Session VII

Road safety, Road marking and street furniture

Pavement skid resistance versus appropriate signage to tackle the visibil-
ity problem on motorway fast lanes along tight left curves
A. Kokkalis & A. Athanasopoulou
Associate Professors, Dpt. of Civil Engineering D.U.Th., 67100, Xanthi, Greece, akokkal@civil.duth.gr
G. Kollaros
Assistant Professor, Dpt. of Civil Engineering D.U.Th., 67100, Xanthi, Greece

ABSTRACT: Obstructed visibility is rather common at tight left downgrade curves along motorway fast lanes
in mountainous terrain. The problem is due to the safety barriers in motorway medians which obscure visibility
to stopping sight distance for the typical obstacle. Stopping sight distance should be provided for V85 along any
motorway section. However, a usual combination of the minimum acceptable radius for the horizontal curves,
together with the maximum allowable downgrades, results to provided visibility to low obstacles correspond-
ing to as much as 50km/h lower than V85 vehicle speeds. To overcome the problem either a troublesome shift
of the barrier to the inside of the curve should be made or excessively anti-skid fast lane pavements should be
provided. Also, certain Design Guidelines accept significant relaxations to the values of vehicle speeds to stop
safely ahead of low objects. In this paper the upper limits of anti-skid properties of motorway pavements are
first evaluated. The gap between required values and the possibly achieved ones is identified. Hence, comple-
mentary vertical signage would be necessary to warn drivers. An evaluation of that signage is made, whether it
should inform the drivers how much they have to reduce their speed under wet pavement conditions, or what
other behaviour have to adopt. The paper advocates that a non-passing signage is the best along these motor-
ways subsections.

1. THE VISIBILITY ISSUE ΑLONG MOTOR-
WAY FAST LANES
Roads crossing mountainous terrain usually have to
follow the lowest acceptable design parameters for
the selected design speed (Ve), i.e. minimum radii
and maximum gradient, or the cost will sharply raised
due to increased earthworks and costly engineering
works. Highway engineers know by experience that
when they design a motorway crossing mountainous
terrain they have to tackle the keen visibility defi-
ciency along the fast lanes. As a matter of fact, long
bridge parapets and safety barriers on a motorway
central reserve area along left curves may obscure
visibility to stopping sight distance to the typical low
height object. The problem becomes keener in down-
grades which increase safe stopping distance.
New Jersey barriers (NJ) are the typical safety
barriers on Greek motorways central reserves. A
(short) NJ has an average typical height of 0,83m.
Considering the vision line for the typical eye height
of 1,06m (OMOE, 2002), the visibility to the vehicle
ahead is unobstructed over the safety barrier. How-
ever, what would be the case if a large object fell
from the leading vehicle or drifted or rolled and
seated on the fast lane? According to Table 10-4 of
(OMOE 2002), the basic obstacle height for
V85=120km/h is 0,35m, where V85 is the 85th higher
percentile speed. That size of obstacles should be
early visible by a driver moving that fast either to
avoid them or brake before them and should not be
hidden by the central reserve motorway barrier.
Highway Design Guidelines, in general, propose for
motorways combinations of horizontal curves and
gradients with inadequate visibility to safe stopping
along fast lanes. Along mountainous terrain it is al-
most inevitable to use tight horizontal curves (R).
Obviously, the term ‘tight’ is relevant. By that term
we mean any radius used corresponding to the design
speed (Ve) or (Ve+10)km/h of the motorway. Thus
for class A roads and Ve = 100km/h, any curve with
R<600m is considered as tight.
The motorway visibility problem can be outlined
(Kokkalis, 2011) by the typical horizontal layout ge-
ometry of the problem (Fig. 1). The various assump-
tions related to distances, widths and positions are
presented there. Considering a typical left hard
shoulder width of 0,975m, the vision line and the
typical profile of a NJ barrier it is calculated that a
typical value for ‘a’ equals to 1,15m. Thus, the dis-
tance of the eye to the NJ side wall equals to:
a+b = 1,15+1,75 = 2,9m (1)
Resolving the problem (for R=600m), we have
that the critical sight distance equals to:
Ssd=2√[(600+2.9)2-6002]=118m
Figure 1. Geometric model to calculate critical visibility length
(BC line) along left turns in dual carriageways
In any place along a road-lane, the available sight
distance should be at least the required total stopping
distance (ssd). Ssd depends on vehicle velocity, usu-
ally V85, road gradient and friction values between
the tyre end the pavement. Obviously, at downgrades
more distance is required for a vehicle to stop safely,
since gravity adds to vehicle inertia. In mountainous
terrain, it is most probable that the maximum accept-
able values for road gradient will be commonly used.
For Ve=100km/h, this maximum acceptable gradient
equals to s=6%. This downgrade slope adds consid-
erably to the required ssd. Figure 2 (OMOE, 2002)
outlines the required visibility lengths for safe stop-
ping. According to that and for a downgrade s=6%
ssd amounts to 240m. This is twice as much as the
available 118m calculated in (2).
Hence, the problem of insufficient visibility to ob-
stacles along motorway fast lanes is keen. Even in
flat road sections, along tight left horizontal curves
the ssd amounts to 215m (Fig. 2), far greater than
the available.
Figure 2. Required visibility length for safe stopping Sh or ssd,
(OMOE, 2002)
(2)
The geometric key to resolve the problem is to in-
crease (a+b) distance (Fig. 1). This requires the both-
ersome and safely ambiguous shift of the NJ barrier
to the inside of the circular curve.
To resolve the problem in the most adverse (but
not infrequent) combination of data (R=480m for
Ve=100km/h & s=6%), we have:
240/2=√[[480+(a+b)] 2-4802]  (a+b)=14,8m (3)
which means many meters shift of the NJ barrier to
the inside of the curve.
2. WAYS TO TACKLE THE MOTORWAY VISI-
BILITY ISSUE
2.1 Shifting the central motorway barrier
As mentioned, adequate visibility could be obtained if
the NJ barrier would be shifted to the inside of the
curve. Considering eq. (1), this means that a 14,8 -
2,9 = 12m shift is required (Fig. 1). In similar visibil-
ity problems at the slow lanes of a mountainous road,
(along right tight curves in cuts), it is common prac-
tice to extend earthworks and level the slopes for
unobstructed sight line. This action results to corre-
spondingly wider construction limits and, usually,
considerably deeper cuts and greater amount of
earthworks, all acceptable to achieve safe drive.
Apart from the above, considering to apply that wid-
ening along a motorway central reserve exhibits cer-
tain design difficulties: the additional pavement width
should be kept without trafficking, usually through
zebra delineation or banking, a practice which may
induce unsafe driving practices. Actually, along Eg-
natia motorway this practice has been applied in only
one case with controversial results. Hence, tackling
the problem by shifting the NJ barrier to the inside
part of the turn is not a good practice.
2.2 Adopting relaxations plus signage
The simplest way Road Design Guidelines in various
countries, tackle the problem is to consider the re-
quired visibility for Ve, instead of, for V85. For the
data of the problem outlined above, this assumption
decreases the required visibility length from 240m to
180m. A 5,4m shift to the inside of the turn is still
required for the NJ barrier. The problem is seriously
alleviated but not resolved.
Cypriot road geometric design standards (2002),
tackle the problem by adopting further relaxations.
According to them, a relaxation may be adopted
where the strict application of design standards
would lead to disproportionately high costs. A re-
laxation means the adoption of design figures for cer-
tain design parameters referring to a one or more
steps lower Ve. It should be mentioned herein, that
the steps for Ve are not the typical 10km/h, but the speed is for V85. According to the Highway Guide-
consecutive values Ve takes are: 120, 100, 85, 70, 60 lines (OMOE, 2002), for motorways there is a rough
and 50 km/h. The relaxation for the required mini- relationship between Ve and V85:
mum stopping sight distance to the typical 0,26m ob- V85 = (Ve + 30)km/h, when Ve < 100km/h and
ject height may be extended to two Ve steps, on the V85 = (Ve + 20)km/h, when Ve ≥ 100km/h (4)
condition that the required (for Ve) sight distance
above the safety barrier to the 1,05m object height Table 1. Vision deficit to low objects along tight left
(representing the stop lights of the front vehicle) is downgrade fast lane motorway curves for V85
available. This means that for a Ve=100km/h, ssd to Ve Min. Max. Avail- V85= Ssd Ssd/
low obstacles should be provided for 70km/h, instead (km/h) R Down able Ve+20 (m) Avail-
for the strictly required V85 speed of ~120km/h. Ac- (m) -grade vision or 30 able
cording to Table 3.5 of the Cypriot road geometric (%) (m) (km/h) vision
design standards (2002), the ‘two step below desir- 80 280 7 81 110 210 2,6
able minimum’ stopping sight distance is, for 90 370 7 93 120 250 2,7
Ve=100km/h, only 120m. This is actually the value 100 480 6 106 120 240 2,3
which the horizontal layout of the problem (Fig. 1, 110 600 5 118 130 270 2,3
eq. (2)) provides. It seems that the Cypriot Standards
resolve the problem and no further action is required. From Table 1 it appears that for V85 the required
Actually, after a parametric study of the problem, it vision length to a low obstacle is, for all cases, more
has been concluded that the Cypriot Standards re- than two times the available ones. Even if consider-
solve the problem in its generalized form: the typical ing V70, (approximately equal to (Ve+10), Kokkalis
horizontal layout of the visibility problem at left et al., 2011) the ratio ssd/available vision line had
curves (Fig. 1) with the minimum required R for a values around 1,8 (Kokkalis, 2011). The question
specific Ve, always provide the required visibility to arises, what would be the target pavement friction
stop ahead of a low obstacle for a two-step, 30km/h, values which would reduce safe stopping to the
Ve-relaxation, on the assumption that the downgrade available vision line?
slope is not too near to the highest permissible. The mathematics behind the calculated ssd values
Hence adopting convenient relaxations is a conven- need to be outlined. Relevant curves have been de-
ient way to radically resolve the issue. Obviously rived connecting tangential friction factors (f) with
relevant speed limit signage is necessary. The relaxa- speeds (Kokkalis & Panagouli, 1998). The shape of
tion-adoption route may have a sound fiscal docu- all these curves is similar. Data correlations result in
mentation but, undoubtedly, leaves serious engineer- a similar polynomial equation of the type:
ing issues half-answered. f = constant term-a1(V)+a2(V2) (5)
where a1 is approximately ~3a2
2.3 The insurmountable friction deficit Because there is a great variety of data (tyres,
pavement surfaces and conditions) and highway en-
Generally, sufficient pavement skid resistance should gineers have to cater and provide adequate safety for
be regarded as one of the most important safety prac- adverse conditions, the (lowest pavement friction)
tice when considering motorway safety. The neces- 85th to 95th percentile distribution curves are adopted.
sary sight distance for safe stopping would be radi- Thus, the actual friction values are higher along the
cally reduced if the values of pavement friction 85% to 95% of the total number of road sections.
would be high. Highway Guidelines (OMOE, 2002), According to Lamn et al., (1999), a new evaluation
require adequate visibility for safe stopping to be data background (for Germany), yields for skid resis-
provided for V85. For normal friction values, the tance the equation:
Guidelines adopt the key parameters which are out-
lined in Table 1. It should be commented that the f =0,708-0,721(V/100)+0,241(V/100)2 (6)
adopted friction values are conservative, to make up where the variable V (speed) is given in km/h.
for these cases where old tyres run on polished wet Also, based on multi-nation data an overall re-
pavements. The adoption of an adverse combination gression equation has been developed (Lamn et al.,
of parameters is normal practice in engineering safety 1999):
issues. f = 0,59-0,485(Ve/100)+0,151(Ve/100)2 (7)
In Table 1 the visibility deficit problem is pre- 2
The coefficient of determination (R ) of the equa-
sented for Ve = 80, 90, 100 and 110km/h, the typical tion equals to 0,731 and the standard error of esti-
design speed range for motorways in mountainous mate to 0,044.
terrain. The minimum horizontal radius as well as the The standard method used to measure skid resis-
maximum permitted gradient has been considered. tance is a test vehicle driven with constant speed. An
Vision lines have been calculated according to equa- additional ‘measurement wheel’ is carried which is
tion (2). Stopping sight distances (ssd), have been braked on wet pavement conditions and the force is
taken from Figure 2. The appropriate operation measured in relation to the weight. SCRIM vehicle,
the Stuttgarter tribometer (SRM) and the Swedish
skiddometer are typical equipment used. The equa-
tion (6) has been derived on measurements taken by
SRM.
The friction quality of Greek pavements is, as ex-
pected, considerably poorer. It is estimated that ap-
proximately 20% of wet pavements in Greece could
be covered by using the overall regression equation
(7). Actually, pavements in national or provincial
roads in Greece are very slippery. This is due to the
fact that limestone aggregates are used and resurfac-
ing works rarely take place. However, if we restrict
only to motorway pavements, where anti-skid surfac-
ing is the rule, a similar equation may not comes
short. Furthermore, the same fact that we restrict
only to motorway pavements, where higher quality
control standards apply, would result in a tighter per-
centile distribution curve, which may permit the use
of a shifted equation yielding slightly higher friction
values: higher constant term due to harsh micro-
texture and even a slightly gentler reduction of f-
values with speed, due to maintained macro-texture.
Also, the good drainage conditions, usually estab-
lished on motorway pavements, ensures that rainfall
water is promptly run off the pavement surface, thus
the water film is not thick enough to cause a sharp
drop in friction with higher vehicle speeds, permits a
polynomial curve of milder slope (like the one meas-
ured in USA pavements (Lamn et al., 1999)), to be
adopted.
In this effort the mathematics involved are out-
lined below (Kokkalis, 2011).
Ssd comprises of two terms, the reaction distance
(S1) and the braking distance (S2). Following the ba-
sic equation from physics, we have:
S 1  Vt R / 3,6 (8)
V Thus, ahead of a tight left turn a composite speed
S 2  V / f (V )  ( s / 100 )  ( D / M ) dV (9)
where, tR is driver perception-reaction time,
0
f(V) is equation (7),
s is grade (%), (+ for upgrade and - for downgrade),
D is vehicle aerodynamic drug force and
M is vehicle mass in kg.
Equations (8) & (9) have been applied to the lines
of Table 1. The following assumptions have been
made:
 The shape of the polynomial equation (5) remains
the same,
 The polynomial equation may have a higher con-
stant term (friction for almost zero speed),
 The operational speed used is V85,
 Overtaking is a maneuver requiring higher driver
alert (even in motorways) than the average design
value of 2 seconds for perception-reaction time
(for category A roads) considered in equation (7).
Perception and reaction time is taken to 1,5sec,
still conservative for an overtaking maneuver.
By equating the available vision line to the re-
quired ssd, the unknown was set the constant of
equation (7). The result is that the constant term
should take values as high as 1 to fulfill equality.
How realistic could be to accept such high initial
friction values, to secure that the available vision line
to low obstacles without any barrier shift to the in-
ternal of the curve is adequate for safe stopping? The
highest values for the constant term of the regression
equations (6), (7) are around 0,8 and they refer only
to circuit pavements. Hence, it seems that this route,
to tackle the insufficient visibility along tight down-
grade left motorway curves by providing very high
friction pavements along the fast lanes has not even a
theoretical resolution.
2.4 Considering friction values in dry and wet
pavement conditions plus signage
It is known that most dry pavements provide ade-
quate skid resistance but in wet pavements conditions
skid resistance values drop considerably: the water
film should be expelled to permit tyre-pavement ad-
hesion to develop. As mentioned, the higher the ve-
hicle speed, the less time is available for that contact
to be made. Thus, available friction drops with speed.
The German road design standards (RAA, 2008) ac-
tually follows a similar approach, justified on avail-
able tyre-pavement friction values on dry and wet
surfaces. On the condition that the required sight dis-
tance above the safety barrier to the 1,05m object
height is available, appropriate vertical signage is
recommended, so that drivers reduce vehicle speed
only under wet pavement conditions, when the avail-
able friction is lower. The height of the central re-
serve barrier should be at most 0,90m and no plant-
ing or anti-glaring fences are recommended along
tight left turns, because they obstruct vision lines.
limit sign is recommended, permitting V85 speed at
dry pavement condition, for example 120km/h, and
restricting the permissible speed to 80 or even
70km/h at wet pavement conditions. A depiction of
this sign (Fig. 3) is presented at page 3/14 of the Ap-
pendix III at the Volume 6 (on motorway vertical
signage), (OMOE, 2002).
2.5 Evaluating signage alternatives
Considering that only via appropriate signage the
keen visibility deficit issue along motorway fast lane
tight left downgrade curves can be resolved, the
question arises, which is the appropriate sign to ef-
fectively and judicially warn drivers?
Posting the relevant K-12 sign (Fig. 4), in combi-
nation with a speed limit sign is out of the question
for motorways, because, as a rule, they do have anti-
skid pavements.
Figure 3 sign is efficient and almost impressive but
also implies that a defect exists at the motorway

