Setra 1998 - Design of Major Rural Junctions
Setra 1998 - Design of Major Rural Junctions
Setra 1998 - Design of Major Rural Junctions
Translation :
Chapters 1 et 2 : Kevin RILEY, Traductions Routes et Transports (Mûrs-Erigné - 49)
Chapter 3 and appendices : Joe BARED, FHWA ( U.S. Washington)
Note: The first chapters have been translated into European English, the third and appendices into
American English.
1
FOREWORD
This technical guide deals with the general design and geometry of at-grade intersections on
major interurban roads. It sets out detailed technical guidelines on this topic which complement the
more general guidelines contained in the publication "Aménagement des Routes Principales".
It replaces two SETRA technical guides: "Les carrefours plans sur routes interurbains" which
dates from March 1980 and " Les Carrefours plans sur routes interurbaines - Carrefours giratoires"
which dates from September 1984.
It covers neither intersections in urban areas - these will be dealt with in a document to be
published by the CERTU ("Guide carrefours urbains") - nor signalized intersections which should not
exist outside built-up areas.
The guidelines in this document should be considered as the rules of good practice the whole
engineering community.
♦ L. DUPONT (SETRA)
♦ L. PATTE (SETRA)
♦ P. BOIVIN (SETRA)
♦ P. FLACHAT (CETE de Rhône-Alpes)
♦ B. GUICHET (CETE de l'Ouest)
♦ J.Y. GIRARD (CETE de l'Ouest)
♦ G DUPRE (CETE de Normandie Centre)
The drawings and diagrams are the work of G. LEPINE (SETRA) and J.Y. LEBOURG (CETE de
Normandie Centre).
The term "major roads" is applied in the ARP ("Aménagement des Routes Principales") to roads
which perform a structural function within the national road network or the Département road networks
(the daily traffic on such roads generally exceeds 1500 vehicles).
2
CONTENTS
u FOREWORD 5
1. GENERAL PRINCIPLES 27
1. GENERAL PRINCIPLES 69
3. SPECIAL ACCOMODATIONS 89
u GLOSSARY 99
u BIBLIOGRAPHY 107
u APPENDICES 111
3
4
INTRODUCTION
u SCOPE
This guide deals with the design and construction of at-grade intersections on major roads
outside urban areas and covers both upgrading of the existing network and the building of new roads.
As at-grade intersections are not compatible with roads that are isolated from their environment,
this document does not deal with the geometric design of type L roads (freeways) or type T roads
(express roads)1. However, the fundamental principles and the approach to be adopted when choosing
the type of intersection, which are explained in the first chapter "General Design", apply to all types of
intersections on all types of roads.
This guide does cover Type R bypasses2. However, it does not deal with ring roads; as the
function of most of these is to link different districts they should be considered as urban roads.
Where buildings are either few in number or widely spaced (perhaps constituting a hamlet but
not a built-up area as defined in the Highway Code (art. R1)), the site will generally be classified as non-
urban and the technical guidelines for rural areas will be applied.
Cross-town routes, irrespective of the size of the built-up area they pass through, are to be
considered as urban roads. For these it is necessary to refer to the texts which deal with urban roads, in
particular the CERTU "Guide Carrefours Urbains". In addition, Chapter 7 of the A.R.P. states the general
principles that apply to the boundary between rural and urban areas and entrances to built-up areas and
contains some remarks concerning roads that pass through small towns and suburbs at the entrances
to large towns and cities.
1
Apart from the very specific exception of the use of a roundabout to provide a "terminal" at the end of a Type T road.
2
A bypass is defined in the A.R.P. as a non-urban road passing round a town and which is mainly used by through traffic.
5
u CONTENTS OF THE GUIDE
This guide contains technical guidelines for the general design and the geometry of intersections
and also enumerates the general principles which the designer should take into account.
However, it does not give detailed descriptions of technical features that are the subject of
specialized guidelines. Although they are mentioned, particularly in the first chapter, preliminary
investigations and study techniques are not covered either. For these topics the text refers readers to
other documents.
It takes into account the most recent developments and accumulated experience with regard to
level of service and the interaction between the infrastructure and road safety3. Consequently, design
rules and the descriptions of the conditions under which the various types of intersection should be
used have been improved or clarified.
This new intersection guide has some major aspects in common with the ARP, and
supplements the recommendations in this. Essentially, these are as follows:
- a definition of technically coherent types of road which are clearly identifiable by drivers;
- acceptance that drivers have the "right to make an error", in particular by designing roadsides devoid of
obstacles and features which are likely to make the consequences of leaving the road more severe.
Consequently, this document modifies the documents it replaces in many important respects (see
Foreword above).
3
The SETRA-CETUR document "Sécurité des Routes et des Rues" presents a survey of this knowledge.
6
u STRUCTURE OF THE GUIDE
A coherent design approach consists of a number of stages which can be divided into two main
phases:
• in the first, the intersection is adapted to its surroundings. This is covered in Chapter 1, "General
Design". This describes the data to be collected, the preliminary studies to be conducted and the
range of facilities which are possible for each type of road. Finally, it states the factors which to be
considered when deciding whether to construct a standard at-grade intersection or a roundabout;
• the second phase is geometric design. This is covered in the second chapter for standard at-grade
intersections and in the third for roundabouts. These chapters both cover general features and the
detailed design of the components which form the intersection.
The principal technical terms used in intersection design are defined in a glossary. Its main
purpose is to eliminate any terminological problems that readers may encounter when reading this
document.
Finally, the appendices contain additional information on safety, the 85th percentile speed and
the design of curbs.
Considerations relating in particular to the safety, level of service and cost of facilities have
resulted in guidelines which deal with choosing the type of intersection, general configuration, sizing and
detailed geometry. In practice, a flexible approach should be adopted towards these, taking account of
local constraints and accident occurrence etc.
In addition, the requirements for existing roads (to which extremely strong constraints may
apply) cannot possibly be as strict as for new roads. For the latter, the principles and guidelines
contained in this document should provide the basis for intersection design.
For existing roads, the design rules which apply for new constructions are to be considered as
medium-term or long-term objectives. Safety, based on a detailed analysis of accidents, should be
taken into account when deciding which improvements should be given priority. However, prioritization
should never interfere with an overall approach to design along a route.
It is generally advisable (for the sake of homogeneity, consistency and performance) to adhere
as closely as possible to the standard layouts in this guide. However, merely doing this blindly will not in
itself guarantee that optimum performance will be achieved. Each intersection is a special case and
specific studies will be required in order to take account of the specific features of the project.
7
8
CHAPTER 1
GENERAL DESIGN
u 1. BASIC PRINCIPLES OF INTERSECTION DESIGN 10
2.1. DATA 12
2.5. SPEEDS 16
3.1. PRINCIPLES 16
9
To operate effectively an intersection must:
- be appropriate for the type of road on which it is located, for the site (environment etc.) and
the conditions of use (types of traffic, etc.);
- have satisfactory geometric design (both of general features and detail).
This chapter deals with the general design of intersections and provides an overview
of the first of the above points, i.e. the main principles to be followed, and describes the
process which is to be used to select the type of intersection.
• identifying the type of road 3 involved (or rather the types of roads involved). To make
such an identification, studies (or perhaps even a planning decision) are required to
establish: (i) its hierarchical position in its network; (ii) the type of road it most resembles
(before construction or after upgrading). If this preliminary work at network level has not
been performed, appropriate studies should be undertaken to determine the type of the
major road (see 3.2.);
10
• collecting and analyzing information concerning the site (topography, environment,
visibility obstructions, layout of existing road or roads) etc. and its operation (accidents,
traffic etc.) (see 2: "Preliminary studies and data");
• geometric design as such; this phase can be divided into three parts:
- determining the general configuration (for example, provision of a left-turn lane, positioning of
the legs of a roundabout, etc.);
- checks on visibility, capacity and general adequacy (simplicity, compactness etc.), which
can bring into question planned measures or earlier decisions.
• it must be compatible with the type of road and the behaviors this generates;
• it must fit in with the rationale of the route (uniformity of facilities, contribution to the
continuity or division of the major road);
• the facility must be legible allowing drivers to recognize the intersection they are entering
easily, quickly and with certainty;
• safety must be optimized for all traffic flows, including very secondary ones;
11
2. PRELIMINARY STUDIES AND DATA
2.1. DATA
The data below should generally be taken into account during design work for a new
intersection or upgrading an existing one. However, systematic data collection should not be
conducted in all cases. Furthermore, this is not an exhaustive list.
• the position of the intersection on the roads involved, the position of these roads in the
network to which they belong, the nature and status of these roads;
• use and operation of the intersection: traffic (volumes, nature, movements, etc.),
saturation phenomena, specific activities at the site (for example a diner, service station,
etc.), location of a bus stop, observable or predictable behaviors (as is the case), and, in
particular, operating speeds on the approaches to the exchange zones;
• accidents: number, types and processes (data only available for existing intersections).
This data provides the main basis for the safety diagnosis that must systematically be
conducted before intersections are modified. In the case of intersections on new
infrastructure, the accident risk can be estimated using "predictive" models, on the basis of
the traffic and the characteristics of the intersection.
The sections which follow (2.2 to 2.5) provide additional information about the
collection and analysis of some types of basic data (i.e. accidents, traffic, measurement of
visibility conditions, speeds).
12
2.2 PRELIMINARY SAFETY STUDIES
With regard to this, reference will be made to the methodological guide "Etudes
préalables à des interventions sur l'infrastructure" (SETRA, 1992). Use can also be made of
the "SECAR" software system for analyzing the safety of non roundabout at-grade
intersections (SETRA; CETE de Normandie - Centre)4 .
What constitutes essential traffic data depends on whether the project involves
modifying an existing intersection or designing an intersection on a new road.
In addition to the volume of traffic on the major road, which is generally known, the
traffic volume on the minor road(s) must be measured.
If accidents involving left turning traffic have occurred, or if there seems to be a high
proportion of exchange movements, the volume of left turning traffic should be assessed
(even if it appears to be small).
However, when traffic volumes are such that capacity problems (long delays for some
movements) are deemed likely, the volumes of all traffic movements must be estimated, for
example when there is heavy traffic on the minor road or very heavy traffic on the major road
(even with light traffic on the minor road). When the possibility of periodic congestion cannot
be completely ruled out more thorough knowledge of peak hourly flows5 must be acquired.
4
See SETRA Note d'information No. 113 (série circulation, sécurité, exploitation); March 1998.
5
The evening peak is generally involved. In some cases the reverse peak can provide useful additional
information; occasionally, a weekly peak hour, or even a seasonal peak can be considered, particularly when
one or other of these is exceeded for at least thirty hours during the year.
13
The graphs below show a few criteria for determining the traffic studies that should be
conducted. These are given for guidance only6, and are above all intended to warn the
engineer that such studies are necessary beforehand.
Fig 1 - A few criteria for the level of traffic studies that should be performed (rural roads with 2 or 3 lanes)
Four leg intersection
5000
4000
3000
the minor
ADT) 3
2000
2
1
1000
0
0 5000 10000 15000 20000
Traffic on major road (AADT)
7000
6000
5000
4000 3
3000 2
2000 1
1000
- The traffic levels may cause congestion problems; it is advised to carry out a check on the basis of the
distribution of traffic flows (and perhaps also peak hourly flows).
® Periodic congestion is highly likely; a detailed study using directional peak hour data is essential.
