| Artxibo OCR | nsnsaan
Objective measures of overtaking
Photo: University of Bath
Background
Collisions as drivers overtake cyclists are
often serions!1!, but little is known about
what affects the distance drivers leave
when passing. This study collected objec-
tive measures of passing behaviour to
investigate effects of riding position,
helmet use, vehicle type and apparent
gender on overtaking proximities.
Method
A bicycle subtly fitted with a video cam-
era and ultrasonic distance sensor re-
corded around 2,300 vehicles overtaking
on a range of urban/suburban road
types similar to those encountered on a
commute. Further data were then col-
lected with the (male) experimenter
wearing a feminine wig, in order that he
appeared to be a woman to drivers ap-
proaching from behind.
Results
Riding position and helmet
0.25m 0.50m 0.75m 1.00m 1.25m
Bicyclist's distance from road edge
—o—Helmet —■—No helmet
contrary to what many experienced
bicyclists believed should happen. Driv-
ers also tended to pass notably closer to
the rider when he wore a helmet (white
circles) than when he did not (black
squares). Riding position and helmet-
wearing accounted for 8% of the vari-
ance in overtaking proximities. The hel-
met effect was due to shifts in overtaking
distributions rather than qualitative
changes. The position effect operated in
a similar way.
A second analysis calculated the likeli-
hood of a given overtaking event being
KEY FINDINGS
When overtaking the test
bicycle, drivers passed closer
when the experimenter:
• rode towards the centre of
the lane rather than the
edge
• wore a helmet
• appeared male rather than
female
The helmet effect is likely the
result of drivers judging cy-
clists’ skill levels from their
appearance and adjusting
their overtaking accordingly
Drivers of buses and heavy
goods vehicles got signifi-
cantly closer than other vehi-
Drivers passed closer to the rider the
further out into the road he was. This is
0.25m 0.50m 0.75m 1.00m 1.25m
Bicyclist's distance from road edge
particularly near to the rider or particu-
larly far away. The graph above shows
the ratio of particularly near to particu-
larly far events, demonstrating that driv-
ers were more likely to get particularly
close when the rider was towards the
centre of the road or wearing a helmet.
A logistic regression was relatively suc-
cessful (15.2% better than chance) at
predicting whether a driver would come
close or not from the rider’s position and
helmet status.
www.bath.ac.uk | www.drianwalker.com/overtaking
University of Bath, United Kingdom
PROFESSIONAL BRIEFING DOCUMENT
AUGUST 2006
Vehicle type
As shown above, professional drivers of
large vehicles tended to get particularly
close when passing.
Close proximities for large vehicles
were particularly seen towards the end of
overtaking manoeuvres. We believe a
reluctance to travel on the wrong side of
the road for the time taken to pass a
cyclist (4-5 s) leads to drivers often mov-
ing back across before it is really safe to
do so.
Gender effects
When the experimenter was wearing a
wig, so that he appeared female to driv-
1.45
1.4
ers approaching from behind, passing
traffic gave significantly more leeway
Time of day effects
Overtaking varied as a function of time.
There was an increasing quadratic trend,
with the curve accounting for 64% of
variance in the graph above. In general,
drivers passed closer earlier in the day.
Note that the graph shows one-hour
bands (so 0700 means 0700-0759). Ac-
cordingly, there seem to be differences
between the morning and evening rush-
hours.
Discussion
Research suggests drivers tend to believe
helmeted cyclists are more serious and
less likely to make unexpected moves
the helmet effect seen here is likely a
behavioural manifestation of this belief.
The gender effect could be the result of
female cyclists being rarer than male
cyclists in the UK, or it may again be
related to drivers’ perceptions of rider
experience and predictability. Motorists
might profitably be warned not to make
assumptions about a bicyclist’s likely
experience, or their likely actions during
an overtaking manoeuvre, based on their
appearance, as in reality gender and
helmet-wearing will not be particularly
valid signs of experience.
The riding-position effect suggests
drivers simply do not change their over-
taking paths very much as a function of
where a rider is: if a cyclist rides further
into the road, they will on average be
closer to passing vehicles as a result.
However, there are also plenty of rea-
sons why riders should not just stick to
the road edge, e.g., debris, car doors, and
drivers’ attention patterns at junctions.
Long vehicles can need to cross into
the opposite lane for several seconds to
overtake a cyclist. Drivers should be
reminded about how they get too close,
and cyclists might also better understand
how difficult it is for these vehicles to
find overtaking opportunities in urban
environments.
References
1. Murphy Jones, C. & Walker, I. (sub-
mitted). How types of pedal cycle acci-
dents in Oxfordshire, England vary with
age and sex of cyclist.
2. Basford, L. et al. (2002). Drivers’ Per-
ceptions of Cyclists. TRL Report 549.
3. Haddad, H. (2005). Cycling Stereotypes
and Identity. Unpublished Master’s thesis,
University of Surrey.
A full account of this study is to be published
soon: Walker, I. (in press). Drive/s overtaking
bicyclists: Objective data on the effects of riding
position, helmet use, vehicle type and apparent
gender. Accident Analysis and Prevention.
The work was funded by the EPSRC and the
instrumented bicycle was designed and con-
structed by Jeff Brewster from the University of
Bath Mechanical Engineering Department.
Ian Walker can be contacted at
i. walkerfpbath. ac. uk
www.bath.ac.uk | www.drianwalker.com/overtaking
University of Bath, United Kingdom | |