Figure 3. A motorway sign prompting drivers to radically re-
duce their speed in wet pavement conditions
Figure 4. K-12 sign declaring slippery road
when raining. On the other hand, a simple P-32 speed
limit (Figure 5) is insufficient. Drivers generally ig-
nore speed limits considering them as arbitrary and
extra-conservative.
Figure 5. A P-32 speed limit sign and a Π-67 typical lane
speed limit sign
The Π-67 Greek Highway Code sign may be an
appropriate solution, if it declares lower speed for
the fast lane. However, is it proper to reversely diffe-
rentiate speed limits and impose lower speed to the
fast lane (i.e. 80km/h for the fast lane and 100km/h
for the slow lane)? This is contrary to Highway Code
principles. Many drivers may think that an abnormal
sign like this is an error, though some others may be
impressed and become suspicious and watchful.
In Figure 6 a different approach is outlined. The
K-25 sign is preferred, which warns for hazards other
than all the previous K-1 to K-24 Highway Code
signs highlight (among which is the Figure 4, K-12
sign). This is true because it is actually the visibility
problem and not the slippery pavement problem
which urges for reduced speed. On the same pole a
P-30 no overtaking sign should exist. Drivers along
mountainous motorways are familiar with P-30 sign
(actually with the similar P-31 sign), which is typi-
cally valid inside tunnels. Both these signs do not de-
clare any infrastructure fault, but the suspension of
the usual overtaking manoeuvre along that motor-
ways subsection due to temporary safety issues. Be-
low the pole with both K-25 and P-30 signs, it is
recommended to exist a tag declaring the length of
the motorway along which the overtaking prohibition
is valid. This length refers only to the central section
of the tight horizontal curve along which the visibility
is insufficient and is normally in the order of only a
few hundred meters and will make the warning more
acceptable. Generally, in motorways, the fast lanes
are used either when traffic volumes are high or dur-
ing an overtaking manoeuvre. In high traffic volumes
speeds are low and no visibility deficit to low height
objects arises. In the second case, if an accident oc-
curs during the overtaking manoeuvre, due to the ex-
istence of an obstacle on the fast lane, liability issues
are clearer, it is the driver who violated the traffic
sign to be blamed for, without any cumbersome and
dubious procedure, i.e. to find out the operating
speed of the vehicle.
Figure 6. K-25 and P-30 signs
3. DISCUSSION AND CONCLUSIONS Lamn R., B. Psarianos and T. Mailaender, Highway Design and
Traffic Engineering Handbook, McGraw-Hill, New York,
ISBN 10 0070382956, 1999.
Safety barriers along tight left horizontal downgrade OMOE, Ministry of Public Works , Guidelines for the Design
curves usually obscure visibility to stopping sight dis- of Roadworks, (Greek acronym: OMOE), 7 Volumes,
tance to the typical low object, although the appro- Greece, 2002.
priate sight distance to the stop lights of leading ve- RAA, German Road and Transportation Research Association,
hicles is maintained above the barrier. Considerable Richtlinien für die Anlage von Autobahn, R1 RAA, ISBN
shift of the barrier to the inside of the curve is a dras- 978-3-939715-51-1, 2008.
tic, but a non-preferred action involving serious im-
plications. The adoption of relaxations and appropri-
ate speed limit signage, which drastically restricts
speeds in wet pavement conditions, are controversial
practices.
From Table 1 it appears that for V85, for the
maximum permitted downgrades, the available vision
length to a low obstacle is less than half the required
for safe stopping. Friction need to take unrealistically
high initial values (around 1) to achieve stopping
ahead of a low obstacle. Although sufficient friction
supply should be regarded as the most important
safety practice when considering motorway safety,
such high friction values is practically impossible to
be achieved.
The German Guidelines (RAA, 2008) proposes
(Fig. 3) a markedly reduced speed in wet pavement
conditions. In this paper the sign combination of Fig-
ure 6 is considered as the right proposal to overcome
motorway fast lanes visibility problems: the K-25
sign of the Highway Code warns for other than the
known hazards, together with a P-30 no overtaking
sign. Actually, to exist a hazardous obstacle along
motorways fast lanes is a rare and specific event,
which is covered by K-25 sign. As far as P-30 sign is
concerned, on the one hand it directly averts moving
along the fast lane and, on the other hand, drivers
moving along mountainous motorways are familiar
with it. This combination of signs does not imply any
pavement defect but only suspends the overtaking
manoeuvre for a few hundred meters along tight
downgrade left curves.

5. REFERENCES

Kokkalis A., Anti-skid pavements to secure safe stopping before

obstacles on motorway fast lanes along tight horizontal
curves, 5th Int. Conf. Bituminus mixtures and Pavements,
Thessaloniki, Greece, 1-3 June 2011
Kokkalis A., G. Mintsis, C. Taxiltaris, S. Basbas and J.
Dimitropoulos, Conclusions drawn by operational velocities
measurements and speed limits enforcement in motorways,
Proceedings of the 3rd Pan-Hellenic Highways Conference,
to be held in Patra, Greece, November 2011.
Kokkalis A., Prediction of skid resistance from texture meas-
urements, Proc. Institution of Civil Engineers, journal of
Transport, No. 129, May 1998, pp. 85-93.
Kokkalis A. and O. Panagouli, Fractal evaluation of pavement
skid resistance variations I & II: surface wear and surface
wetting’, Journal of Chaos, Solitons & Fractals, Elsevier
Science Ltd., Vol.9, No. 11, 1998, pp.1875-1899.
Brighter and better all-weather road markings in Malaysia
Mohd Hizam HARUN
Nur Syaza MOHAMAD NOH
Public Works Department Malaysia

Wan Radhiah WAN HANAFI

Roadcare (M) Sdn. Bhd.