In addition to the traffic on the major road, traffic on the roads crossed by the new road
must be measured, and forecasts must be made of the future traffic on these roads once the
new road section has been opened and, in spite of the difficulties involved, of all the turning
movements too. More detailed investigation of peak hourly flows may be conducted,
especially when daily traffic volumes mean that the possibility of periodic congestion cannot
be ruled out.
6
These graphs are, in particular, based on simple hypotheses about the peak hour, the distribution of turning
movements etc.
14
2.3.3. TRAFFIC DATA COLLECTION METHODS
Essentially, the following traffic data collection methods are available for existing
roads:
- counts at permanent counting stations (in particular SIREDO stations) - these techniques
are usually limited to the structural network;
- automatic counts using temporary counters (often pneumatic tubes) which are installed for
one or more periods during a week, with the results then extrapolated for an entire year;
- manual peak hour counts, extrapolated to give the annual daily average.
With regard to these last two points, where traffic data is incomplete, details of the
methods for estimating the annual average daily traffic (AADT) and additional information are
given in the technical appendix of the "Circulaire relative aux recensement de la traffic
routière".
Directional traffic counts use the same techniques as comprehensive counts, but
there is no method for extrapolating non-permanent hourly counts to give an average hour or
an average peak hour.
The design of a safe project generally involves checking the entry approach visibility
(see 2.1.), and the crossing visibility for vehicles on the non-priority road (see Chap.2, 1.2.1).
Of the two, the second generally generates the most constraints. It is expressed either in
terms of a visibility distance or a crossing time. Several methods can be used to ensure that
visibility conditions is satisfactory.
This method can only be applied to existing intersections and consists of measuring
the actual intersection crossing time in situ. It is convenient in that it uses the speeds and
visibility distances that are involved in the problem of visibility. Instructions are given in
Appendix 5.
The method requires knowledge of the operating (V85) speeds of vehicles approaching
the intersection (see 2.5). The required sight distances depend on these (see Chap. 2, 1.2.2.)
and can be compared with those actually provided (because of lateral masking, road layout,
etc.). In the case of an existing road, visibility distances are measured in situ. Otherwise, they
are estimated a priori, using the plans.
15
2.5. SPEEDS
Generally, speeds are measured, in accordance with international practice, using the
85th percentile speed (V85), which is the speed below which 85 percent of drivers travel under
free-flow (uncongested) conditions. The V85 speed can be estimated experimentally or
theoretically (see appendix 4).
In the case of an existing road, the 85th percentile speed can be deduced from the
distribution of observed speeds. Measurements are usually made using one of the various
types of vehicle detectors or hand-held radar equipment7 . The latter offers operational
flexibility and is able to discriminate easily between free-flow and impeded vehicles, but is
difficult to use over long periods.
The V85 speed can also be estimated with formulae that take account of the main
features of site geometry (cross-section, horizontal alignment, longitudinal profile, etc.). The
"DIAIVI software (SETRA) can also be used to estimate the operating speed at each point of a
project.
3.1. PRINCIPLES
The types of intersections installed on a road help to make it clear to drivers what type
of road they are on. When intersections are too varied or inconsistent they are a source of
harmful ambiguity. For example, the construction of interchanges on an ordinary major road
generates behaviors downstream which are incompatible with the operating conditions of the
road in question (because of frontage access, at-grade intersections, etc.).
7
The document "Mesure des vitesse et ses applications" (SETRA 1997) lists available methods for measuring
vehicle speeds.
8
Irrespective of the representativeness of the measurement period (and the accuracy of measurement device),
about 25 measurements are required to estimate the V85 speed to an accuracy of ± 10 km/h, and 70
measurement to estimate it to within ± 5 km/h.
16
Furthermore, drivers on a certain type of road expect certain types of intersections.
For example, drivers on an express road, where there is normally no frontage access and
where intersections are grade-separated, would be unprepared for frontage access or an at-
grade intersection and fail to react appropriately and rapidly if a non-priority vehicle were to
cross the road.
Finally, the manner an intersection operates must be compatible with the operating
conditions of the type of road on which it is installed. For example, the difficulty of non-priority
movements (crossing or left turns) on a standard at-grade intersection is incompatible with
the width and speeds that exist on some roads (for example divided highways).
The document "Aménagement des Routes Principales" identifies the following types
of interurban road:
- "freeways" and similar roads (type L): these are divided highways with no at-grade access
or intersections;
- "express roads" (type T): these have a single roadway and no at-grade access points or
intersections
- divided type R roads ("interurban major roads"): there may be access to roadside plots,
which are usually not built on (but traffic is not allowed to cross the central reservation), and
at-grade intersections (roundabouts or partial intersections only);
• secondary roads in rural areas: the same characteristics as undivided type R roads, but
with less traffic (rarely more than 1,500 v/d) and smaller width (generally less than 5 m to
5.50 m).
As with the types of road, the classification of intersections into homogeneous groups
of designs must favor the clearest possible perception by drivers. Here too, the appearance of
an intersection must suggest a mode of operation which is readily and rapidly
understandable. In practice, classifying intersections into major technical families (on the
basis of the shape and type of operation) means this requirement can be met satisfactorily
17
From the outset, two large groups of intersections can be identified:
• at-grade intersections in which all the exchanges between the roads take place on the
same plane. These are of two main types: standard at-grade intersections and roundabout
at-grade intersections.
This results directly from the principles set out in section 3.1. Grade-separated
intersections create a powerful impression of a road which is isolated from its environment;
there should be no at-grade intersections on these roads and, apart from a few exceptional
cases, grade-separated intersections should only be used on these roads. In contrast, at-
grade intersections are the normal type of design on roads which have a close connection
with their environment. The choice of the type of at-grade intersection (roundabout or
standard) depends above all on the volume and distribution of traffic and the number of
exchanges which exist or will be created on the roads involved.
"Type T roads are designed with the objective of providing a high level of service and favoring
long-distance traffic. Therefore, the design of intersections and access points should attempt
to limit disruption and speed loss caused by exchange movements between the road and the
secondary roads or the road's environment. It is therefore logical to prohibit frontage access
and grade-separate all points of exchange and crossings. The retention of a small number of
standard at-grade intersections is incompatible with safety. Roundabout intersections, which
are incompatible with the high level of service objective, can only be considered as a
provisional or permanent "terminal" option at the end of the road, in particular at the entrance
to a built-up area (the only other satisfactory arrangement at the end of such a road being a
"provisional end of freeway" design with the driver leaving via a deceleration lane and a
transition road before joining the ordinary road system)."
(Aménagement des Routes Principales)
18
If a section of existing road is converted into a type T road, the exchange and access
points must receive appropriate treatment: grade-separation (with or without exchange of
traffic) or removal of the intersection and transfer to a nearby intersection, removal of frontage
access, provision of access to and creation of a parallel road system (for traffic which is not
authorized to use the type T road).
Table 1 - Suitable alternatives for type T roads and the general conditions for their use.
"Type R roads carry long-distance traffic without favoring this over shorter-distance
traffic for which access and ease of exchange movements are important. Frontage access,
generally to non built-up land, can be allowed (except when safety criteria are not met, for
example when visibility is inadequate) and there should be a large number of exchange
points. The most appropriate intersections for these constraints are at-grade, either
roundabout or non-roundabout (4-way intersection or T-intersection). A grade-separated
intersection can be justified in exceptional circumstances (when a roundabout is saturated,
etc.), as can removing or displacing an intersection (in the event of safety problems
associated with its location). "
(Aménagement des Routes Principales)
19
Table 2 - Suitable alternatives for type R roads and the general conditions for their use.
In view of the specific problems affecting these roads as regards the safety of
standard at-grade intersections (the amount of crossing traffic, frequently unfavorable
location), the measures to be taken are as follows:
- no frontage access to ensure the road retains its primary purpose of carrying through-traffic
and to avoid urbanization which would result in an ambiguous identity.9
- standard at-grade intersections, with nearside priority or a "STOP" or "YIELD" priority rule.10
- roundabout intersections (generally small) for slightly larger intersections
9
In the case of a high-traffic road, this is compulsory (see article L152-1 of the "Code de la voirie routière").
10
The systematic (or quasi-systematic) giving of priority to a minor road is generally not recommended.
20
3.5. THE CHOICE BETWEEN TWO TYPES OF INTERSECTION
(WHICH ARE COMPATIBLE WITH THE TYPE OF ROAD)
The type of intersection for a given site must not be chosen without reference to the
general design policy for intersections along the route. 11 The consistency of the facilities
along a route is important and should be achieved immediately in the case of a new road and
taken into account, at least as a long term aim, in the case of existing roads.
On type R major roads, the choice of a type of intersection (from among the possible
options) is usually between the two types of at-grade intersections (standard or
roundabout).12 However, the choice of facility only takes this form when the other possible
solutions (removal of the intersection, grade separation with exchange, for example) have
been rejected.
The choice is based on local conditions at the site. It may be assisted by multicriteria
analysis, taking not only safety into account, but also user costs, financial analysis, cost
benefit appraisal, etc. 13
In the case of a new construction, the decision will frequently be based on general
knowledge about how a given type of intersection affects the number of accidents, delays etc.
At a site with an existing intersection, a safety diagnosis is an essential basis for decision-
making. In particular, there is no need to transform an intersection into a roundabout when no
accidents have occurred there.
The safety and journey time benefits depend mainly on the traffic on the major and
minor roads, in particular the nature of through and exchange traffic. This must be known in
order to make an informed choice.
a) Safety
This is a priority criterion. On a major road, a roundabout is safer than a standard at-
grade intersection, accidents are generally less frequent and less severe. Information
concerning the performance of the two types of at-grade intersection is given in Appendix 1.
21
b) Cost
The costs of at-grade intersections vary greatly depending on local conditions, the
extent to which the existing pavement is reutilized (in the case of reconstruction), the amount
of road furniture installed, the construction of feeder roads, etc. Some aspects of design
(illumination, landscaping, choice of materials, etc.) can considerably increase the cost of a
project. Furthermore, operating costs (maintenance, electricity consumption if applicable)
must also be considered.
c) Delays
This criterion is also important on roads that carry long distance or medium distance
traffic (which although rarely predominant can be deliberately favored). Local traffic must also
be considered on roads of secondary importance.
Delays are, basically, of two types, the relative importance of which depends on the
traffic at the site:
- traffic delay (also known as congestion delay). This is due to not having priority and
interactions between vehicles. It can be considered as being the time spent in a queue and at
the front of a queue.
- geometric delay. This is the delay experienced by a vehicle when crossing the facility,
when not impeded at all by traffic. This exists because an intersection forces some traffic
streams to slow down. 14
Delays are usually negligible for rural roundabouts. If this is not the case there is
probably a capacity problem which may be detected by the GIRABASE software.15 Figure 2
below illustrates the field of application for roundabout intersections with regard to traffic, in
particular their capacity limit.
14
In the case of a roundabout too, drivers need to negotiate the central reservation which makes their trajectory
slightly longer than a straight line, but the corresponding delay is small, at least for the dimensions
recommended in this document (see Chap. 3).
15
Produced by the CETE de l'Ouest. In addition, this also estimates delays.
22
Fig. 2 - Field of use of roundabout intersections with regard to traffic.
12 000
10 000
8 000
6 000
4 000
14 000
12 000
10 000
8 000
6 000
4 000
2 000
AADT on major
0 road
The light area represents a reserve capacity of more than 30%. Delays are generally
low.
