of pedelecs together with their similarity to
conventional bicycles could indeed result in road safety problems. Simple measures that
mainly aim at increasing cyclist visibility do not seem to be suitable to change gap
acceptance and speed judgement of pedelecs. The results indicate that observers base
their speed judgements and turning/crossing decisions on their experience with
conventional bicycles. Therefore it may be more appropriate to develop a unique pedelec
design to clearly distinguish it from a conventional bicycle. In the future, it might be that
with even more pedelecs on the road other road users will adapt and the difference
between pedelecs and bicycles in speed judgments and turning / crossing decisions will
disappear. At this point, however, this is mere speculation.