ABSTRACT: Road markings are essential in providing delineation to the road users. They should be visible
especially under poor lighting. The most critical time is at night when it rains. Malaysia have started using all-
weather thermoplastic as road marking materials since 2010 which incorporate special reflective elements
which are capable in giving relatively high retro-reflection even though when they are submerged in thin film
of water. However, it was observed that the brightness of this more costly road marking material at some lo-
cations does not last long. So, what could have gone wrong? This paper highlights some good and bad prac-
tices in laying the road markings which result in satisfactory and poor performances respectively. Correlations
between the brightness and durability of the road markings with the various procedures of heating and laying
the thermoplastic powder, different qualities of the powder, inter-mix and drop-on reflective elements, differ-
ent techniques in incorporating drop-on reflective elements, variation in the heating and laying temperatures,
and the conditions of the road surface during laying, are discussed. These findings will be used to improve the
standard specification for brighter and longer lasting road markings.

1 INTRODUCTION than the ordinary thermoplastic road lines. So what

could have gone wrong?
Thermoplastic road markings will appear bright and
glowing after installation. The ability to reflect light
from vehicles’ headlamps at night could help motorists
to see the road markings clearly, and thus feel safe
and comfortable. However, when it rains, conven-
tional glass beads could not reflect light effectively
when submerged in thin film of water on the road
surface, resulting in the road markings appear dull
and difficult to see. Consequently, all-weather ther-
moplastic (AWT) road markings have been intro-
duced and included in the new specification of Pub-
lic Works Department (PWD) Standard
Specification for Road Works JKR/SPJ/2012 Sec-
Figure 1. All-weather thermoplastic road markings.
tion 6: Road Furniture, Sub-Section 6.3: Road Mark-
ings (2012), hereinafter referred to as SPJ. Malaysia
is the first country in Southeast Asia that uses the
AWT road markings, its inception in October 2010. 2 AWT ROAD MARKINGS
Befitting the tag 'all-weather', the road markings
could be clearly visible in all weather conditions, es- AWT road markings comprise of high performance
pecially at night and when it rains. Fish (1996) de- thermoplastic powder, inter-mix reflective beads and
fines all-weather road markings as markings that are drop-on beads. The performance of the AWT road
visible at night under dry conditions and also under markings depends on three parameters, namely;
rainy conditions up to 0.25 inches per hour of rain- 1. Materials.
fall. 2. Applicator machines.
However, there are complaints that the AWT road 3. Workers.
markings are not able to last long. This would cause
discontent among the road authorities as the price of
AWT road markings is three times more expensive
2.1 Materials ramic beads (refractive index > 2.4) with the remain-
ing composition consists of 70% high performance
2.1.1 Material specifications glass beads (refractive index > 1.9).
As specified in clause 6.3.2 of the SPJ, thermoplastic
materials (ie. the white or yellow powder) shall
comply with BS EN 1871 (2000). High performance
thermoplastic powder which complies with BS EN
1871 and certified by recognised agencies should
always be used; materials of inferior quality will not
hold the drop-on beads firmly even though the beads
are properly embedded during application. If the
white thermoplastic powder change its colour to
grayish or yellowish if overheated, that is a clear in-
dication of inferior quality thermoplastic material
that should be immediately rejected from sites. Sub-
stantial different in cost (~ 35%) between the supe-
Figure 2. Glass beads.
rior (RM2,700/ton) and inferior (RM2,000/ton)
product may encourage the contractor to use the
cheaper product if enforcement is lacking. Materials
without certification should not be accepted.

2.1.2 Reflective beads

The thermoplastic material should contain inter-mix
reflective beads, and these beads should comply
with BS EN 1424 (1998). In addition, during appli-
cation on to the road surface, drop-on reflective
beads are dropped (usually by gravity) shortly after
the heated thermoplastic material is spread, and
these beads should comply with BS EN 1423 (1998). Figure 3. Micro-crystalline ceramic beads.
Both BS EN 1424 and 1423 refer to glass beads,
with a minimum value of refractive index (ability to
refract light in water) of 1.5.
Other types of reflective material such as micro-
crystalline ceramic beads have higher refractive in-
dex of 2.4. The ability of ceramic beads to reflect
light towards the source of the light (ie. retro-
reflection. In the context of road lines, it is the abili-
ty of the beads to reflect light towards vehicles’
driver) is said to be similar to diamonds. For normal
glass beads, the refractive index is relative low
(usually about 1.5), but there are high performance
glass beads with refractive index of 1.9. Beads of in-
ferior quality will not be able to reflect light towards Figure 4. Comparison of size between glass beads and mi-
the vehicles’ drivers, or reflect the light towards cro-crystalline ceramic beads.
many other directions (diffuse reflection). During
the day and when it does not rain, normal glass
beads are sufficient, but at night or when it gets
dark, especially when it rains, the superiority of ce-
ramic beads will be more pronounced. Ceramic
beads are said to be seven times brighter than glass
beads.
The quantity of inter-mix beads is not specified in
the SPJ. It is recommended that it should be not less
than 10% by weight of the thermoplastic powder
with at least 3% ceramic beads. For drop-on beads,
the quantity is specified of not less than 400 g/m2.
However, the proportion of each type of beads is not
specified. It is recommended that the drop-on beads Figure 5. Inter-mix micro-crystalline ceramic beads and glass
should contain at least 30% of micro-crystalline ce- beads inside road markings.
 mcd/m2/lux. Apart from this, Fish (1996) stated that
most research has come to a conclusion that a mini-
mum acceptable level of retro-reflectivity is around
100 to 120 mcd/m2/lux.
Asdrubali et al (2013) observed that while the on-
ly performance parameter considered was RL, at ur-
ban level the reflection in daylight condition Qd is
equally (or more) important.

Figure 6. Drop-on micro-crystalline ceramic beads and glass

beads on road markings.

2.1.3 Performance specifications

SPJ clause 6.3.2 stipulates that the performance of
the thermoplastic road line material shall comply
with BS EN 1436 (2007). Among other things, the
BS EN 1436 specifies;
1. Luminance factor β in dry condition not less
than 0.60. Figure 7. Retro-reflectometer.
2. Luminance coefficient under diffuse illumi-
nation Qd not less than 160 mcd/m2/lux.
(Note: If the luminance factor β can be
achieved, Qd does not need to be evaluated,
and vice versa).
3. Coefficient of retro-reflected luminance RL
in dry condition (ASTM-E1710) not less than
300 mcd/m2/lux.
4. Coefficient of retro-reflected luminance RL
in wet condition (ASTM, E2177-11) not less
than 75 mcd/m2/lux.
(Note: Qd measures the brightness of the
road markings in day light illumination and
street light, while RL measures retro- Figure 8. Coefficient of retro-reflected luminance.
reflection when lighted by vehicles’ head-
lamp.
The SPJ specifies that the RL shall be not less 2.2 Applicator Machines
than 300 mcd/m2/lux (dry condition) and not less
than 75 mcd/m2/lux (wet condition) within 7 days 2.2.1 Drop facility for beads
after the road markings are laid. However, after ac- If a mixture of glass beads and micro-crystalline ce-
quiring more knowledge on good and bad practices ramic beads is used as a drop-on reflective element,
in laying road lines, the Superintendant Officer is the applicator machine should be equipped with a
encouraged to request the contractor to carry out the double-drop facility with two units of dispenser. The
testing on the 6th or 7th day to allow any loose drop- larger ceramic beads (typically 1.1 - 1.3 mm) should
on beads, if there is any, to detach prior to testing. be dropped first, followed by the smaller glass beads
Consequently, the requirement on the RL will be re- (usually 0.7 - 0.8 mm). If the smaller glass beads fall
viewed and changed accordingly. It has been sug- first, it would impede the larger ceramic beads from
gested that the requirement for the test to be carried filling up the space in between the smaller beads,
out ‘within 7 days’ be changed to ‘within 10 – 20 and this can lead to the larger beads from being easi-
days’ while the minimum value of RL remain un- ly detached from the road markings by the passing
changed. traffic. Similarly, the larger ceramic beads would be
In the opinion of the European Union Road Fed- loosely embedded into the road markings if both
eration, ERF (2013), a good road markings are those types of bead drop simultaneously from a single-
with performance level under dry conditions not less drop facility with one unit of dispenser.
than 150 mcd/m2/lux, whereas for wet conditions,
the minimum performance level should be 35