The dark area represents a reserve capacity16 of between 10% and 30%. Delays can
become very long in certain cases
However, all the users passing through the roundabout are subjected to geometric
delay. The duration of this will vary according to the site. Its average value is 12s for light
vehicles (further information on geometric delay is given in Appendix 3).
16
See Appendix 2.
23
The delay experienced by non-priority drivers at a standard at-grade intersection is
generally greater than at a roundabout. It can be estimated using the OCTAVE software17 ,
but only become really high when demand on the minor road approaches capacity. Under
such circumstances, the traffic levels would normally justify the construction of a roundabout
intersection on safety grounds. Figure 1 in section 2.3.1. gives an approximate idea of the
capacity limit of a non-roundabout at-grade intersection.
Geometric delay mainly affects the minor road. It is of the same order of magnitude as
for roundabouts, but tends to be slightly longer (see Appendix 3).
- the need to mark the boundary between two types of roads (between a divided highway and
a single roadway, between a rural road and an urban or suburban road, etc.), to enable
drivers to identify the type of place they are entering and modify their behavior accordingly;
- a situation at the entry of a built-up zone: the "gateway" effect encourages a change in
driver behaviors and attention levels which are beneficial to safety in built-up areas; in addition
a roundabout provides a functional or aesthetic way of treating public space;
- a large number of legs (>4) which makes it difficult to design an acceptable standard at-
grade intersection;
- concerns about access, as a result of which turning movements can be facilitated to the
detriment of journey times on the road;
- more specifically, the need to provide opportunities for turning back, in particular on 3-lane
roads (where turning left towards accesses is always a difficult maneuver, particularly for
trucks) or on divided highways.
However, these criteria do not remove the need to consider the principle selection
criteria (see 3.5.2.). On their own they are not sufficient to justify the construction of a
roundabout. In particular, a roundabout is unnecessary when traffic demand on the minor
road is insignificant.
17
The OCTAVE software (SETRA, 1998) deals with the capacity of unsignalized intersections.
24
CHAPTER 2
STANDARD AT-GRADE INTERSECTIONS
u 1. GENERAL CONSIDERATIONS
1.1 DESIGN PRINCIPLES
4.3. ILLUMINATION
25
This chapter deals with the construction and geometric design of non-roundabout at-
grade intersections, known as standard at-grade intersections.
Standard at-grade intersections provide a lower average level of safety than other
types of intersection (roundabout, grade-separated). The priority in their design must therefore
be to maximize safety - capacity problems are comparatively rare in rural areas.
This chapter lays down guidelines to enable this objective to be achieved (while at the
same time adapting the facility to the type of traffic) and also states the rules and parameters
for constructing and sizing intersection components.
Minor leg
(non priority)
Exit lane
Island nose
Entry lane
Exit radius (R s)
Entry radius (Re)
Nez d’îlot
Major leg
Splitte risland ( priority)
Splitter
Left-turn lane
Through lane
Island tips
26
1. GENERAL CONSIDERATIONS
The conceptual and geometric design phases of standard at-grade intersections must
take into account the basic principles enumerated in Chapter 1. This essentially involves
compatibility with the type of road, integration with the rationale of the route, legibility of the
facility, optimization of safety, provision of a high rate of flow for priority flows and taking
account of specific types of road users.
In addition to these fundamental principles which apply to intersections of all types, the
construction and design of a standard at-grade intersection requires the following specific
precautions.
- special attention must be given to changes in the type of intersection or the priority rule along
a route.
- the geometry or the environment should be modified to encourage speeds which are
appropriate for the type of facility and the priority rule;
- visible and legible advance signing which informs drivers in the clearest possible terms of
the type of facility they are about to encounter and the priority rules which apply there
(directional and priority signing).
- the use of certain types of intersection which are compact, straightforward, tested and
rapidly identifiable and whose operation is well understood by drivers: T-intersections or 4-
way intersections (in addition to roundabouts);
- homogeneous geometrical features along a route; with facilities which comply as closely as
possible with the standard layouts described in the sections below;
27
- features which assist good legibility (it is generally only necessary to comply with the rule of
simplicity to achieve this);
- the simplest possible signing, which is consistent with the layout and placed where it is
clearly visible (particularly in the case of directional signing).
Fig. 2- Principal recommended and not recommended layouts for non-roundabout at-grade intersections.
1. Recommended
28
Several types of intersection must be rejected because they do not comply with these
principles and frequently cause accidents:
- "bulb" type intersections that are too large and rather complex, in which non priority traffic
flows too freely;
- Y-intersections which often have similar shortcomings and which suffer from a degree of
ambiguity; these should be replaced by either a T-intersection or a roundabout depending on
which is more appropriate.
- intersections in which there is a "left turning transition lane coming from the right", the
operation of which is ambiguous (left turning movements are unusual at at-grade
intersections);
- many other atypical types of intersection, frequently large, where the large number of islands
and transition lanes confuses drivers.
In addition, intersections with a nearside priority rule must not be allowed on type R
roads, because there is a risk they will not be understood by drivers on the major road, who
have generally had priority for long distances upstream of the intersection. Such intersections
are prohibited on roads which are classed as trunk roads.
Signalized intersections, which may surprise drivers and have a poor safety record,
should also be ruled out in rural areas. It is generally advantageous to replace them by
roundabout intersections, even in suburban areas or on cross-town routes.
• new roads: the principles set out above provide the basis for the design of planned
intersections;
• existing roads: priority should be given to facilities which improve safety, and to a lesser
degree capacity. Any plan to reconstruct an intersection should be preceded by accident
and traffic analysis (see Chap 1). The design rules given in the remainder of this Chapter
should be considered as general guidance for improvements to existing roads.
1.2.1. VISIBILITY
2
This includes maneuvers to cross the intersection and merge into the traffic on the major road.
29
Drivers turning left into the minor road must be provided with a similar length of time
with regard to opposing traffic on the major road.
The time required to cross the priority road, known as the "crossing time", naturally
depends on its width.
Table 1: Crossing time3 according to the width of the crossed road and the priority rule (to be taken into
account when calculating the visibility distance).
recommended time 8s 9s 8s
STOP
absolute minimum 6s 7s 6s
recommended time 10 s 11 s 9s
YIELD
absolute minimum 8s 9s 7s
recommended time 8s
Left-turn lane
into minor road absolute minimum 6s
N.B. These times should be increased by 1 s when access is via an upward grade of more than 2%,
which should, furthermore, be avoided (see 3.1.3.).
It must also be ensured that approaching vehicles have adequate visibility of the noses
of splitter islands on the major road and the secondary legs. This generally constitutes less of
a design constraint than the crossing time condition, and is set out in 2.1. for the major road.
b) Provision of visibility
This relates to the intersection crossing time and involves clearing a sight triangle for
each conflict between two traffic streams: there must be no visual obstruction within this
triangle.
The triangle is located 1 m above a plane which passes through the centerline of both
roads. Its corners are located as follows: (i) the conflict point between the two traffic streams
in question, (ii) an assumed observation point on the non-priority road beyond which a driver
must be able to see a vehicle traveling on the major road, and (iii) an observed point on the
major road. These elements vary depending on the priority rule (see Fig. 4 and 5).
3
The recommended times provide a greater margin of safety and are better suited to slow-starting vehicles (trucks, two-
wheel vehicles).
30
The assumed observation point is 2 m from the right edge of the non-priority road, set
back 4 m from the stop-line4 at a STOP-controlled intersection and set back 15 to 20 m from
the YIELD line5 where this rule applies. In the case of left turns into the minor road the position
of the assumed observation point is to be determined on a case by case basis depending on
the configuration of the intersection.
The observed point is at a height of 1 m above the centerline of the priority road,6
and at a distance from the conflict point which corresponds to the distance traveled by priority
vehicles during the crossing time (f); this distance is known as the crossing sight distance
(D). The 85th percentile speed (V85) is used to calculate D (see Appendix 4). 7
The distance D is given by: D = V85 x t, where the value of t is given below (Table 1)
and varies according the width of the road to be crossed and the priority rule.
Fig. 3 - Crossing sight distance (D) on the basis of the 85th percentile speed on the major road and the
crossing time (t).
400
t =11 s
350
t =10 s
300 t= 9s
t= 8s
250
t= 7s
200 t= 6s
150
100
Sight distance ( m) .
50
0
50 60 70 80 90 100 110 120
V85 on major road ( km/h)
Site triangles will be provided (4 for a 4-way intersections, 2 for a T-intersections, 1 for
a partial intersections on a divided highway) on the basis of the elements described above
and the priority rules which apply. The visibility for left turning maneuvers from the major road
will be checked later (this generally imposes fewer constraints than left-turning movements
from the secondary road).
4
This is the point where 95 % of drivers actually stop.
5
This is the position where road users gain information when enterring an intersection where the YIELD priority rule applies.
6
If the major road is a two-way road on which overtaking is allowed the lane in question is the left-hand lane (for traffic
coming from the right) for sight triangles to the right of the observation point. In all other cases it is the right-hand lane.
7
The V85 speed used to calculate the crossing sight distance is computed using the speeds of all vehicles, including those
exceeding the posted limit.
31
Fig. 4 - Sight triangles for drivers on the minor road, depending on the priority rule.
Fig 5 - Visibility for the left-turn maneuver into the minor road
32
These crossing sight distances will be used in geometric design and the management
of roadsides.
On a new road, intersections and access points must not be located on curved
alignment.8 It is, however, acceptable to install a T-intersection or an access point on the
outside of a bend whose radius is such that no more than the normal cross-fall is required, on
condition that sight distances are adequate. Right turning movements from the major road
must not be too tangential.
In the vicinity of an intersection, anything located near the road (signs or road
furniture9, slopes, trees, crops or other vegetation, buildings, engineering structures, walls,
parked vehicles, etc.) can potentially mask visibility. Thus, the visibility conditions set out
above demand a zone that is free of lateral masking with sufficient guarantees that it will
remain so. Exceptionally, localized masking may be tolerated, on condition that it does not
interfere with vision.
In order for traffic signs to be outside the sight triangles, they should be set back
roughly 200m from a YIELD line and at least 50 m from a STOP line. 10
e) The case of an existing facility where sight distance requirements are not
met
On an existing road, when it is not possible to remove the masking that impairs
visibility at an intersection, other measures must be considered. There are a number of ways
in which the visibility requirements stated above can be attained; we shall mention the
following:
- realigning the minor roads - for example, this can transform a 4-way intersection into two T-
intersections, known as a staggered intersection (see 3.1.2.) - in some cases with a small
radius in a salient angle, it can be advantageous to move the centerline of the minor (non-
priority) road to the central point of the curve;
- in exceptional cases, modifying the layout of the major road (horizontal alignment,
longitudinal profile);
8
Apart from the adverse effects on visibility, it is more difficult to judge speeds on a bend and information gathering is more
difficult when the non-priority branch joins the major road on the inside of a curve.
9
Safety barriers can also mask visibility, particularly when there is a summit curve on the major road.
10
If it is assumed that vehicles are 0.70 m away from the right shoulder (itself 2 m wide) and the 85th percentile speed is 100
km/h.
33
- in the case of partial intersections installed on 4-lane divided highways, installing or
extending a acceleration lane;
1.2.2. LEGIBILITY
Drivers arriving at an intersection must understand easily and rapidly how it operates ,
the behavior that is expected of them (for example slowing down and yielding) and what the
other drivers are doing or will do.
Finally, to ensure that the initial legibility provided when the facility is constructed has
some durability, the following should be considered from the design phase: foreseeable
changes in the use of surrounding land, the needs of residents, maintenance constraints, etc.
This means involving local partners (Commune, residents, etc.) in any project.
11
See ARP, Chap 2.
34
Fig. 6 - Sighting of the crossed minor road as a result of a transverse line.
35
1.2.3. THE DISTANCE BETWEEN INTERSECTIONS
The impact on the above factors depends on the type of intersection, operating
speeds (V85) on the route, etc. There is therefore no universally valid formal rule for the
minimum distance between two successive intersections. However, a minimum distance of
250 m can generally be considered to be satisfactory12 , but specific site characteristics can
make this very inadequate.
The presence of a center lane can also lead to a minimum distance being
recommended between two successive intersections, in order to allow sufficient opportunity
for safe overtaking. In the case of an existing road, however, this recommendation should not
mean that the possibility of a central facility is rejected if it is necessary for the safety of the
intersection.
Table 2 - Minimum recommended distance between successive intersections, and residual length for
overtaking13, depending on operating speeds.
- for safety, as it limits the number of potential conflict zones, and grouping existing
intersections together makes it possible to retain only those with the required approach and
crossing sight distances;
- for traffic flow and the comfort of drivers on the major road, as the number of zones where
flow is disturbed by exchanges between roads is limited and there is more opportunity to
overtake slow vehicles.
12
"Staggered" intersections should be considered as a single facility rather than two T-intersections close together.
13
This is a theoretical distance which may be reduced by other factors (for example the layout).
36
2. DESIGN OF THE PRIORITY ROAD
The general rules for the installation of intersections are given in section 1.2. of this
Chapter.
It is particularly important for drivers traveling on the priority road to have an adequate
sight distance with respect to the noses of raised traffic islands (the so-called "entry approach
sight distance"). This sight distance must be at least equal to the stopping distance14 for a
vehicle traveling at the V85 speed of the approaching traffic on the major road.15 For new
roads this condition will usually be met at all points along the route. For improvements to
intersections on existing roads, the table below shows the stopping distances used for design
purposes.
Generally, the wider the major road, the more dangerous the intersection. This means
that at the intersection the number of through lanes should be limited in both directions.
At a standard at-grade intersection, only one through lane should be retained in each
direction. Widening to 3 or 4 lanes at an intersection must therefore not occur: such
configurations are dangerous and result in little improvement in traffic flow on the major road.
14
The stopping distance d is made up of the braking distance (distanced covered during braking which reduces the speed
from V85 to 0 under specified wet pavement conditions) and the distance covered during the reaction time (taken as 2 s at
speeds of 100 km/h and less). In order to calculate the stopping distance d, readers are referred to § 4.2.b. in the A.R.P.
15
It is also possible for only speeds below the speed limit (typically 90 km/h) to be used for calculating the 85th percentile
speed.
16
The stopping distance on a bend should be used for radii R< 5 x V85 (where V 85 is in m/s and R is the radius of the bend).
37
2.2.2. INTERSECTIONS ON 3-LANE (AND EXCEPTIONALLY 4-
LANE) ROADS
It is strongly recommended that the 3-lane (and even more the four-lane) configuration
be abandoned at intersections. On such road links traffic should be reduced to a single lane in
each direction upstream of the intersection.
The guidance given below on the choice of facilities for left turning movements (lateral
widening, central facility, construction of a roundabout), are essentially the outcome of safety
considerations, but also take account of driving comfort and the cost of facilities. The traffic
levels which are stated should not be taken rigidly but considered in relation to local
constraints and any specific traffic peaks.
Furthermore, on an existing road, detailed accident analysis will reveal the relative
proportion of accidents involving left turns and criss-crossing movements and evaluate
whether it is appropriate to install a left-turning facility and what form this should take.
17
Opportunities for turning back must be frequent, about 5 km apart to avoid lengthening jounrneys excessively (this distance
should, however, depend on real needs and whether or not access roads have been constructed).
38
2.3.1. ON A 2-LANE ROAD
The scale of design measures at a standard at-grade intersection on a 2-lane road will
depend on the type of intersection (T-intersection or 4-way intersection)18 and the traffic
levels.
For T-intersections with low left-turning demand (less than 100 v/d) or for resident
access points, constructing a shoulder opposite the minor road (or access point), reduces
the risk of accidents caused by the left-turning movement (by providing for an avoiding
maneuver to the right of the turning vehicle). In the absence of a continuous (and sufficiently
wide) shoulder along the major road, local treatment as shown in the diagram below (Fig. 11)
is satisfactory.
≥5 m
10 m 15 to 35 m 10 m 10 m 1,50 to 2,00 m
The aim should be to achieve a width of at least 5 m between the centerline of the
road and the edge of the widened section. This generally involves widening the road by
between 1.50 m and 2.00 m.
A total length of less than 40 m should be avoided (a short facility of this type is likely
to encourage hesitation and impede the avoidance maneuver). The total length may be
increased to 65 m when there are trucks turning left. Nothing should be done which might
deter drivers from using the shoulder for an avoidance maneuver: the surface must be of
good quality, without special markings (normal edge marking highlights the boundary between
the roadway and the shoulder, etc.).
However, use of the shoulder as a traffic lane should not be over encouraged: the
shoulder should not be treated as a lane (for example with marking on its outer edge, or a
long entry taper), and any signing which attempts to specify a particular use of the shoulder
should be avoided.
N.B. Parked vehicles do not generally cause a problem, as parking demand tends to be very low in rural
areas. Where problems arise a prohibitory sign may be installed (off the usable part of the shoulder).
Very occasional parking does not reduce the usefulness of this facility.
18
Staggered intersections are covered in 3.1.2.
39
b) creation of a left-turn lane
For 4-way intersections, the construction of this type of lane reduces the accident
risk for left turns but increases the risk of a collision in the second part of the maneuver
between a vehicle from the minor road crossing the major road and a priority vehicle. A facility
of this type is therefore only recommended when left-turning demand is quite high (more than
200 v/d), and when the traffic crossing the intersection is not too heavy. The geometric
characteristics of left-turn lanes are given in 2.5.
On safety grounds, the center lane must be systematically be closed at all at-grade
intersections and resident access points into which left turns are possible. The central space
should be converted into a (perhaps short) left-turn lane or a central refuge in the case of an
access point. This means that the number of access points must be restricted in order to
maintain a sufficient number of zones where the 3-lanes can be used to increase overtaking
capacity.
40
2.3.3. CRITERIA FOR ACHIEVING THE BEST COMPROMISE
Table 4 - General rules for facilities for left-turn movements from the major road.
2-lane roads
no change or
< 8000 v/d no change left-turn lane
surfacing shoulder left-turn lane
or same or or
> 8000 v/d left-turn
surfacing of shoulder roundabout
lane
3-lane roads
left-turn lane left-turn lane
< 8000 v/d or or left-turn lane left-turn lane
removal of access point removal of intersection with left-turn lane or
> 8000 v/d (and provision of new transfer to an adjacent or roundabout roundabout
access road dans le cas intersection
d’un créneau de
dépassement)
2. For a 4-way intersection
2-lane roads
3 lane roads
left-turn lane
< 8000 v/d or left-turn lane left-turn lane
41
2.4. CROSS-SECTION WHERE THERE IS A CENTRAL FACILITY
ON THE PRIORITY ROAD
Traffic lanes should generally have the same width within an at-grade intersection as
in the links between intersections. However, if the roadway's typical cross-section is between
5 and 6 m wide, the traffic lanes at the intersection will be 3 m wide. Furthermore, at an
intersection there is no justification for having through lanes wider than 3.50 m.
The central island should be separated from the continuous line (3 u wide)19 that runs
adjacent to it by an unpainted area with a width of at least 2 u: the shoulder on the left will
therefore have a minimum width of 5 u, but a considerably greater width is preferable (0.50 m
for a road with a width of 6 m or more wide, for example).
For a roadway whose typical cross-section is 7 m wide, this would give, for each
traffic direction, a minimum paved width between the edge of the central island and the line at
the outside edge of the pavement of 3.80 m, which would break down as follows: 3.50 m of
traffic lane + continuous line of 3 u + space of 2 u (minimum), where u = 6 cm.
The right shoulder (which carries the edge marking) has the same width as it does in
the road's typical cross-section (with, however, a minimum of 1 m). When there is a high
level of light two-wheel traffic, it is preferable to surface a 1.25 m width of shoulder.20
Fig.12 - Cross-section at a standard at-grade intersection as a function of the width of the typical
cross-section.
1. Width of between 5 and 6 m (typical cross-section)
"LC" "T’3"
(3 u ) (3 u )
0,30 à ≥ 1,00 m
0,40 m 3,00 m
0,30 à ≥ 1,00 m‚
3,00 to 3,50
0,50 m
m‚
left shoulder carriageway ( 1 through lane) right shoulder•
7,00 m ≥ 1,00 m‚
¬ A minimum 1.25 m width of the right shoulder must be paved is if there is a large amount of two-wheel
traffic.
- The lane and the right shoulder should be at least as wide as in the typical cross-section of the road.
® The left shoulder and the median strip are the same width as in the typical cross-section of the road
19
u is the width unit (see 4.1.6).
20
These features are to be considered is minima and do not remove the need to comply with the rules set out in the ARP.
42
2.5. SPLITTER ISLANDS (OR CENTRAL RESERVATION)
The main function of splitter islands on the major road is to protect vehicles turning left
from rear-end collisions by physically separating the left-turn lane. Correctly designed splitter
islands (in particular with advance signing of the island nose by means of stripes as laid down
in the regulations) improves overall perception of the intersection for priority traffic (warning
effect generated by introducing a clearly identifiable image into the "roadscape".
Raised splitter islands on the major road are essential in the following situations:
- on all types of roads when there is a left-turn lane, except in the specific case of narrow
roads (see 2.8);
- on 3-lane roads when there is a considerable amount of crossing traffic (<100 v/d);
- on 4-lane roads, in order to reduce each roadbed to one lane upstream of the intersection.
2.5.2 DISPLACEMENT
A splitter island should always cause the roadbed leading to the intersection to be
displaced to the right. For reasons of legibility, it is preferable for displacement to be
symmetrical with respect to the centerline of the priority road. Clearly apparent displacement,
which creates some visual constraints, should also be preferred to a gradual change. A
displacement21 of 1/15th is satisfactory as far as safety is concerned.22
The length of the displaced zone can be reduced, but the advance signing of the
island nose by means the symmetrical marking that widens upstream of the island must have
a length of L/2 for the island to be clearly perceived.
21
The displacement (or "inclination of the island") is the angle formed between the curb of the island between its nose and its
widest point and the centreline of the road upstream of the island's nose.
22
On narrow roads (≤ 5 m) a displacement of 1/10th can be considered.
43
Fig. 13 - Geometric design criteria for theapproach to splitter islands on major roads with reference to the
typical cross-section of the road.
Table 5 - Values of L (length of advance signing; see "Instruction Ministérielle sur la signalisation
routière; Livre 1 - Partie 7.".