Figure 9. Applicator machine with double-drop facility.
Figure 10. Applicator machine with single-drop facility.
2.2.2 Height of spreader shoe
The height of the spreader shoe above the road sur-
face should not be too high. It is recommended that
the height should be 40 – 50 mm. If it is higher,
there is a possibility that up to 10% of the drop-on
beads will be blown away before they reach the road
markings by the wind or passing traffics. It should
not be too close because the drop-on beads will not
satisfactorily embed into the road markings. The tar-
get is to get the beads embedded into the road mark-
ings layer by about 50 – 60%.
A study elsewhere has shown that high retro-
reflectivity could be achieved if the degree of sink-
ing of glass beads is between 55 and 60 % of their
diameter. A sinking degree less than 50% weakens
their resistance to tyres, while a sinking degree
higher than 60% limits the retro-reflection properties
(ERF, 2013).
Figure 11: Embedment of drop-on beads.
2.2.3 Speed of application
The applicator machine should not be pushed too
fast as some of the relatively light glass beads or ce-
ramic beads will be blown by the wind.
2.3 Workers
2.3.1 Laying thickness
The optimum thickness of the overlaid thermoplastic
powder is typically 2.0 - 2.5 mm. If thinner, the
drop-on beads would not stick firmly to the road
markings. If thicker, most of the drop-on beads
would sink into the road markings.
Figure 12. The optimum thickness is 2.0 – 2.5 mm.
2.3.2 Road surface condition
Prior to the application of the road markings, the
surface of the road should be clean and dry. Surface
which is dirty, sandy and wet would cause the road
markings to easily peel off.
After rain, it is advisable to wait for at least 12
hours for the moisture on the road surface to dry.
The formation of small holes (pinholes) on the road
markings indicates that the road surface has not ade-
quately dried when the road markings are installed.
It is also recommended that road markings are in-
stalled not less than 72 hours after paving works to
let the bitumen sufficiently harden.
2.3.3 Overheating
If the thermoplastic material is inadvertently over-
heated above the recommended mixing temperature,
the viscosity during application, and thus the em-
bedment of the drop-on beads, will be affected even
though the material is allowed to cool down.
2.3.4 Mixing and laying temperatures
The mixing temperature of the thermoplastic materi-
al should be as recommended by the supplier. Usual-
ly the temperature is 200 - 220 °C. However, if the
road markings are installed in the evening when the
temperature is a bit cooler, or in cooler highlands,
the mixing temperature should be increased up to
230 - 240 °C. Laying temperature (the temperature
of the thermoplastic material as it flows out from the
spreader shoe) should be similar to the mixing tem-
perature. If applied at cooler temperatures, the drop-
on beads would not stick firmly while if the tem-
perature is higher, the drop-on beads would sink.
While the RL could be achieved when the test is car-
ried out within seven days, the loosely attached
drop-on beads would eventually detach from the
road markings by passing traffics. In contrast, the RL
could not be achieved when a lot of drop-on beads
sink into the road markings if the laying temperature
is too hot. The applicator team should be equipped
with hand-held contactless thermometers to rapidly
check the temperatures.
Figure 13. The mixing temperature should be as recommended
by the material’s supplier.
3 FUTURE WORKS
The rate of degradation of the road markings varies
with traffic intensity. However, this is still unknown.
Consequently, the SPJ presently only specify the ini-
tial minimum coefficient of retro-reflected lumin-
ance RL of 300 and 75 mcd/m2/lux, in dry and wet
conditions respectively, which are to be achieved
within seven days after laying the road markings.
The performance of the AWT road markings, which
will be installed transversely to expedite the degra-
dation, will be periodically monitored before the re-
quirements of RL after, say, six and/or 12 months are
incorporated into the SPJ.
4 CONCLUSION
For countries like Malaysia, being located in the re-
gion with tropical climate and experiencing frequent
heavy rainfall, it is envisaged that it is imperative to
have all-weather thermoplastic road markings for the
safety and comfort of the road users. Some good
practices which are based on experience have been
highlighted which need to be adhered to in order to
consistently produce high quality road markings.
Nonetheless, further studies need to be carried out to
ensure that the road markings are not only visible at
all times, but also durable.
5 REFERENCE
ASDRUBALI Francesco, Cinzia Buratti, Elisa Moretti, Fran-
cesco D’ Alessandro and Samuele Schiavoni. Assessment
of the performance of road markings in urban areas: The
outcomes of the CIVITAS RENAISSANCE Project. CI-
RIAF, Interuniversity Centre of Research on Pollution by
Physical Agent, University of Perugia, Italy. The Open
Transportation Journal, 2013.
ASTM INTERNATIONAL. Standard test method for mea-
surement of retro-reflective pavement marking materials
with CEN-prescribed geometry using a portable retro-
reflectometer. E1710 – 11.
ASTM INTERNATIONAL. Standard test method for measur-
ing the coefficient of retro-reflected luminance of pavement
markings in a standard condition of wetness. E2177 – 11.
BRITISH STANDARDS INSTITUTION. Road marking mate-
rials – Physical properties. BS EN 1871: 2000.
BRITISH STANDARDS INSTITUTION. Road marking mate-
rials – Premix glass beads. BS EN 1424: 1998.
BRITISH STANDARDS INSTITUTION. Road marking mate-
rials – Drop on materials - glass beads, antiskid aggregates
and mixtures of the two. BS EN 1423: 1998.
BRITISH STANDARDS INSTITUTION. Road marking mate-
rials – Road marking performance for road users. BS EN
1436: 2007 +A1: 2008.
EUROPEAN UNION ROAD FEDERATION. Marking the
way towards a safer future. An ERF position paper on how
road markings can make our road safer.
FISH, Joseph K. Evaluation of all-weather pavement markings:
Report on two years of progress. Semisequicentennial
Transportation Conference Proceedings, Iowa State Univer-
sity, Ames, Iowa. May 1996.
PUBLIC WORKS DEPARTMENT, MALAYSIA. Standard
specification for road works, Section 6: Road furniture –
Sub-section 6.3: Road markings. JKR/SPJ/2012-S6.
Safety audit using operating speeds V85 at rural road in Northern Greece
G. Mintsis, S. Basbas, C. Taxiltaris, N. Domoktsis & K. Labropoulou
Faculty of Rural & Surveying Engineering, Aristotle University of Thessaloniki

ABSTRACT: Before the introduction of the new Guidelines for the Design of the Road Network (National
Road Design Standards) in the early 2010, road design in Greece was lacking of a clear methodology to secure
continuity in driving performance. This paper describes the research carried out in order to study the safety
level provided by older designs. For that, the Safety Criteria introduced by the new Guidelines were tested in
road sections of the rural network of Northern Greece.

1 INTRODUCTION by the new Standards (OMOE) and whether the im-

For more than forty (40) years road design in Greece
was guided by principles and rules set in the begin-
ning of 1960 decade adopted in the National Road
Design Standards. One of the most important prin-
ciples of that Manual was the distinction between
tangents as curves as far as the operational role of an
alignment element is concern. The dynamic element
of vehicle-driver system was taken into account only
when curves were to be examined. Tangents were
considered simply as the connecting links between
curves. This led to alignments that in many cases
forced drivers to develop driving profiles characte-
rized from sudden changes in speed therefore highly
demanding of continuing attention and concentration.
The new Greek Design Standards (OMOE), that
were recently introduced, introduced the principle of
harmony and continuity in driving profile by recog-
nizing tangents as being alignment elements that af-
fect and in a way when combined harmonically with
curves do not create a highly demanding environment
in terms of ability and continuously paid attention.
The principle of harmonic design is implemented
of a set of rules that concern geometry and operation
most important of which are the safety criteria I, II
and III.
However the majority of road link in Greece have
been designed and built before the introduction of the
new Design Standards (OMOE) and obviously do
not offer the driving environment envisaged by
OMOE.
This paper tries to answer a question raised many
times as to what extent order designs and construc-
tions fulfill the operation and safety requirements set
plementation of the Safety Criteria exercise could as-
sist authorities in defining road alignment elements
that need to change in geometry and so combined to-
gether with other existing road elements to create an
infrastructure that secures safe and comfortable driv-
ing.
Three (3) road sections of different category of
the rural primary and secondary road network of the
Prefecture of Serres in Northern Greece were used to
be examined under the above presented concept.
2 STUDY AREA
Prefecture of Serres was defined as the study area
and specific road sections of the primary and the sec-
ondary rural road network were selected for the col-
lection of the data required. For the final selection of
the road sections surveyed the existing data base of
EGNATIA ODOS S.A. was used that was created
from data collected from different Road Safety Im-
provement Studies developed by this Organisation.
In all 18 road sections of the rural road network were
examined with lengths ranging from 2,3km to
25,3km. Finally, three of those roads sections were
selected and studied in terms of geometry and opera-
tion. The selection methodology is described below.
3 METHODOLOGY – SELECTION OF ROAD
SECTIONS
The following criteria were defined and used for the
selection of the three (3) road sections of the rural
network of the Prefecture of Serres.
1. The road sections had to belong to the pre-
vailing road categories of the Prefectures road net-
work auditing to the current National Road Design
Standards (ΟΜΟΕ-ΛΚΟΔ).
2. To have reasonable length.
3. To show small width variations along their en-
tire length.
For the application of the first criterion (a.) the 18
initially considered road sections were classified in
road categories according to the road classification
system introduced by the National Road Design
Standards. All road sections belong to Road Group
A as being sections of the rural network of the Pre-
fecture of Serres. In addition to that and by the use
of the EGNATIA ODOS S.A. data base the road
sections were also classified ranked according to
their operational character and an operational ranking
was attributed to each sections. The combination of
the above described grouping and ranking resulted to
the categorization of the sections to road categories
from AII to AIV.
The final three (3) road sections used in the study
were selected when the other two selection criteria
(band c) were also considered. Table 1 shows the
specific road sections and their basic characteristics
where their exacted location in the entire road net-
work is presented in Figure 1.
Speed data collection and safely analysis were
performed at specific road sub-sections that were se-
lected out of the three study road sections selected
with the use of the methodology described above.
For that a detailed analysis of the geometry of the
three road sections was performed that led to the
identification of a series of road sub sections selected
for study which also were grouped in the following
four groups.
Group 1: Road sections with low gradient (≤5%)
and low bendiness (comfortable values of the hori-
zontal curvature).
Group 2: Road sections with low gradient (≤5%)
and limited bendiness (limited horizontal curvature).
Table 1. Road sections considered in the study.
Road code Road section
EO*-O2 Boundaries of Prefecture of Thessaloniki – Strimonas
River Old Bridge
EΠΟ**-10 Kala Dendra – Heraklia
ΕΟ-59 Amfipolis – Mesoraxi Junction
*EO: National Road
**ΕΠΟ: Regional Road
Group 3: Road sections with high gradient (>5%)
and low bendiness (comfortable values of the hori-
zontal curvature).
Group 4: Road sections with high gradient (>5%)
and limited bendiness (limited horizontal curvature).
All road sections considered are aligned at a plain
terrain. The classification criteria to the above four
groups were quantified according to the provisions
made by the Design Standards for the calculation of
the operating speed (V85) for signal carriageways and
the definition of the limited (allowed) values of the
radius of the horizontal curvature in a flat terrain for
design speed of 80 km / h and 90 km / h.
In particular the grouping of the sub-sections ac-
cording to bending criterion to low and limited ben-
diness was performed by the use of the limited value
of curve radius increased by two (2). Thus sub-
sections with road curve radius smaller than 500 m
for design speed (Ve) 80 km / h and 660 m for design
speed (Ve) 90 km / h were grouped in groups 2 and
4. All other sub-sections were placed in groups 1 and
3.
Additional criteria for the selection of the road
subsections were defined as:
• existence of reasonable length to allow valuable
results,
• existence of both tangents and curves in the ho-
rizontal alignment,
• longitudinal profile with slopes of about the
same magnitude.
Exception to the above third rule constitute the
selected sub-sections of the road section “Amfipolis -
Mesoraxi” where in one of the two finally selected
sub-sections both steepness criteria (>5% and ≤5%)
have been applied.
The application of the above described methodol-
ogy led to the final selection of the road sub-sections
where the speed data were collected. Altogether
twelve (12) sub-sections where identified and studied
along the three (3) road sections. The alignment cha-
racteristics of the road sub-sections are given in the
Table 2.
Length Lane Road Road Design speed
width grou category
p
m m Km / h
10.210 4,00 A AIV 80
20.684 3,40 A AIII 80
24.084 4,25 A AII 90
Table 2. Selected road sub-sections for speed measurements.
Road section Lengt Gradient Alignment Alignment elements
h
m % m
Low ben- Limited Indepen- Depen- Horizon-
diness bendiness dent tan- dent tan- tal curves
gents gents
Boundaries of Prefecture of Thessa- 10.210 ≤ 5% 2.205 - 2 - 1
loniki – Strimonas River Old ≤ 5% 2.648 - 3 1 3
Bridge ≤ 5% - 1.045 2 - 1
Kala Dendra – Heraklia 20.684 ≤ 5% - 1.430 1 1 2
≤ 5% - 963 1 3 4
≤ 5% 846 - 2 1 3
≤ 5% 863 - 2 1 4
Amfipolis – Mesoraxi Junction 24.084 ≤ 5% 968 - 1 1 3
≤ 5% & > 5% 2.046 - 3 - 3
≤ 5% & >5 % 1.423 - 3 - 5
≤ 5% 2.695 - 2 - 1
≤ 5% 1.792 - 2 - 2
TOTAL 54.978 15.486 3.438 24 8 32

Figure 1. Selected road sub-sections for speed measurements (Source: ©Contribution by OpenStreetMap).

4 DATA COLLECTION Vehicle speed data were collected at carefully

Site visits and inspections of the road sub-sections
revealed that there were inconsistencies among the
horizontal alignments provided by EGNATIA ODOS
S.A. and the physical layout of the particular sub-
sections. Therefore, site alignment surveys were con-
ducted with the use of GPS device. The so collected
survey data were input in the road design software
ANADELTA and the required alignment element
values were calculated. Superelevation measurements
were made on site with the use of appropriate survey
instruments.
predecided location in each of the study road sub-
sections with the use of radar speed meter. Addition-
al care was given in order not to influence drivers’
behaviour if they had seen the observer. Therefore
the observer remained sitted in the vehicle that was
carefully placed by the side of the road. A sampling
limit of 100 vehicles per data collection period was
imposed in order for the analysis to produce scientifi-
cally sound results.
5 DATA ANALYSIS – RESULTS
The aim of the analysis was tri-fold.
1. To produce values for different speed statis-
tical parameters that allow the recognition of the
speeding bahaviour of drivers at the particular road
sub-sections and their variations along sub-sections.
Emphasis was given in the calculation of the operat-
ing speed (V85) of the different elements (30 tangents
and 32 curves) of the twelve (12) sub-sections.
2. To apply the three (3) distinctive safety crite-
ria adopted in the Road Design Standards (OMOE-
X).
Figure 2. Distribution of mean speed values along the Amfipolis – Masoraxi Junction road section.
3. To examine the existing speed limits on the
basis of the adopted speed behaviour from the road
users.
In all speeds were measured at 64 road elements
(tangents and curves) as it is shown in Table 2. Mean
speed, standard deviation and operating speed (V85)
were calculated for each of the 64 different elements
of the road alignment.
Typical examples for the distribution of mean
speed and standard deviation values are given in Fig-
ures 2 and 3 accordingly for the road section Amfi-
polis – Mesoraxi Junction.
Operating speed (V85) values also calculated as
shown in Table 3 for the different alignment elements
of the same road section.
Figure 3. Distribution of standard deviation values along the Amfipolis – Masoraxi Junction road section.

Table 3. Operating speed for road alignment elements along Amfipolis – Mesoraxi Junction road section.
Amfipolis – Mesoraxi Junction
Road alignment elements
K8-K13 K25-K31 K32-37 K57-K59 K80-K85
Elements* V85 Elements* V85 Elements* V85 Elements* V85 Elements* V85
km / h km / h km / h km / h km / h
K8-K10 101 K25 96 K32 95 K57-K58 119 K80 77
K10 89 K25-K27 95 K32-K33 100 K58 113 K80-K83 105
K10-K11 87 K27 112 K33 104 K58-K59 112 K83 78
K11 80 K27-K30 103 K33-K35 99 K83-K85 87
K11-K13 69 K30 104 K35 97
K30-K31 106 K35-K36 86
K36 94
K37 81
* Kx-Ky: tangent, Kz: curve

Table 4. Safety criteria values / acceptance level for the road section Boundaries of the Prefecture of Thessalo-
niki – Strimonas River Old Bridge.
Subsection Alignment V85 Criterion I Acceptance level Criterion II Acceptance level Criterion III Acceptance level
element
K6-K10 K6-K7 106 26 not accepted
K7 96 16 moderate 10 good 0,01 good
K7-K10 108 28 not accepted 12 moderate
K16-K22 K16-K17 94 14 moderate
K17 85 5 good 9 good -0,15 not accepted
K17-K19 81 1 good 4 good
K19 87 7 good 6 good 0,03 moderate
K19-K21 99 19 moderate 12 moderate
K21 97 17 moderate 2 good 0,01 good
K21-K22 106 26 not accepted 9 good
K32-K35 K32-K34 84 4 good
K34 60 20 moderate 24 not accepted 0,12 not accepted
K34-K35 73 7 good 13 moderate

Table 5. Safety criteria values / acceptance level for the road section Kala Dendra - Heraklia.
Subsection Alignment V85 Criterion I Acceptance level Criterion II Acceptance level Criterion III Acceptance level
element
K30-K35 K30 75 5 good 0,00 good
K30-K34 79 1 good 4 good
K34 69 11 moderate 10 good -0,12 not accepted
K34-K35 87 7 good 18 moderate
K55-K59 K55 69 11 moderate -0,20 not accepted
K55-K56 78 2 good 9 good
 K56 74 6 good 4 good -0,16 not accepted
K56-K57 80 0 good 6 good
K57 73 7 good 7 good -0,29 not accepted
K57-K58 78 2 good 5 good
K58 82 2 good 4 good -0,12 not accepted
K58-K59 82 2 good 0 good
K75-K78 K75 90 10 good -0,10 not accepted
K75-K76 87 7 good 3 good
K76 82 2 good 5 good -0,13 not accepted
K76-K77 80 0 good 2 good
K77 58 22 not accepted 22 not accepted -0,24 not accepted
K77-K78 71 9 good 13 moderate
K96-K100 K96 75 5 good -0,18 not accepted
K96-K97 84 4 good 9 good
K97 73 7 good 11 moderate -0,21 not accepted
K97-K98 69 11 moderate 4 good
K98 64 16 moderate 5 good -0,02 moderate
K98-K100 79 1 good 15 moderate
K100 77 3 good 2 good -0,09 not accepted

Table 6. Safety criteria values / acceptance level for the road section Amfipolis – Mesoraxi Junction.
Subsection Alignment V85 Criterion I Acceptance level Criterion II Acceptance level Criterion III Acceptance level
element
K8-K13 K8-K10 101 11 moderate
K10 89 1 good 12 moderate -0,06 not accepted
K10-K11 87 3 good 2 good
K11 80 10 good 7 good -0,02 moderate
K13 69 21 not accepted 11 moderate -0,01 moderate
K25-K31 K25 96 6 good -0,10 not accepted
K25-K27 95 5 good 1 good
K27 112 22 not accepted 17 moderate -0,11 not accepted
K27-K30 103 13 moderate 9 good
K30 104 14 moderate 1 good -0,09 not accepted
K30-K31 106 16 moderate 2 good
K32-K37 K32 95 5 good -0,13 not accepted
K32-K33 100 10 good 5 good
K33 104 14 moderate 4 good -0,10 not accepted
K33-K35 99 9 good 5 good
K35 97 7 good 2 good -0,12 not accepted
K35-K36 86 4 good 11 moderate
K36 94 4 good 8 good -0,09 not accepted
K37 81 9 good 13 moderate -0,02 moderate
K57-K59 K57-K58 119 29 not accepted
K58 113 23 not accepted 6 good -0,30 not accepted
K58-K59 112 22 not accepted 1 good
K80-K85 K80 77 13 moderate 0,03 good
K80-K83 105 15 moderate 28 not accepted
K83 78 12 moderate 27 not accepted -0,23 not accepted
K83-K85 87 3 good 9 good

The level of service in terms of safety and opera-

tion provided by the road sections studied were ex-
amined by the use of the safety criteria adopted in the
National Road Design Standards. The scope of this
exercise was not only to access the level of safety
and operation level provided by the road sections,
but mainly to examine the level of compliance of old
road designs to the safety control procedures intro-
duced in the Design Standards.
In particular, three (3) safety criteria examine the
deviation between design (design Speed – Ve) and
operation (operational speed – V85) with the use of
Safety Criterion I (Ve - V85), the existence of har-
monic driving profile along the road with the use of
Safety Criterion ΙΙ (V85 i - V85 i+1) and the safe and
comfortable driving around road curves with the use
of Safety Criterion III (fexist. – fdemand.). A common
ranking is used in all three safety criteria expressed as
“good”, “moderate” and “not acceptable” case.
The above mentioned procedure was applied in all
twelve (12) sub-sections using data defined (Ve), cal-
culated (V85), measured (superelevation) and ac-
cepted (f – Design Standards) and the values of the
three safety criteria were calculated for all the differ-
ent alignment elements of the three road sections.
The results of this exercise for the three road sec-
tions are presented in Tables 4, 5 and 6 for the re-
spective road sections.
Figures 4, 5 and 6 present the results for the three
distinctive safety criteria for all 64 road alignment
elements where considered together.
Figure 4. Acceptance level of the Safety Criterion I for all road
elements.
Figure 5. Acceptance level of the Safety Criterion II for all
road elements.
Figure 6. Acceptance level of the Safety Criterion III for all
road elements.
In the final stage of the analysis the speed limits
prevailing for the particular road sections according
to the Greek Road Traffic Code were compared with
statistical values of vehicle speed such as V75 and V85
that consider to express driving attitude to safe driv-
ing along the specific road sections.
The cumulative distribution of vehicle speed val-
ues was produced for selected alignment elements of
the road sections and compared to the prevailing
speed limits.
In all but three cases 75th percentile of speed (V75)
exceeded speed limit by 6% to 25%. All the same the
85th percentile of speed (V85) exceeded speed limit in
all two cases from 16% to 35%. These findings raise
a question to whether the imposed speed limits are,
in most cases, too conservative and also if a reconsi-
deration of the limits should be put under discussion.
6 CONCLUSIONS
The question set to be answered by this research
work was whether old road designs in Greece offer
to drivers safety and comfort and therefore harmony
in driving profiles in the way that new National De-
sign Standards (OMOE) consider harmonic driving
along roads. To answer this question driving beha-
vior as expressed by speed profiles was measured
and studied at three road sections of the rural prima-
ry and secondary road network in the Prefecture of
Serres in Northern Greece.
In all 64 alignment elements were examined sepa-
rately and combined. They belong to three (3) dis-
tinctive road sections categorized as roads of AII,
AIII and AIV category respectively. With the use of
speed and geometry data measured and calculated
the three (3) safety criteria I, II and III introduced by
the new Standards (OMOE) where applied at twelve
(12) road sub-sections to examine whether there
were consistency in speed variation between design
and operation, along the different alignment elements
and also where curves could have been driven safely
and comfortably with given speeds.
The analysis showed that there was a rather good
relation between alignment and safety checks when
criteria I (Ve - V85) and ΙΙ (V85 i - V85 i+1) were ap-
plied. Out of all cases 59% of the elements satisfied
the Criterion I requirement of “good” and 27% as
“moderate”. All the same the respective percentile
values for Criterion II were 70% and 22%. However
the results were quite contrary when Safety Criterion
III was examined and 72% elements (curves) were
found to be “not accepted” and only 12% to be ac-
cepted (“good”).
Once more it was found that curves and their high
value of curvature is the weak point of the old road
design in Greece and need to be look after. On the
other hand it is interesting to note that as far as com-
pliance between design and operation is concerned
the findings are much better and even more when
continuity and harmony in driving profile is ex-
amined.
Finally the results showed that there is degree of
conservatism in the definition and establishment of
the legal speed limits. Discrepancies even of 35%
where observed between speed limits and operating
speed V85.

7 REFERENCES

Ministry of Infrastructure, Transport and Network. 2005.