A 5 m wide splitter island is always sufficient for installing a left-turn lane. Greater
width has a number of disadvantages: it leads to a marked deviation in the paths of priority
vehicles and can be confused with an island on a minor road, or even an island at a
roundabout entry; under such circumstances the priority rule is unclear for both priority
vehicles and left turning vehicles, which are provided with very good conditions of flow. In
addition it increases the width of the intersection crossing, and therefore increases the risk of
a criss-crossing accident. The width of the central facility should therefore be limited to what
is strictly necessary in order to install a left-turn lane; reducing the width of a left-turn lane
does not reduce safety.
Table 6 below shows the recommended range of widths for the principal scenarios.
44
2.5.4. LENGTH OF SPLITTER ISLANDS
The length of the left-turn lane (without taper) does not affect safety, and can therefore
be limited to the length required to store left-turning vehicles (generally quite short), and
occasionally reduced to just a central refuge. Other factors may, however, impose greater
lengths (for example the guaranteed level of comfort along a route, or the choice of a large
capacity reserve).
The total length (for one traffic direction) of a splitter island (including its advance
marking) will be determined by the length of the displacement, the tapered lane and the left-
turn lane. It can vary between 90 m in the case of intersections with low volumes of left-
turning traffic to more than 170 m in the case of the largest intersections. The length of the
advance marking of the splitter island nose will always be equal to L/2 when the lanes are
displaced symmetrically with regard to the centerline of the road and equal to L in the case of
unilateral displacement.
Table 6 - Length (in m) of the components of splitter islands, in the principal scenarios (with
symmetrical displacement, depending on the width of the typical cross-section of the road and
the composition of left turning traffic.
¬ a is the width of the part of the island between the left-turn lane and the through lane for opposing
traffic
- b is the total width of the central island, i.e. the sum of a and the width of the left-turn lane.
45
The nomograph below (Fig. 14) shows the number of vehicles that should be
considered when designing the storage zone. The storage length is obtained by simply
considering the average space that vehicles occupy. The following formula, for example,
could be used : Ls = (7 + 10p).Ns (where Ls is in meters, p is the proportion of trucks in the the
stream in question, and Ns is the number of vehicles obtained from the nomograph. 23
Fig 14 - The storage capacity to be provided (number of vehicles) on the left-turn lane (values given by
the OCTAVE software which deals with the capacity of unsignalized intersections, see Appendix 2).
10 12
400 8
6
350
300 5
250 4
200
3
150
100
2
Left turning traffic (v/h) .
50
0
0 100 200 300 400 500 600 700 800 900 1000 1100 1200
The splitter islands on the major road must (i) assist vehicle guidance (to obtain an
optimum path), in particular in the case of left-turning movements (from the minor road into
the major road or vice-versa) and (ii) not be located in the path of straight ahead or left-turning
vehicles on the minor road. In practice, the position of the island heads is given by the
tangency point of the splitter island turning circles when exiting the minor road and from the
left edge of the relevant traffic lanes on the major road (see 3.2 and Fig. 15).
Fig 15 - the central part of the splitter island in the case of a 5 m central reservation on a road with a 7 m
wide typical cross-section.
Taper : 20 m to 30 m Storage : 20 m to 50 m R is
* 0,30 mini
23
This formula can be modified to take account of specific sizes of turning vehicles.
46
2.5.6. DETAILED DESIGN OF SPLITTER ISLANDS
It is always preferable for raised islands to have mountable curbs (see Appendix 6
"Design of the curbs"). This is important because the island markings become less visible in
wet conditions, and, also, some drivers do not obey continuous lines).
Islands must be free from aggressive obstacles (road lighting columns, large sign
poles, etc.). All singing (priority, regulatory, or directional signs, J5 markers) which is normally
located on splitter islands, must be installed with a minimum distance of 0.70 m between the
edge of the sign and the edge of the nearest lane.
Raised islands must be constructed from materials with a different surface from the
roadway or which are painted uniformly (no striping) in order to contrast with the roadway,
both during the day and at night. A mineral surface on an island is therefore much preferable
to grass, not only because the latter does not comply with the above requirement but also
because of maintenance related factors.
If the surface of the island is permeable, the necessary drainage facilities should be
provided, particularly when the island has embedded curbs.
If it is not possible to construct raised islands, the use of color can provide a solution.
The coloring should dissuade vehicles from crossing the part of the roadway it delineates and
improve the overall legibility of the intersection (as compared with the use of painted islands,
for example).
When there are very strong constraints which apply to the land on which the facility is
located, the part of the island which separates the storage lane for vehicles turning left and
the opposing through lane can simply be painted (5 u continuous line).
The creation of "right-turn" deceleration lanes does not generally improve safety at an
at-grade intersection.24 Furthermore, such lanes can indirectly reduce safety.25 The general
rule is, therefore, not to provide right-turn lanes at at-grade intersections on undivided type-R
roads: the transition lane for the right turning movement from the major road consists of a
circular arc (see 3.3). In those cases where the right turning movement must be favored the
turning radius will be considerably increased (by about 25%).
24
Accidents which involve right-turning movements are always rare and less serious than accidents of other types.
25
On the grounds that a wide road encourages higher speeds at an intersection, joining the minor road at excessive speed
or moving masking created by certain vehicles travelling on the highway.
47
2.6.2. THE CASE OF PARTIAL INTERSECTIONS ON DIVIDED
HIGHWAYS
- a straight exit taper 80 m long (between the point where the striping that marks the exit
island narrows to1 m and the start of the taper) and 4 m wide at the end;
2m R=25 m
Lb=80 m
Ld≈25 m
4m
1,50 m 1m
When exiting traffic demand is low, the transition for the right-turning movement from
the major road will consist of a ciscular arc with a radius of 25 m, usually preceded by a
clothoid with a length of approximately 25 m.
26
The change in direction imposed by the circular arc must be at least half the total change in direction (a) between the exit
of the taper and the secondary road; the general condition to be satisfied is Lcl ≥ 0.44 x a (where Lcl is the length of the
clothoid and a is in degrees).
48
2.7. ACCELERATION LANES ON A MAJOR ROAD
a) On an undivided road
For standard at-grade intersections, the improvement in comfort and travel time is
never sufficient to justify provision of a right side acceleration lane. A cost-benefit analysis for
such a facility would only be positive with flow levels above those which would easily justify
creation of a roundabout intersection. Furthermore, in the absence of a central facility on the
priority road, right side acceleration lanes should be forbidden as they can create ambiguity in
the road as perceived by drivers.
b) On a divided highway
Right side acceleration lanes are therefore only an option for intersections installed on
divided highways (type R); these intersections are designed as partial intersections (see
Chap.1.). Saturation which is partly due to delay resulting from the loss of priority, or a visibility
problem due to a particular type of configuration can justify the creation of a right side
acceleration lane. The geometry for the right turn movement when no acceleration lane is
provided is given in 3.5.
The acceleration lane must be parallel to the major road and short. It allows vehicles
to join the major road with a low angle of conflict, or to stop when the non-priority vehicle does
not encounter an acceptable gap in the traffic.
To achieve the greatest possible uniformity between facilities, adhering to the diagram
below is recommended.
Fig 17 - Design of partial intersections with right side acceleration lane on a four-lane divided type R
road); this example shows a minor road with a 6 m wide typical cross-section.
6m
r=200 m
r=300 m
≥30 m L/2=39 m
le=4m 2m
Re=25 m
70 m 40 m
2m
1,5 m
3,5 m
1m
Central reservation
49
2.7.2. LEFT SIDE ACCELERATION LANE
Generally, a left side acceleration lane should not be constructed to cater for vehicle
movements turning left from the minor road into the major road. Such vehicles find it difficult
to assess the opportunities for merging (an obvious difficulty due to the need for visibility
behind the vehicle), and the fact that the priority rule cannot be displayed, etc.
An arrangement of this type can only be considered for certain T-intersections when
there are obvious merging difficulties due to heavy traffic on the major road, or for minor
intersections which service particular situations (for example, works exits where most of the
traffic merging into the major road consists of trucks).
If need be, this left-turn lane should be considered as a refuge allowing crossing in two
stages rather than as an acceleration lane. In addition, neither the simultaneous storage of
several vehicles (which could impede each other), nor overtaking on the major road in the
vicinity of the intersection should be encouraged. The left-turn lane should therefore be as
short as possible - a length of 30 m with a taper of 30 to 40m may be adequate. This
arrangement, which should only be considered when there is a central facility, lengthens the
splitter island on the major road by an equivalent amount.
- where there is displacement this should be as little as possible (with a maximum angle of
1/10th).
- the total width of the cross-section at the intersection should be at least 7 m (two 2.50 m
wide traffic lanes and a 2 m wide left-turn lane)27
27
However, if trucks will be making the left turn a minimum width of 2.0 m is required.
50
Fig. 18 - Minimum characteristics for intersection design on secondary roads with a narrow roadway (less
than 5 m).
R=7 m
2,5 m
2 ou 2,50 m27
L/6 L/6
L=39 m (mini) 6,5m 10 m 10 m 10 m 10 m 5 m 10 m 6,5m L=39 m (mini)
Fig 19 Detailed features of painted islands at intersections on secondary roads with narrow roadways
(less than 5 m).
2,50 m to 3,00
Right
Central island Pavement shoulder*
* At least 1 m of the right shoulder must be paved if there is a large amount of two-wheel traffic.
51
3. DESIGN OF THE NON-PRIORITY ROAD
Whenever possible, the alignment of the non priority road should be such that its angle
of skew (θ)28 with the priority road is as small as possible.
If the angle of skew of the secondary road deviates significantly (more than 20°) from
the perpendicular of the major road, this must be rectified by realignment. Under these
circumstances, one of the solutions below should be selected, depending on the constraints
but bearing in mind the order of efficacy, which is given in decreasing order from 1 to 3 on the
basis of the resulting approach conditions, cost, feasibility etc.
N.B. the solutions numbered (3) transform a 4-way intersection intersection into a staggered intersection
and replace through movements by a right-turning movement onto followed by a left-turning movement off
the major road (see 3.1.2.).
28
The angle of skew is measured with respect to the perpendicular of the axis of the major road, and is therfore zero for a
perpendicualr intersection.
52
3.1.2. PLACEMENT OF STAGGERED INTERSECTIONS
There are two configurations of staggered intersections, depending on the order in
which the major road intercepts the legs of the secondary road: a right-turning movement
onto followed by a left-turning movement off the major road and the other way round. The first
configuration is the only one that assists the continuity of the secondary road link, and thus
the only one that should be considered;29 there is no benefit in installing the second type over
retaining a 4-way intersection.
Table 7 - Distance between centerlines depending on the width of the typical cross-section of the major
road.
Width of the typical cross-section
of the major road >6m ≥6m
* Longer staggered intersections can be constructed when two T-intersections or a 4-way intersection do
not meet the distance conditions set out in 1.2.3.
Fig. 21 - Distance between the centerlines of the two minor legs of a staggered intersection.
Before it joins the major road, the longitudinal profile of the non-priority road should
contain a zone with a sufficient length (10 to 20 m) with a low grade (maximum 2%) in order
to make it easier for vehicles to start and considerably improve visibility of the priority road. On
topographically difficult roads, 30 constructing a short level section may be considered (at
least 5 or 6 m).
Steep downgrades must be avoided where a minor road intersects with a major one,
particularly when the former is used by trucks.
29
This configuration has the advantage that vehicles cross the intersection in two stages, the first movement being a right-
turn (which is generally safe) and the second a left turn which is protected by a central facility. However, nothing proves
that it is better than a properly designed crossroads.
30
"A section of road can only be considered as a "topographically difficult road" if problems occur continuously or frequently
over a distance of at least 10 kilometres. Local difficulties should not cause a road to be considered as belonging to this
category." (A.R.P.)
53
3.2. SPLITTER ISLANDS ON THE NON-PRIORITY ROAD
A raised splitter island will help improve safety for crossing movements and merging
movements into the main traffic stream from the left. An island of this type constitutes the
main feature on each leg of the non priority road.
- to interrupt the continuity of the non-priority road so as to warn approaching traffic (in other
words, to improve driver perception of the intersection and loss of priority);
- to play a role in slowing considerably (or stopping) the non-priority traffic stream, so it
complies with the priority signing;
- to guide vehicles in the non-priority road towards an optimum position at the STOP or YIELD
line;
- to guide vehicles leaving the major road to join the minor road.
The shape the splitter island on the secondary road is determined by these functions.
In particular, it should apply a considerable constraint on vision and vehicle path.
54
3.2.2. GENERAL DESIGN PRINCIPLES
Drivers should be able to identify easily the various types of facility they encounter and
the manner in which exchanges operate. It is therefore advisable to introduce some
standardization of the components of standard at-grade intersections, in particular the splitter
islands on the non-priority road, in view of their vital function.
Conventionally, the size of the island is defined as the space between the lanes
carrying each movement, i.e. the so-called "envelope" of the painted island. The island thus
includes the edge markings of the lanes it separates. Its dimensions depend on the
characteristics of the cross-section of the non-priority road and the specific features of the
major road. However, the constructional arrangements are identical for STOP and YIELD
signs (decisions based on the crossing visibility).
Marking
Curb Centre of island Curb
Fig. 23 - Effect of the angle of skew of the minor road on the general shape of the splitter island.
55
It is not advisable to place channelization islands beside the splitter island, except if
there are deceleration or acceleration lanes (which are, furthermore, generally discouraged
on undivided type R roads).
On extremely minor secondary roads, splitter islands still have an important safety
function. However, in order to make cost savings,31 very much smaller islands can be
constructed (see 3.4.). Nevertheless, it must not be forgotten that a certain minimum width is
required in order to install any new signing: a 500 mm J5 marker cannot be installed on an
island which is less than 1.90 m wide. In the light of this, raised islands that are smaller than
standard islands (defined below) are acceptable, so long as the minimum dimensions are
exceeded: below these minimum dimensions a painted island or a traversible island must be
constructed (no traffic signs can be installed on these).
To be effective (as regards perception and compliance etc.), the island must be raised
and surrounded by low beveled curbs whose shape renders them less aggressive (see
Appendix 6).
It must not contain any aggressive obstacle (lighting columns, small walls, large sign
poles, etc.). All signing (priority, regulatory, or directional signs, J5 markers) which is normally
located on splitter islands must be installed with a minimum distance of 0.70 m between the
edge of the sign and the edge of the nearest lane and also in a manner which does not impair
visibility on the minor road.
Raised islands must be constructed from materials with a different surface from the
roadway or which are painted uniformly (no stripes) to create a contrast with the roadway by
both day and night. A mineral surface on an island is therefore much preferable to grass, not
only because the latter does not satisfy this objective but also because of maintenance
related factors.
If the surface of the island is permeable, the necessary drainage facilities should be
provided, particularly when the island has embedded curbs.
31
Intersections of this type considerably outnumber large intersections.
56
3.2.4. GEOMETRIC CONSTRUCTION OF SPLITTER ISLANDS
a) Construction triangle
The axis of the triangle is parallel to the centerline of the secondary road. It is offset
0.50 m from the centerline of the pavement, in order to be aligned with the center of the island
nose where a J5 marker is installed. In addition, the base of the triangle is aligned with the
edge of the roadway of the major road.
L/2
b1
b2
H = 2× B
The diagram below (Fig 25) shows how a splitter island should be designed on the
basis of the construction triangle. It should be noted that:
- the left-turn radii at the entry(Rie ) and exit (Ris) depend on the width (l) of the minor road, but
also on the width of a central facility (if any) on the major road;
- the tip of the island will be set back 1 m from the right edge marking of the major road.
The design of the raised island will be based on this drawing, with a set back of at
least 5 u (preferably 0.50 m), except that the part which is tangential to the right edge of the
roadway of the major road which will be set back 1.50 m from this edge.
57
Fig 25. Splitter island and minor road entry and exit lanes (see Table 8).
l
s
le Rs
Re J5
R is
R ie
Rre Rrs
l
l s R ie
e
R is
rn=1m
Rs
Re
1,50m
a
b
58
Table 8 shows the parameters for splitter island design, with reference to the
approach speed, the characteristics of the priority road and the width of the roadway of the
non-priority road (l). The input parameter for this table is the width l. To simplify matters, it has
been assumed that the approach speed on the minor road depends on this width. An advance
signing distance Lp can therefore be given for each shape of island (see 4.1.6.).
Table 8 - Summary of the main construction parameters for a splitter island and the entry and exit lanes
on the minor road (values in meters).
Notation Parameters Typical values
Island parameters
Height of construction triangle H 4l 20 24 28
Radius of island to the left of the exit Ris 2l + a 10+a 12+a 14+a
Radius of island to the left of the entry (*) Rie 2l + b 10+b 12+b 14+b
* where entry radii are small, it is necessary to ascertain that trucks on the minor road have appropriate turning
conditions (turning movement template) and take any necessary measures to assist them (in particular a wider
paved area, or moving the raised splitter island nearer the major road).
32
However, a standard at-grade intersection should not normally be installed at the meeting point between two roads whose
link sections are 7 m (or more) wide. A roundabout will normally be constructed at the intesection between two major
roads occupying the same position in the road hierarchy.
59
The splitter island marks the farside (left edge) of the entry and exit lanes. The position
of the outside edges will be established by drawing a line parallel to the inside edge. The
transition to the major road which widens the end of the lanes is the arc of a circle that is
tangential to this line and the right edge of the major road.
Where approach speeds on the major road are high, the exit radius (Rs) can be
introduced gradually (using a clothoid) with a length Lcl = 6 Rs0.4.
The geometry of the exit of the minor road must not allow side-by-side storage of
vehicles at the STOP or YIELD line, as this would be detrimental to visibility.
The entry lane should be the same width as lanes in the typical cross-section of the
road (l/2), with a minimum value of 3 m. Where the width l exceeds 5 m, it is strongly
recommended to make the transition to 5 m gradually, by a lateral displacement of one
thirtieth the longitudinal distance" from the beginning of the island approach (which could be
reduced to a minimum of 20 m).
Small islands are to be reserved for extremely secondary non-priority roads (whose
roadway is stricly limited to 5 m). Their characteristics do not meet the turning circle
requirements for all types of public transport vehicles and trucks.
Such roads will have a 1 raised splitter island which is 1.5 m wide and 4 m long,
inclined at an angle of 10° (to the centerline of the secondary road) to make it more
perceptible, and set back 3 m from the major road. It will be designed so as to be occasionally
traversible by vehicles with large turning circles and must therefore be free from signs or
markers.
10°
20 m
3m
1,50 m
4m
3m R s=15
e=7,50 m
3m
60
3.5. PARTIAL INTERSECTIONS ON DIVIDED HIGHWAYS
The situation regarding speed change lanes on divided highways is dealt with in 2.6.2.
and 2.7.2. § b.
In addition to the functions described in 3.2.1., in partial intersections the splitter island
on the secondary road must also reduce the danger of "wrong-way" entry onto the divided
priority road. However, the desire to achieve this should not give rise to excessively tangential
paths which hamper information gathering and perception of loss of priority.
The construction principles for the splitter island are similar to those at a conventional
T-intersection, except for the tip of the island $this case the island has a $concave tip p61. In
practice the curves which widen the island are parallel to the outer edges of the
corresponding roadway - these are circular arcs which are tangential to the edge of the major
road with a radius of 2 l (minimum 12 m) for the entry and 25 m for the exit.
The diagram below shows the construction principles for the splitter island at a partial
intersection using the construction triangle as a basis. The following points should be noted:
- the noses of the raised islands are rounded (radius between 0.50 and 1.50 m) and set well
back from the points of divergence and convergence;
- the raised island is set back 1.50 m from the right edge of the major road;
- the exit lane is 4 m wide; to the right of this is a 2 m paved shoulder. The width of this
shoulder is rapidly reduced to the normal width for a secondary road, in practice at the point
where the alignment becomes straight;
- the entry lane has the same width as on the typical cross-section (with a minimum value of
3 m), but widens to 4 m at its end.
61
Fig 27 - Standard layout for a splitter island and the minor road entry/exit lanes at a partial intersection.
Rrs
Rre
ls=4m
le
0,50
2m
Re Rs*=25 m
1,50 m
médian
* The transition to the circular arc can be made by a clothoid approximately 25 m in length.
62
4. ROAD FURNITURE AND SIGNING
In the case of both existing roads and new roads, traffic management devices
(signing, restraint devices, road furniture, etc.) should be considered at the earliest possible
stage of project design to ensure the layout is compatible with requirements and the
performance of the equipment.33
The main considerations that apply to road furniture and signing at standard at-grade
intersections are stated below; more detailed explanations (concerning nomenclature,
conditions for use, etc.) are to be found in instructions, circulars and other technical
documents (Guide des Equipements des Routes Interurbaines, etc.).
4.1. SIGNING
Traffic signs must be installed with a minimum distance of 0.70 m between the edge
of the sign and the right shoulder or, in the case of signs on raised islands, the edge of the
nearest lane.
All signing (priority signs, regulatory, or directional signs, J5 markers) which is within
the intersection or on roadsides or splitter islands in its approach must be installed in a
manner which does not impair visibility. To ensure that the signs on the major road are
outside the sight triangles they must be set back roughly 200m at a YIELD sign and at least
50 m at a STOP sign.34
The size of islands should take into account the constraints affecting the installation of
signs and the rules that apply to road markings. In particular, a minimum width is required in
order to install a given sign: a 500 mm J5 marker cannot be installed on an island less than
1.90 m wide, for example.
33
The interdependence between layout and road furniture at or in the vicinity of intersections frequently means
that design work on signing (horizontal, vertical and directional) and road furniture must be conducted at the
same time as general design work on layout.
34
Where they are set back 0.70 from the edge of the right shoulder (which is itself 2,00 m wide) and the 85th
percentile speed is 100 km/h.
63
4.1.2 PRIORITY SIGNING
A STOP or YIELD priority rule applies at non-roundabout intersections on interurban
sections of major roads as the nearside priority rule and traffic signals are to be avoided in
rural areas.
The choice between the STOP or YIELD priority rule depends on the visibility
conditions on the minor legs of the intersection (see 1.2.1).35 The STOP rule is the most
constraining for drivers on the minor road who must halt for a time. Also, in order for the
STOP sign not to become discredited, the YIELD rule must be that which applies generally. In
particular, when the visibility conditions for a YIELD sign are satisfied it would be unjustifiable
to impose a STOP rule.
On roads which are classed as trunk roads, an AB6 sign (PRIORITY ROAD) should
not be installed systematically after each intersection, but nevertheless one should be
installed after a large intersection (with large numbers of exchange movements and therefore,
normally, a central facility). When the major road does not have trunk road status an AB2
(LOCAL PRIORITY) sign is compulsory upstream of the intersection.