Guidelines for the Design of the Road Network (National
Road Design Standards) – OMOE, Athens.
Domoktsis, N. & Labropoulou K. 2014. The use of operational
speed in Road Safety. Diploma Thesis, Faculty of Rural &
Surveying Engineering, Aristotle University of Thessalo-
niki, Thessaloniki.
Touliatos, P., Evaggelidis, D., Lyberopoulos, G., Vaniotou, M.
& Anagnostopoulou, A. 2009. Speed counts analysis and
utilization of results in the determination of speed limits,
Proc. of the 4th Pan-Hellenic Greek Road Safety Confe-
rence, 5-6 November 2009, Athens.
Kokkalis A., Mintsis G., Taxiltaris, C., Basbas, S. & Dimitro-
poulos I. 2012. Conclusions from operational speed counts
for the imposition of speed limits in motorway segments.
Proc. of the 3rd Pan-Hellenic Highway Engineering Con-
ference, 9-10 February 2012, Athens.
The contribution of roundabouts to road safety: the case of 4 roundabouts
in the municipality of Thermi Greece
D. Spanou, F. Kehagia & M. Pitsiava-Latinopoulou
Department of Civil Engineering, Aristotle University of Thessaloniki, 54642 Thessaloniki, Greece

ABSTRACT: Roundabouts are circular intersections with specific design and traffic control features, offering
improved road safety, operational efficiency and level of service. Compared to other types of intersections,
roundabout installation is strongly promoted as an effective safety intersection treatment, mainly because the
potential conflict points between vehicles, pedestrians and other prone users are limited and speeds into and
through the intersections are decreased. According to statistical studies relating to fatal accidents and injuries,
roundabouts are proved to be safer in low or moderate traffic capacity conditions. The main objective of this
paper is to examine the safety aspects of roundabouts, focusing on the effectiveness of roundabouts in reduc-
ing intersection crash frequency and severity, discussing crash and accident statistics based on international
data resources and presenting the case of 4 roundabouts in the municipality of Thermi, Greece.

1 INTRODUCTION wards this purpose, the results of a case study con-

Roundabouts are designed in such a way so as to ef-
fectively control the traffic flow at intersections
without the use of stop signs or traffic signals. In re-
cent years researchers have shown particular interest
on the benefits of roundabouts, which constitute ba-
sic elements in both urban and rural traffic networks.
Improvements in capacity and road safety are in-
cluded in these benefits and make roundabouts an
effective measure for improving the overall traffic
conditions. The main attributes contributing to these
benefits are the reduction of conflict points and the
prevailing lower speed levels in comparison to other
types of at-grade conventional intersections. How-
ever the improved safety of roundabouts is related to
their geometric design, which determines their effi-
ciency and operation level.
This paper examines the main features and advan-
tages that roundabouts offer, as an alternative to oth-
er types of conventional intersections. At the same
time, the basic geometric design features are de-
scribed, with special reference to their contribution
to the level of road safety that roundabouts offer to
all their users. Towards this direction the paper
presents and analyses accident statistics in rounda-
bouts from various countries relating them with the
type of road users and highlights their benefits on
road safety providing the findings from before and
after case studies, concerning the reduction in the
number of accidents after the conversion of tradi-
tional intersections into roundabouts. Moreover to-
cerning the benefits derived from the conversion of a
number of junctions of other type to roundabouts in
the municipality of Thermi of the Prefecture of
Thessaloniki are given.
2 BASIC ELEMENTS OF ROUNDABOUTS
2.1 Definition
Roundabouts are circular intersections with specific
design and traffic control features. Basic elements of
these intersections are the counterclockwise circula-
tion around a central island (for right-hand traffic
countries, like Greece), yield control of all entering
traffic to all oncoming circulation already moving
into the roundabout, channelized approaches and ap-
propriate geometric curvature, ensuring the desired
speed into the circulatory roadway.
2.2 Key features
For a better understanding of the following sections
the basic components of a roundabout design are
shown in Figure 1, such as:
 Intersection approaches
 Central island (9)
 Splitter islands (3)
 Circulatory roadway (8)
 Apron (10)
 Entrance line (5)
 Lane use horizontal markings (4)
 Outer borderline of the circulatory roadway (6)
 Accessible pedestrian crossings (2)
 Pedestrian refuge on splitter islands (1)
 Bicycle treatments
 Landscaping buffer (7)
 Road lighting pole (A)
 Sidewalk (B)
 Regulatory sign of circular path (C)
 Exit sign (D)
 Yield control sign (E)
Figure 1. Key roundabout features (MITN, General Secretariat
of Public Works, Directorate of Road Work Studies, 2011)
2.3 Roundabout categories
Depending on the size and number of approach
lanes, roundabouts are divided into three main cate-
gories:
 Mini-roundabouts
 Single-lane roundabouts
 Multilane roundabouts
Taking into account the external environment of
the roundabout, namely, depending on whether a
roundabout is constructed in urban, rural or subur-
ban area, there is a further differentiation in the fol-
lowing subcategories:
 Mini-roundabouts
 Urban compact roundabouts
 Urban single-lane roundabouts
 Urban double-lane roundabouts
 Rural single-lane roundabouts
 Rural double-lane roundabouts
Generally, urban roundabouts are characterized
by smaller inscribed circle diameter, due to spatial
constraints. Moreover, special attention should be
given to the choice of the geometric characteristics
of the roundabouts according to its location securing
thus the safety of vulnerable users i.e. pedestrians
and cyclists. More specifically in rural areas ap-
proach speeds are higher and therefore great impor-
tance should be given on adequate visibility and sec-
tion detail issues. Generally it could be said that,
keeping the external environment constant, rounda-
bouts with the same number of lanes tend to show
similarities in specific design elements.
3 GEOMETRIC DESIGN OF ROUNDABOUTS
Roundabouts require specific geometric design and
show some peculiarities in their operation. The gen-
eral idea of their design is based on the reduction of
conflict points of traffic and of prevailing speeds.
Roundabouts’ main characteristics are the curved
entries and exits to ensure low speeds, the giving
way to the already moving within the roundabout
traffic, the deflection of traffic when entering the
roundabout, the traffic circulation around the central
island, the existence of splitter islands along the ap-
proaches and the pedestrian traffic only through the
appropriate crossing locations.
The basic principles that should govern the geo-
metric design of each roundabout category are:
 Adequate traffic deflection to ensure low entry
speeds and consistent speeds within the rounda-
bout.
 Proper alignment of the approaches, in order to
avoid exiting-circulating conflicts.
 Appropriate number of lanes to achieve the de-
sired traffic capacity.
 Smooth channelization aimed at comfortable driv-
ing.
 Adequate accommodation for the design vehicle.
 Design to meet the needs of pedestrians and cycl-
ists.
 Adequate visibility, for driver recognition of the
existence of the roundabout and the conflicting
traffic.
The main design elements relating to roundabouts
are listed below and are shown in the following fig-
ure (Fig. 2):
 Inscribed circle diameter
 Circulatory roadway width
 Entry width
 Exit width
 Entry radius
 Exit radius
In countries where roundabouts are constructed
and used mainly for reasons of road safety and pro-
tection of vulnerable users, there are no guidelines
for design elements that increase traffic capacity and
roundabouts are constructed much smaller compared
to those of other countries. The opposite occurs in
countries in which the priority is given to increased
traffic capacity, where roundabouts are having less
compact form.
Figure 2. Key roundabout geometric dimensions (Federal
Highway Administration, 2000)
4 ROAD SAFETY AT ROUNDABOUTS
4.1 General
The level of safety provided by a roundabout is
mainly related to its particular design characteristics.
In specific, their design converts all traffic move-
ments into right turns minimizing thus potential con-
flicts (Fig. 4). The reduction of conflicts due to the
physical geometric characteristics of roundabouts
has proved to be more effective in comparison with
the expected, but not certain, driver’s obedience to
the traffic regulations (Fig. 3). A second key advan-
tage that is achieved by the operation of a rounda-
bout and ensured by the appropriate geometry is
speed management, which provides multiple bene-
fits for users, regarding the safety factor such as:
 Reduction of the frequency and severity of colli-
sions for vehicles, pedestrians and cyclists as well
as for other vulnerable groups (elderly, children,
people with disabilities, etc.).
 Provision of more time for drivers approaching
the intersection and consequently better visibility
in order to adjust their speed and carefully enter
the roundabout (particularly at the intersections of
urban with rural roads, where speed limits are dif-
ferent).
 Insurance of safer penetration of incoming circu-
lation into the roundabout.
 Provision of more time to all users so that they
could identify and correct any traffic behavior
faults, which is very crucial especially for new
users.
Figure 3. Vehicle conflict point on exit because of wrong lane
selection of the moving vehicle into the roundabout (National
Cooperative Highway Research Program, 2010)
Figure 4. Vehicle and vehicle-pedestrian conflict points in a
crossing junction and roundabout respectively (4 approaches)
(MITN - General Secretariat of Public Works - Directorate of
Road Work Studies, 2011)
4.2 Statistics referring to the impact of roundabouts
on road safety
Numerous "before and after" studies are internation-
ally undertaken, in order to investigate the positive
impacts on road safety derived from the conversion
of a conventional intersection to a roundabout.
The results of a study by Robinson et al. (2000)
given in Table 1 reveal significant reductions in ac-
cidents after the conversion of conventional intersec-
tions to roundabouts in various countries. Specifical-
ly, the reduction for the United States rose to 51%,
for the United Kingdom ranged between 25-39%,
for France between 57-78% and for Australia be-
tween 45-87%. In addition the same study for single-
lane roundabouts resulted in a reduction of 73% in
the number of accidents involving injuries, while in
multi-lane roundabouts, the accidents involving inju-
ries have been decreased by 31%. Also Ogden
(1996) demonstrated that the construction of roun-
dabouts can reduce accidents by 60-80% in high-
speed intersections and by 50-80% in low-speed in-
tersections. Later, Elvik (2003) found slightly small-
er road safety improvement rates, in cases of roun-
dabouts which replaced signalized intersections
instead of intersections regulated only with STOP
and give way signs (i.e. reductions of 59% and 46%
instead of 64% and 53% for fatal and serious acci-
dents respectively).
 Finally, a 2004 study in the UK showed that out of accidents involving cyclists and pedestrians re-
of a total of 207,400 road accidents involving inju- spectively. Brüde & Larsson (2000) found that the
ries, the 18,000 (8.7%) occurred at roundabouts with rate of reduction for vulnerable users is highly re-
the rate of fatal accidents at roundabouts being lated with the roundabout category. More specifical-
0.35%, while at other crossings 0.88% (The High- ly for accidents involving cyclists they found reduc-
ways Agency, Transport Scotland, Welsh Assembly tions of the order 21% at single-lane roundabouts
Government, The Department For Regional Devel- and 112% for multi-lane roundabouts, while the re-
opment Northern Ireland, 2007). spective reductions for accidents involving pede-
strians were 79% and 12% for the two roundabout
Table 1. Mean crash reductions in various countries categories. Stone et al. (2002) showed an even lower
after the construction of roundabouts. reduction percentage in accidents involving pede-
Country Mean reduction (%) strians, of around 7%.
All crashes Injury crashes In addition a survey conducted in Great Britain
United States 37 51 found that accident rates involving pedestrians are
United Kingdom - 25-39
Germany 36 -
much higher at signalized intersections than rounda-
France - 57-78 bouts (Table 2).
Australia 41-61 45-87
* Source: Robinson et al., 2000. Table 2. British crash rates for pedestrians at roun-
dabouts and signalized intersections.
However, the statistics that have emerged from Intersection type Pedestrian crashes
these studies do not show the same level of road per million trips
safety improvements for the various countries. This Mini-roundabout 0.31
Conventional roundabout 0.45
is likely due to the existing differences from country Flared roundabout 0.33
to country: Signals 0.67
 Ιn traffic volumes. * Source: Federal Highway Administration, 2000.
 Ιn the design of the various geometric and opera-
tional elements (operation of only single-lane However a French survey in 15 cities concerning
roundabouts in some countries). the comparison of accident distribution per user cat-
 In the definition of accidents occurred at intersec- egory, between roundabouts and the rest conven-
tions as each country defines different distance, tional intersections showed that accidents involving
over which an accident is not considered to be as- bicycles and mopeds constitute almost the one quar-
sociated with the specific intersection. ter of the total accident percentage occurring at the
 In the definition of accidents involving injuries roundabouts, while the respective value for the rest
and more specifically in the definition of death of intersections is around 15% (Table 3). The above
due to accident as this varies among countries de- result shows that nevertheless the overall reduction
pending on the period in which death occurs (just of accidents involving vulnerable road users at roun-
on the spot or some days after staying in hospit- dabouts they continue to remain significantly high, a
al). fact that should be treated properly.
 In cultural differences which influences driving
behavior. Table 3. Crash percentage per type of user at roun-
In any case, the comparison of statistics should be dabouts and conventional intersections in 15 towns
made for roundabouts with similar geometric and in western France.
operating characteristics, taking into account as well User category Other crossings Roundabouts
any existing differences. In this way, wrong assess- (%) (%)
ments are avoided and correct and reliable conclu- Pedestrians 6.3 5.6
sions can be made, in order to proceed with the ap- Bicycles 3.7 7.3
propriate measures that should be taken towards the Mopeds 11.7 16.9
improvement of road safety at roundabouts. Motor cycles 7.4 4.8
Regarding the impact of roundabouts on vulnera- Cars 65.7 61.2
ble users’ safety, the survey results are not so clear. Utility vehicles 2.0 0.6
Schoon & van Minnen (1993) reported that the re- Heavy goods vehicles 2.0 3.0
duction in the number of accidents involving cyclists Bus/coach 0.8 0.6
at roundabouts reaches 30%, well below 47%, which Miscellaneous 0.4 0.0
represent the overall road safety improvement rate Total 100.0 100.0
that roundabouts induced. On the other hand, the * Source: Federal Highway Administration, 2000.
same survey indicated that roundabouts reduce the
number of accidents involving at least one pede-
strian by 89%. Hydén & Várhelyi (2000) also dem-
onstrated reductions of 60% and 80% in the number
Table 4. Percentage of accidents by accident type at roundabouts in various countries.
Country Single-vehicle Approaching Entering-circulating Other vehicle Pedestrian
accidents (%) accidents (%) accidents (%) accidents (%) accidents (%)
United Kingdom 30 25 20 18 6
Germany 28 17 30 - -
France 28 7 37 15 10
Australia 18 22 51 9 9
* Source: Kennedy, 2007.
** Values for Germany and Australia refer to accidents that caused only property damage.

4.3 Distribution of accidents by accident type in ratio leads to greater chances of accidents due to
relation to geometric design elements large traffic width within the circulatory roadway.
 The angle between the centerlines of two succes-
The main accident types that occur in a roundabout
sive approaches. A wide angle reduces the neces-
are:
sary vehicle path deflection and thereby increases
 Single-vehicle accidents, mainly due to crashes
the risk of an accident, both at the entrance and at
towards fixed objects and elements of the roun-
the exit.
dabout.
 The approach width relative to the entry width, as
 Other accidents between vehicles, including colli-
this indicates the level of widening and therefore
sions between vehicles already moving into the
the possibility of an accident due to collisions
roundabout, collisions between vehicles which
with adjacent vehicles.
are already moving into the roundabout and those
 The visibility to the left. Increased visibility at the
exiting the roundabout, collisions between ve-
hicles leaving and vehicles entering the rounda- entrance and to the left leads to increased entry
speed and eventually causes more single-vehicle
bout, collisions at the entrance and exit etc.
accidents, but also accidents between incoming
 Accidents involving pedestrians and other vulner-
and already moving into the roundabout traffic.
able users, when one of them is hit by a passing
vehicle.
 Accidents between the incoming and already 4.4 Accident frequency and severity
moving into the roundabout traffic.
Kennedy (2007) tried to summarize, from a number
 Accidents during approach (e.g. due to lane
of surveys that took place in various countries, the
changing and path overlapping when approaching
average accident frequency values at roundabouts-
the roundabout).
expressed in accidents per year. The results of this
Table 4 gives the accident distribution by acci-
study are given in Table 5 and they are based on the
dent type, for various countries.
sample of roundabouts selected for the surveys refer-
The accident distribution by accident type de-
ring to four-arm roundabouts with single-lane circu-
pends on the roundabout category i.e. small rounda-
latory roadway. Specifically it was found that in the
bouts with small diameter of the central island have
United States the frequency was 1.5 accidents per
many more accidents between incoming and already
year for a small number of roundabouts (11) and
moving into the roundabout traffic, while double-
with reference to the year 1998. In the UK, a 2007
lane roundabouts have more single-vehicle accidents
survey gave a frequency of 1.77 accidents per year,
and accidents during the approach to the roundabout.
while accident severity, expressed as the percentage
The most important geometric elements affecting
of fatal and serious accidents of total accidents, was
accidents and are related directly or indirectly to
7%. In France it was found a frequency of 0.05 acci-
their incidence are:
dents per year, based on data of a 2005 survey for
 The curve of the entry path. By increasing this
about 27,000 roundabouts, while in Australia a fre-
curve and thereby the deflection of the vehicle
quency of 0.6 accidents per year for a sample of 290
path when entering (reduction of the entrance
roundabouts, according to a research undertaken in
path radius), accidents between incoming and al-
1998.
ready moving into the roundabout traffic are be-
According to surveys and as expected, the fre-
ing reduced, while accidents at the entry and sin-
quency of accidents at roundabouts increases with
gle-vehicle accidents are increasing.
the number of approaches and the number of lanes,
 The entry width. The larger the entry width, the
due to the increased traffic flow and hence the in-
greater effect it has on the increase of accidents
crease of conflict points for both cases. Table 6
between incoming and already moving into the
presents the results of a survey conducted in 2004 in
roundabout traffic, while reducing the accidents
Great Britain, on the frequency and severity of acci-
at the entry.
dents, for a sample of 1162 roundabouts and for a
 The ratio of the inscribed circle diameter to the di-
five-year period, depending on the number of ap-
ameter of the central island. Great value of this
proaches and lanes.
Table 5. Accident frequency and severity at rounda- Table 7. Accident rates at roundabouts in different
bouts in different countries. countries.
Country No. of roun- Accident Accident Country No. of Accident rate Mean total
dabouts in frequency severity (% roun- (accidents per vehicle
study (accidents fatal and dabouts 100 million inflow
per year) serious) vehicles) (vehicles per
United States 11 1.5 - day)
United Kingdom 1162 1.77 7 United States 11 8 -
France 27000 0.05 - United Kingdom 44 36 28000
Australia 290 0.6 - Germany - 53-162 5000-25000
* Source: Kennedy, 2007. France 179 4.5 -
Australia - 4-8 -
Table 6. Average accident frequency and severity at * Source: Kennedy, 2007.
UK roundabouts by number of arms, for the period
1999 to 2003.
No. No. Accident frequency 5 CASE STUDY: FOUR AT GRADE
of of (accidents/year) ROUNDABOUTS – MUNICIPALITY OF
arms sites Single Double Grade Accident THERMI
cway cway separatedseverity (%
roads roads junctions fatal and In this case study the design elements of four roun-
serious
3 326 0.63 1.28 0.79 9.3 dabouts in the Municipality of Thermi are described
4 649 1.08 2.65 1.79 7.1 in relation to the benefits derived for the area on the
5 157 1.72 3.80 3.66 7.1 topic of road safety and traffic capacity. Municipali-
6 30 2.11 4.62 5.95 5.2 ty of Thermi is located on the southeast side of the
All 1162 1.00 2.60 1.87 7.2 city of Thessaloniki, Greece and north to the airport
* Source: The Highways Agency, Transport Scotland, Welsh "Macedonia" and according to the National Statistics
Assembly Government, The Department For Regional Devel-
opment Northern Ireland, 2007. Office, in 2001 its population was 11,412.

Here it should be pointed out that accident rate-

defined as accidents per 100 million vehicles travel-
ing through the roundabout- is a more representative
indicator for comparisons between sites but its esti-
mation depends on the availability of traffic flow da-
ta, which is not feasible in most cases. Thus, the ta-
ble below (Table 7) provides values of the accident
rate for the countries studied so far, where data ac-
quisition was feasible. The values given refer to sin-
gle-lane roundabouts, except for the United King-
dom, where they refer to larger roundabouts. In the
United States, the accident rate is 8% for a sample of
just 11 roundabouts, based on data for the year 2003
and in the United Kingdom it appears to be quite
high, reaching 36%, based on a 2007 survey refer- Figure 5. Municipality of Thermi, Thessaloniki Prefecture,
ring to a sample of 44 roundabouts with traffic flows Region of Central Macedonia, Greece (http://el.wikipedia.org/)
ranging between 10,000 and 50,000 vehicles per
day, with an average value of 28,000 vehicles. In
Germany the accident rate ranges between 53 and The Municipality had the highest population
162 according to data from a research undertaken in growth in the decade 1991-2001. This increase is
2005, a value too high which is due to the inclusion due to the decentralization that took place during
of only property damage accidents. The average ac- this decade as a large part of the population moved
cident rate in 179 urban roundabouts in France ac- to the suburbs of the city of Thessaloniki. The road
cording to data of a 1991 research is only 4.5%, network of the settlement of Thermi is radial with
while in Australia it ranges between 4 and 8 acci- Paramana square at its center, being crossed by 11
dents per 100 million vehicles, also including only roads.
property damage accidents, based on a 1993 survey.

Figure 6. Roundabouts position (https://maps.google.gr/)
According to the police statements the conversion
from a junction of other type to roundabout resulted
to an enhanced effect on safety for all examined
junctions. This positive effect, which varies from
junction to junction according to its design elements,
is mainly due to the reduced speed and the increased
awareness aiming to deflect from ahead path. Addi-
tionally, before and after remarks show that rounda-
bouts’ application genuinely improved the level of
service and the functionality of the intersections by
increasing the circulatory capability.
5.1 Ap. Kougiami Street - P. Tsaldari Street
intersection
Road safety interventions that have taken place in
the municipality of Thermi include the construction
of a roundabout at the site where Ap. Kougiami
street (1) intersects with P. Tsaldari street (2), in July
2014, replacing a non-signalized intersection. The
main reason for the construction of the roundabout
under consideration was the emergence of severe ac-
cidents involving pedestrians at this site, although it
is sparsely populated.
1 This at-grade roundabout was constructed in early
2 giami street (2) in the expanding urban plan of the
Figure 7. Intersection before the construction of the
roundabout (https://maps.google.gr/)
Figure 8. Roundabout design (Directorate of Technical
Services, Municipality of Thermi, Thessaloniki, Greece)
The intersection was converted into a single-lane
roundabout with one lane of traffic in each direction
per approach. The inscribed circle diameter is 26m.
and the central island’ diameter is 16.8m. resulting
to the circulatory roadway width of 4.6m. For the
Ap. Kougiami street splitter islands were also con-
structed, while remained the existing islands on the
P. Tsaldari street.
Figure 9. Intersection during the construction of the
roundabout (Directorate of Technical Services, Municipality of
Thermi, Thessaloniki, Greece)
5.2 Charilaou - Thermi Street / Ap. Kougiami Street
intersection
2014, on the provincial road of Charilaou - Thermi
(1), at the position of the intersection with the Kou-
settlement of Thermi.
The provincial road of Charilaou - Thermi con-
nects the settlement of Thermi with a non-residential
area where the prevailing operational speeds are
high with at the same time a significant presence of
pedestrians. Also, it is worth mentioning the increase
in traffic volumes in this site due to the attraction of
the area for business as well as residential purposes.