The AB3a or AB4 signs must be installed "as close as possible to the transverse line",
on condition that this does not mean they become visibible too late. Thus, in view of the
widening of the minor road when it joins the major road it may occasionally be helpful to move
them a few meters forward. Furthermore, these signs should not repeated to the left on
splitter islands: this would be superfluous and could adversely affect the visibility of drivers on
the minor road.
35
A different priority rule may apply on different minor legs of a given intersection.
36
In the specific case of a partial intersection on a divided highway, recommendatory signing can also include B2a and B2b
signs.
64
4.1.5 MARKERS
A J5 marker must be placed in the center of the nose of raised islands on both major
and minor roads. The regulations do not make this compulsory in the strict sense, but it is
recommended to install one systematically because it improves perception of the island.
In order not to interfere with the visibility of left turning vehicles, the J5 markers
installed on the tips of splitter islands on the major road should not be more than 1 m in
height.
J3 markers may be of value when the intersection is not sufficiently perceptible to
drivers on the major road. However, there is normally no need to install one when design
measures have already been implemented at the intersection (splitter island on major road,
etc.).
Fig. 1 Location of road markings (the case shown is a minor road with marking between
intersections).
Lp
V85
T3 (3u) LC (3u)
T’3 (3u)
T2 (5u)
37
u is the width unit. It differs according to the type of road: 7.5 cm for divided highways, 6 cm for major roads, 5 cm for all
other roads.
65
4.2. SAFETY BARRIERS (RESTRAINING DEVICES)38
Safety barriers are obstacles in themselves (their supports and rails are particularly
aggressive for the riders of both light and heavy two-wheel vehicles.
Particular attention needs to be paid to the ends of safety barriers (highly aggressive
"fish tail" or "quarter circle" arrangements are prohibited). The rails should be lowered or
buried in an embankment, in compliance with the instruction concerning the approval and
utilization of restraint devices. When lowering the barrier along the leg on which it is installed
means that certain obstacles or hazardous configurations cannot be protected, the barrier
may be extended to the adjacent leg using a large radius (> 10 m) circular arc; the barrier will
then be joined to the next leg, or if there is none lowered a few tens of metres further down the
branch. Readers should nevertheless be aware that this arrangement is not really
satisfactory.
4.3 ILLUMINATION
In general, roads in rural areas are not illuminated, and this applies to their
intersections to, even when these have raised islands (retroreflective curbs and J5 markers in
particular provide adequate visibility). Indeed, except when there are illuminated zones in the
vicinity, there is no proof that illumination improves night-time safety.40 In fact, illumination has
certain drawbacks: high capital, maintenance and energy costs, poles that are agressive
obstacles which are difficult to isolate adequately at intersections.
38
Safety barrier is the term employed in French standards (and this document) to designate all restraining devices.
39
It must be considered that generally it is not possible to protect obstacles in the immediate vicinity of the intersection to a
satisfactory degree from all the possible paths of a vehicle that leaves the roadway (on its own or after colliding with
another vehicle).
40
Illumination does not make vehicles more visible to each other.
66
CHAPTER 3
RURAL ROUNDABOUT INTERSECTIONS
u 1. GENERAL CONSIDERATIONS..........................................................................69
1.1. DESIGN PRINCIPLES................................................................................................ 69
2.4. EXITS.......................................................................................................................... 83
67
This chapter deals with the construction and geometric design of roundabout
intersections, commonly known as "roundabouts."
A roundabout is an at-grade intersection that provides the highest level of safety.
However, its performances can be degraded if precautions are not taken, be it during the
design phase (selection of the size and location of the roundabout, attention given to its
comprehensibility and visibility, trajectories of the various legs, layout of all the elements that
make up the facility, etc.), or during the actual implementation phase (construction of the
central island, selection and positioning of the signing, etc).
The adherence to a certain number of guidelines will ensure a good level of security
and optimize the capacity of the planned intersections, even if the conditions that favor safety
and capacity are not always compatible with each other. First of all, because it forces a
major slowdown, if not a complete stop, a roundabout must be well perceived by all users
who approach it. It must be rapidly identified as such, well before reaching the area where
calculating stopping distance is needed.
This chapter describes the principles that must be observed in order to obtain both a
good level of safety and a concurrence with the features of the types of traffic. It also
specifies the rules and parameters required for the construction and sizing of the various
components of the facility.4
Figure 1: The Major Components and Parameters of a Roundabout
Spliter island
Radius of the Roundabout (Rg )
Central island
Radius of entrance (Re)
Branche
The radius of a roundabout intersection (Rg) consists of the radius of the marking on
the outer perimeter of the circulating roadway, namely the radius at the right curb of the
circulating roadway. Defining a roundabout according to its outside radius provides a much
better notion regarding the intersection's footprint, and offers a better understanding of what
the actual turn movement constraints on the most disadvantaged vehicles will be (trucks,
especially).
4
In the case of the adaptation of an existing intersection, it might sometimes be quite difficult to confirm some
of the recommendations in this chapter. This does not mean that the roundabout concept should be
abandoned when it does turn out to be necessary, particularly from a safety standpoint (as other types of at
grade intersections generally don't perform any better). As the case may be, special attention must be given to
all the aspects of the facility and to its special features (e.g., illumination and signing).
68
1. GENERAL CONSIDERATIONS
The conceptual and geometric design phases of roundabout intersections must take
into account the basic principles enumerated in Chapter 1. Beyond that, as with any other
type of intersection, the construction of a roundabout must abide by the following guidelines:
- Exclude all road hazards from the likely trajectories of vehicles that could
accidentally leave the roadway. In particular, the central island should not
include any obstacle or device that could aggravate the impact of an out-of-
control vehicle at the intersection's entrance;
- Provide the user with a good perception during the approach of the
intersection: i.e., appropriate geometric design, visible and legible advance
signing that states as clearly as possible the type of facility that one is
approaching;
- Confirm that the capacity of the facility is adequate to manage the type of
traffic involved. Saturation of a roundabout (i.e., of one its entrances) is rarely
reached in rural areas. A cursory analysis can be performed when traffic flow
is low; a more accurate analysis is required in cases of higher traffic flows
(See 1.2.);
69
1.2. ADAPTATION TO TYPES OF TRAFFIC
For this purpose, one should always seek nothing more than an adequate capacity
reserve, calculated according to realistic traffic data. Facilities aimed at improving capacity
should only be built if one entrance is saturated (by adding special lanes or a second entrance
lane for example); these facilities generally tend to reduce safety.
Traffic flow forecasts may eventually force the consideration of capacity changes. A
phasing-in process conceived at the design stage (transition from 1 to 2 lanes at the
entrance, for example) can help to optimize safety conditions at each evolutionary stage of
the facility. In this case, the initial sizing will match the existing traffic flows at the time of
commissioning, and long-term traffic forecasts will be used to reserve the footprint required to
implement the later phases, as well as other possible arrangements.
Planning a roundabout with the option of grade separating it at a later date is not
recommended: for whatever dubious savings may be made, this solution is excessively large
and may cause safety problems in the long run.
Even roundabouts with moderate radii can allow enough room for trucks to complete
their turns (See 1.4.4.). However, if there are severe footprint constraints or extremely
cumbersome vehicles, special accommodations (traversable sections) offer an acceptable
tradeoff (safety vs. cost vs. turning conditions).
Under the rule of "right of way to the circulating roadway," there is (theoretically) no
longer any interaction between the entrances. On the other hand, a disruption downstream
from the roundabout can always cause a dysfunction of the intersection by blocking traffic on
the circulating roadway.
In order to perform a capacity analysis, the traffic at peak hours must be evaluated;
regular peaks should preferably be chosen over exceptional peak hours. In some
configurations tied to extreme swings in the traffic flow, the traffic flow of the so-called reverse
peak must also be taken into account. Strong seasonal variations must also be factored in.
2
Appendix 2 features more information on capacity.
70
Evaluation of the Studies to be Conducted
A brief observation of the traffic entering the roundabout during the peak hour (QTE ) will
make it possible to determine the required level of research into the facility's capacity:
• QTE between 1,500 and 2,000 v/h: a flow distribution analysis is required. In this
case, if the sum of the entering and turning traffic flows on the most frequented leg
exceeds 1,000 v/h, a capacity test is recommended;
Generally, a capacity reserve greater than 30% (for the entry under consideration) can
be deemed adequate. 3
A high capacity reserve (> 80%) on a major entrance should cause one to verify if its
width (especially its number of lanes) is not excessive. If the capacity reserve of all the
entrances is suitable (between 30% and 80%), the width of the circulating roadway may
possibly be reduced (within the limits specified in 2.2.).
If the capacity reserve is low (between 5% and 20%), the focus will be on the wait
times (for a possible cost-benefit study), the length of the queues (proximity of other
intersections, loss of visibility during the approach) and on the evolution of the traffic flow in
the years to come. In this case, weekly or seasonal peaks will have to be studied.
If the capacity reserve is under 5%, all the more so if it is negative, major disruptions
can to be expected: one therefore has to develop solutions aimed at improving capacity:
• widening of the entrance (without deviating from the values listed in 2.3.);
71
1.3. RULES OF PLACEMENT
The facility must produce a break in any excessively linear trajectory of the road, as
soon as it is built. The screen produced by the intersection and its landscaping must be
effective by day and by night. Furthermore, the D42b directional sign (diagrammatic advance
directional sign) is a fundamental component of the intersection's identification process. It is
not mandated by law, but its placement as a matter of course is recommended (except on
extremely secondary legs). It must be perfectly visible, and located at least 150 m from the
entrance to the roundabout on two-way roads, and at least 200 m away on roads with two
lanes in each direction (See 4.1.3.).
The components of the roundabout (splitter island of the leg involved implemented by
a J5 marker, as well as the central island) should become visible at a distance of 250 m,
according to the standard methods of calculation of the visibility of an obstacle (line of sight at
a height of 1 m and at a distance of 2 m from the right curb, object observed at a height of
0.35 m). At any rate, the rules of stopping distance must be observed (See chapter 2, Table
3). If it is absolutely impossible to provide an adequate sight distance, and if the roundabout
solution is not reconsidered, a lengthening of the islands may constitute a remedy.
The geometry of the roundabout must be legible. After having become aware of the
presence of a roundabout, the user must be able to quickly recognize its various
components: the central island, the splitter island at the entrance, the outside curbs, the
circulating roadway, as well as other entrance and exit lanes.
There are some aspects of the design of a roundabout that can impair its perception
or legibility. They should be avoided (∆), or excluded (1):
∆ if the intersection is located on a curve or at the end of a curve, m if the radius is less
than the non-slanted radius (See 1.4.3.);
∆ if the axes of the legs are off-center (m to the right) in relation to the central island (See
1.4.3.);
∆ if the intersection is on a convex curve along one of the roads, even if it has a very
large radius in an outward projection,6 especially after a crest. If this configuration
cannot be avoided, placement at the crest is often preferable, and special attention
must be given to conditions of visibility during approach (See above);
1 trees in alignment with one of the roundabout's nearby legs--all the more so on either
side--giving the impression of a continuous trajectory;
72
1 slanting the circulating roadway toward the inside of the roundabout.7
1.3.2. VISIBILITY
Drivers approaching a roundabout must see the vehicles that have the right of way
soon enough in order to yield or even stop. A large sight triangle is however not required; a
complete view over the left quadrant of the circulating roadway at a distance of 15 m
(approximately) from the entrance, is enough.8
Furthermore, the central island must not include any visual obstacles (high plants)
within less than 2 m of its peripheral curb (or, if there is no curb, at 2.50 m from the edge
marking surrounding the central island).