1 6m. in order to achieve simultaneous circulation of
Figure 10. Intersection before the construction of the roun-
dabout (http://gis.ktimanet.gr/wms/ktbasemap/default.aspx)
Thus the construction of this roundabout in the
place of a previously signalized intersection
emerged as necessary aiming to slow down the ob-
served high speeds and consequently to avoid acci-
dent occurrence as well as to facilitate the traffic
from and to the settlement of Panorama through a
new non-signalized street.
Figure 11. Roundabout design (Directorate of Tech-
nical Services, Municipality of Thermi, Thessaloni-
ki, Greece)
The construction included a double-lane rounda-
bout with two lanes of traffic in each direction per
approach and for the circulatory roadway. The inter-
section point of the roadways corresponds to chai-
nage 0 + 887.66 on the main road (Route Thessalo-
niki - Thermi) and to chainage 0 + 962.24 for the
perpendicular street. The content between the axes
angle is â = 91.211grad. The inscribed circle diame-
ter is 36m. and the central island has a diameter of
24m. As a result, the circulatory roadway width is
traffic in two lanes and avoid delays at the entries
and exits. All four splitter islands are kerbed so as to
avoid their use from traffic.
Figure 12. Intersection during the construction of the
roundabout (https://maps.google.gr/)
5.3 B. Tavaki Street / Sot. Petroula Street
intersection
B. Tavaki street (1) constitutes the entrance to the
settlement of Thermi for vehicles entering the Thes-
saloniki - Polygyros highway through its intersection
with the Sot. Petroula street (2). Thus, this intersec-
tion is characterized by the high speed of entering
vehicles in the settlement. In addition conflict points
were observed due to the drivers’ violation of the
prohibited left turns to Sot. Petroula street during
their entry to the settlement.
1
Figure 13. Intersection before the construction of
the roundabout
(http://gis.ktimanet.gr/wms/ktbasemap/default.aspx)
 For this reason, in August 2012 a roundabout was
constructed, in the place of a previous, insufficient
to regulate the traffic, triangular island with two tri-
angular splitter islands at the entrances of B. Tavaki
and Sot. Petroula street aiming to achieve the reduc-
tion of conflict points between traffic flows and in-
crease pedestrian safety.
The roundabout constructed is a single-lane roun-
dabout with one lane of traffic in each direction for
the two main approaches and one for the circulatory
roadway. The inscribed circle diameter is 28m. and
the central island has a diameter of 17m. As a result,
the circulatory roadway width is 5.5m. For the two
main approaches splitter islands were also con-
structed.
Figure 14. Roundabout design (Directorate of Technical Ser-
vices, Municipality of Thermi, Thessaloniki, Greece)
Figure 15. Intersection after the construction of the roundabout
(https://maps.google.gr/)
5.4 Capetan Chapsa Street - El. Venizelou Street
intersection
The Capetan Chapsa (1) (continuation of the Chari-
laou - Thermi street into the settlement) - El. Veni-
zelou (2) intersection was problematic in terms of
road safety, since it was a non-signalized site with
high pedestrian use, where eight approaches ended
with only two one-way streets. For this reason, the
chosen solution was a single-lane roundabout con-
struction, completed in 2012, aiming to improve
traffic capacity and increase pedestrian safety. How-
ever, the roundabout was poorly designed with the
centerlines of the approaches not intersecting in its
center and with the absence of pedestrian crossings
in all its approaches.
2
1
Figure 16. Intersection before the construction of the rounda-
bout (http://gis.ktimanet.gr/wms/ktbasemap/default.aspx)
Figure 17. Intersection after the construction of the roundabout
(https://maps.google.gr/)
Figure 18. Photo looking north-east along El. Venizelou street
(https://maps.google.gr/)
6 DISCUSSION AND CONCLUSIONS
Roundabouts are often a standard solution when it
comes to the conversion or the original design of an
intersection, in both urban and rural environment.
The attitude of the public is favorable towards them,
as both drivers and pedestrians benefit from the mi-
nimization of delays, even at peak hours, as opposed
to signalized intersections. This improved level of
service reflects the contribution of roundabouts to
the increase of traffic capacity and to the general
functional benefits.
Technically, the geometry of roundabouts aims to
reduce speeds, contributing thereby to a higher safe-
ty level, which is considered to be the most impor-
tant advantage roundabouts offer. Examples include
the reduction of the number of accidents and in
many cases the limitation of their seriousness only to
property damage. Crash reductions resulting from
conversion of conventional intersections to modern
roundabouts can be attributed primarily to two fac-
tors: reduced traffic speeds and elimination of spe-
cific types of motor vehicle conflicts that frequently
occur at angular intersections. These conflicts in-
clude left turns against opposing/oncoming traffic,
front-to-rear conflicts (often involving the lead ve-
hicle stopping or preparing to stop for a traffic signal
or stop sign), and right-angle conflicts at traffic sig-
nals and stop signs.
The adaption of roundabouts in the landscape and
their potential aesthetic treatment are also consi-
dered to be positive elements, because a smooth con-
tinuation of the current situation is maintained,
without elements that do not match and do not be-
long to their wider installation environment being
observed. Another benefit that favors the environ-
ment is the reduction of the energy consumed by ve-
hicles and the gases emitted, due to the smoother
flow of traffic, with fewer accelerations and decele-
rations, since absolute stops of the vehicles are
avoided.
The data collected for the case study in the Muni-
cipality of Thermi showed a general safety and traf-
fic capacity benefit from converting conventional in-
tersections, either from stop sign or traffic signal
control, to modern roundabouts. It was pointed out
that the construction of the four roundabouts im-
proved significantly road safety at these junctions,
eliminating conflict points between traffic flows and
ensuring their safer use by drivers and pedestrians.
However, the construction of roundabouts is not a
panacea. Like any transportation project, so the in-
stallation of roundabouts has drawbacks and causes,
where appropriate, negative effects (large space re-
quirements, expropriations of significant resources,
increased construction cost, etc.). Often, indeed, sig-
naling an intersection outweighs the construction of
a roundabout, depending on the circumstances, the
users that will be served and the objectives set for
any treatment. For this reason, caution and care on
the part of engineers is required, to fully understand
the needs and constraints of the area and take the
right decision, to meet the drivers’ and pedestrians’
needs and avoid blind construction of improper
treatments.
7 REFERENCES
Brabander, B., Nuyts Ε. & L. Vereeck 2005. Road safety ef-
fects of roundabouts in Flanders. Journal of Safety Re-
search 36: 289 – 296.
Brabander, B. & L. Vereeck 2006. Safety effects of rounda-
bouts in Flanders: Signal type, speed limits and vulnerable
road users. Accident Analysis and Prevention 39 (2007):
591–599.
Brilon, W. 2011. Roundabouts: A State of the Art in Germany.
Paper presented at the 3rd International TRB-roundabout
Conference. Carmel, Indiana.
Brüde, U. & Larsson, J. 2000. What roundabout design pro-
vides the highest possible safety? Nordic Road Transport
Res. 2, 17–21.
Directorate of Technical Services, Department of Transporta-
tion, Traffic, Hydraulic and Reclamation Projects, Munici-
pality of Thermi 2013. Project study: "Road safety interven-
tions on Ap. Kougiami Street of the settlement of Thermi",
No. of Study: 41/2013. Thessaloniki, Greece.
Directorate of Technical Services and Environment, Depart-
ment of Technical Works and Urban Planning Applications
2010. Project: "Paving regeneration on V. Tavaki and D.
Karaoli Streets of the municipality of Thermi", No. of
Study: 97/2010. Thessaloniki, Greece.
Efraimides E. & Iordanopoulos P. 2008. Thesis: "Evaluation of
implementing Traffic Calming Measures in urban areas:
The case of Thermi". Thessaloniki, Greece.
Elvik, R. 2003. Effects on road safety of converting intersec-
tions to roundabouts. Review of evidence from Non-U.S.
studies. Transportation Research Record 1847, paper 03-
2106.
Federal Highway Administration 2000. Roundabouts: An In-
formational Guide. U.S. Department of Transportation.
Publication No. FHWA-RD-00-067.
Hydén, C. & Várhelyi, A. 2000. The effects on safety time con-
sumption and environment of large scale use of rounda-
bouts in an urban area: a case study. Accid. Anal. Prev. 32
(1), 11–23.
Institute of Transportation Engineers 2008. Enhancing Inter-
section Safety Through Roundabouts: An ITE Informational
Report. Publication No. IR-127. Washington, D.C..
Insurance Institute for Highway Safety 2000. Crash Reductions
Following Installation of Roundabouts in the United States.
Kennedy, J. 2007. International comparison of roundabout de-
sign guidelines. Published Project Report PPR206.
Ministry of Infrastructure, Transport and Networks (MITN),
General Secretariat of Public Works, Directorate of Road
Work Studies 2011. Road Work Studies Instructions, Vo-
lume 10, Part 2: Roundabouts (RWSI). Consulter: NAMA
Consulting Engineers & Planners SA (Project).
National Cooperative Highway Research Program 1998. Mod-
ern Roundabout Practice in the United States. NCHRP
Synthesis 264. Α Synthesis of Highway Practice. Georges
Jacquemart, P.E., AICP, Buckhurst Fish & Jacquemart Inc.,
Transportation Research Board - National Research Coun-
cil. Washington, D.C..
National Cooperative Highway Research Program 2010. Roun-
dabouts: An Informational Guide. Second Edition. NCHRP
 Report 672. U.S. Department of Transportation, Federal
Highway Administration (FHWA). Washington, D.C..
Ogden, K. W. 1996. Safer roads: A guide to road safety engi-
neering. Monash University: Institute of Transport Studies.
Robinson, B. W. et al. 2000. Roundabouts: An informational
guide. Washington: U.S. Department of Transportation.
Federal Highway Administration. Report FWHA-RD-00-
067.
Schoon, C. & van Minnen, J. 1993. Ongevallen op rotondes II:
tweede onderzoek naar de onveiligheid van rotondes vooral
voor fietsers en bromfietsers. Stichting Wetenschappelijk
Onderzoek Verkeersveiligheid (SWOV), rapport R-93-16.
Leidschendam.
Spanou, D. 2014. Thesis: "Roundabouts: Geometric design,
Adaption in landscape and Road Safety - Case study in the
city of Kavala". Aristotle University of Thessaloniki. Facul-
ty of Engineering. Department of Civil Engineering. Post-
graduate Specialization Programme "Environmental Protec-
tion and Sustainable Development". Thessaloniki, Greece.
Stone, J.R., Chae, K. & Pillalamarri, S. 2002. The effects of
roundabouts on pedestrian safety. Southeastern Transporta-
tion Center, working paper. Tenessee.
The Highways Agency, Transport Scotland, Welsh Assembly
Government, The Department For Regional Development
Northern Ireland 2007. Design Manual for Roads and
Bridges: Geometric Design of Roundabouts. Volume 6:
Road Geometry, Section 2: Junctions, Part 3. TD 16/07.