In general, similar sight triangles should be designed on all the legs of a same
roundabout (in order to induce consistent behaviors).
2,50 m
Leg A
7
Except perhaps, on certain areas of the circulating roadway, in the very special case of a roundabout
located on a road with strong grade (See 1.3.6.)--a solution which, by the way, is not recommended.
8
Too much visibility to the left can even be harmful to the facility's safety. Indeed, drivers approaching the
roundabout may focus their attention on the open spots on the entrance directly to their left, while neglecting
other movements that are not so visible (if there is heavy masking from the central island, for example).
73
1.4. GENERAL GEOMETRIC CONFIGURATION
In a rural area, a roundabout can have between three and six legs. Furthermore, it is
always preferable to add a leg to the roundabout rather than maintain or create a secondary
intersection nearby.
A regular distribution of the legs around the circulating roadway is preferable (this
issue however is not crucial, considering item 1.4.3.): a well-designed distribution can
markedly improve the facility's legibility.
The central island is optimally located when all the axes of the legs cross through the
center of the roundabout. Since this configuration isn't always possible, the island should first
be centered on the main axis and then, as much as possible, on the axis of the secondary
legs. Although it is always desirable to have the axes of the secondary legs pass through the
center of the island, a slight offset to the left is acceptable. However, the direction of the
secondary leg should never produce an excessively tangential entrance.
74
On a new facility, a radial alignment over a length of approximately 250 m should be
sought (350 m in the case of a 2X2 entry lane). When converting an existing intersection into
a roundabout, this distance can be reduced to 150 m (250 m in the case of a 2X2 entry lane).
Figure 4: Radial alignement of the legs
Alignment
Alignment
P2
The latter configuration has the advantage of allowing the placement of large visual
screens. However, for roundabouts with radii (Rg) greater than 15 m, this solution produces
approach configurations of the "curve/reverse curve" type, likely to impair the facility's
perceptibility.
Figure 6: Configuration of the legs of a roundabout in a "Y" shape
75
Figure 7: Configuration of the branches of an offset "T" shape
Designing a very large facility is seldom useful. Experience shows that large
roundabouts do not provide greater safety than small ones (for often minute improvements in
capacity and much higher costs).
a) On a single-lane road:
- however, on secondary roads with very little truck traffic, a radius (Rg) between 12 and
15 can be considered;
- on more important roads, and if the constraints of the project make it possible, a value
of approximately 20 m should be sought for the extra level of comfort they provide to
trucks;
- if there are many legs (> 4), radii in the range of 20 to 25 m may considered, but rarely
higher, considering that some minor legs do not require much space (See 2.3. and
2.5.2.).
b) On a road with two roadway beds: a radius (Rg) of 25 m (no more) is generally
advisable.
In all cases, the width of the circulating roadway cannot be less than 6 m.
76
1.4.5. DEFLECTION
The deflection of vehicles' paths through a roundabout (the trajectory traced by two
opposing or two adjacent arms of the roundabout) is a major factor determining the safety of
the facility. Indeed, the overall geometric design should not allow the most «stretched out»
trajectories to be taken at speeds in excess of 50 km/h.
A trajectory's deflection is the radius of the arc that passes at a 1.5 m distance away
from the edge of the central island and at 2 m from the edges of the entry and exit lanes. The
radius of such an arc should be less than 100 m.
Figure 8: Deflexion
R < 100 m 2m
2m 2m
1,50 m
77
1.4.6. GRADES
Between 3% and 6%, some designs can adversely affect safety, notably by impairing
the stability of trucks (steep banking, high entrance speed, etc.).
If the grade is higher than 6%, it is generally acknowledged that this type of facility
can create some serious problems. However, under the same conditions, another type of at-
grade intersection does not always perform better and offers a lower level of safety. The use
of a roundabout can therefore not be excluded in principle at grades of 6% or higher, through
the modification of existing roads. In the case of a new facility, giving up the roundabout
solution should not lead to the acceptance of another type of intersection, but to the
elimination or relocation of the intersection, or to a modification of its vertical profile.
In all cases, at no point should the banked areas on the outside of the circulating
roadway or the normal banked areas on the entry and exit lanes exceed a grade of 3%,
including the merge lanes to the left. For roundabouts on a sloped plane, no slope should be
added to the normal transverse grade of the circulating roadway (1.5% to 2%).
For steep grades, (5% to 6%), the slope may change around the circulating
roadway, for example between +2% at the crest of the circulating roadway (road tilted toward
the inside) and -2% at the low point (road tilted toward the outside).
If the intersection is located on the slope, or at the low point of the vertical profile of the
roads involved, a smaller facility can reduce the slope of the circulating roadway by about 1 to
2%.
78
2. GEOMETRY OF THE ELEMENTS OF A ROUNDABOUT
2.1.1. SHAPE
The central island must be circular; if it has another shape (oblong shapes or shapes
made up of arcs of circles and connecting segments), its level of safety is oftentimes
seriously degraded.
2.1.2. SIZES
There is no maximum recommended radius for the central island, but designing it too
large is unnecessary; it does not improve the intersection's operation (no or only minor
capacity gains) and oftentimes produces negative effects (higher speeds on the circulating
roadway, higher costs...). In fact, moderate length radii should be chosen (See 1.4.4.).
However, a central island with an inscribed radius of about ten meters is generally
desirable if there is major semi-trailer traffic (always the case on main rural roads), to ensure
a decent comfort level for the movements of these vehicles.
The traversable strip allows, among other things, to enhance the trajectories of
trucks, whose rear wheels could damage the facility or simply the curbs of the central island,
and provides improved turning conditions for more cumbersome vehicles. The traversable
strip is tilted toward the outside of the circulating roadway at a grade between 4% and 6%. It
is recommended that this strip be separated from the road by a relatively substantial and
inconvenient raising (low, half-buried curbs, of which 3 cm at most are visible, for example.
See Appendix 5, "Curb Design"). Very light dynamic effects or visual devices do not offer
lasting improvements on trajectories or safety. It must also be covered (with a "coated"
material, concrete cement or cobblestones, etc) and offer contrast against the circulating
roadway that is perceptible by day as well as by night.
79
2.1.4. LANDSCAPING
In general, the purpose of landscaping is to differentiate the roundabout from the "road
environment" and the immediate surroundings. The landscaping of the central island can
improve the perception of the roundabout from a distance, and block the perspective of the
incoming user on the circulating roadway. It may also help to beautify the road environment,
and underscore the transition process of the arrival into town, etc. On the other hand, the
nature or position of some types of vegetation can degrade the facility's perception. The
central island may be slightly elevated, but the slopes of the landscaped area should not
exceed 15%.
In rural areas, no obstacle that is aggressive or likely to cause the sudden stop of an
out-of-control vehicle (tree, massive sculpture, stone block, pole, lighting fixture, little wall,
earthen wall, embankment at a grade above 15%, ditch, etc) should be located on the central
island. This does not prohibit the use of non-hazardous materials (shrubs, flowers, water
fountains,9 sculptures made out of fragile and low-mass materials, etc).
The maintenance requirements of the central island are discussed in paragraph 3.2.
4à6%
2m 1m
1,5 à 2 %
6 cm
15 % 3 cm <14 cm
2. for Rg = 20 m
1m
2m
15% 6cm
1,5 à 2 % <14cm
0,50 m 0,50 m
(ou 5 u )
12,50 m 7,00 m
R g =20 m (circulating roadway)
9
Translator's note: footnote cropped from photocopied page.
80
2.2. THE CIRCULATING ROADWAY
2.2.1. OPERATION
The circulating roadway should not be seen as a one-way road with 2 or 3 lanes
separated by lane markings responsible for their allocations, but as a single lane, wide
enough to allow trucks among others to complete their turns.
The width of the circulating depends on the Rg radius, and on the width and the
number of lanes at the widest entrance. The width must be consistent (no extra wide areas
between two legs should be allowed). It is 20% wider than the widest entrance, with 6 m as a
minimum width and 7 m being the standard width. A width of 8 m is justified for roundabouts
that are highly used by semi-trailers. In the case of two-lane entries, the width of the
circulating roadway shouldn't even exceed 9 m,10 the standard value being 8.50 m.
The circulating roadway must feature a consistent banking between 1.5 and 2%. The
slope is tilted toward the outside of the intersection for three main reasons: to improve the
perception of the circulating roadway, to avoid breaks in the slopes on entrance and exit lanes
(a cause of discomfort and even instability to some vehicles), and to help the drainage of
surface water.11
These provisions do not apply to the special case of a roundabout located on a highly
sloped road (See 1.4.6.)-- a situation which should be avoided in all cases. However, at no
point on the circulating roadway should the transverse slope exceed 3%.
2.2.4. MARKINGS
In general, lane markings are not recommended; they should only be used for
circulating roadway widths in excess of 9 m or more, as long as the Rg radius is at least 20
m. Therefore, the circulating roadway is most often delineated, on the inside as well as the
outside, by a continuous line (except at the entry and exit lanes).
10
This width can sometimes be higher on existing roundabouts in high traffic areas located in the
outskirts of a town, where there are often 3 lanes on the circulating roadway: these very special cases
(such as a site listed as a historical landmark on which grade-separating is not allowed) can be
tolerated.
11
The aforementioned grade is entirely adequate for surface water drainage.
81
2.3. ENTRANCES
At each leg, the entry lane must be separated physically from the exit lane by a
projecting splitter island. Simple pavement markings are not enough (except perhaps to
delimitate extremely secondary legs).
Entries are generally single-lane, except when the capacity calculated at the time of
commissioning requires the creation of two-lane entries. If the capacity study shows the
need to build more than two lanes on one entrance, the choice of a roundabout for the
purpose of solving the problems of exchanges between lanes may have to be reconsidered.
Nonetheless, under these conditions, it is generally not acceptable to chose a standard at-
grade intersection or one with traffic lights, which would be more unsafe than a roundabout.
Grade-separated solutions must then be contemplated, with one roundabout on each axis:
the good thing about them is that they don't encourage high speeds, and avoid transferring
safety problems to intersections located downstream. Otherwise, the roundabout solution
can be chosen in spite of the periods of saturation it will generate.
The recommended entry widths (le) (measured between pavement markings) are as
follows:
The entrance radii (Re) must always be less than or equal to the outside radius of the
roundabout (Rg). They generally range between 10 and 15 m (depending on the
configurations of the legs around the circulating roadway).
The entrance lanes are delineated by pavement markings (T3 type on the outside
perimeter) and by (traversable) curbs on the inside, located most often at the edges (See
Appendix 5). For the smallest entry radii (Re £ 12 m), the outside curb can be replaced with
cobblestones abutting the road surface, whose surface properties will act as a deterrent
against crossing.
Rr
le =4 m
Ri
la =7 3,75 m 3,50 m
1,0 m
m 3,50 m
4
ls =4,50
Rr
Rs
12
To obtain the maximum benefit from two-lane entries, it should be recalled that the circulating roadway
should be at least 20% wider than the widest entry lane, with a limit of 9 m.
82
2.4. EXITS
Exits are always designed with one lane, except in the following cases:
- the exiting traffic (Qs) is higher than 900 uvp/h, as well as 3 times higher than the
circulating traffic (Qt).14
The width of the exits (ls) is set at 4 or 5 m for one lane (depending on the value of
(Rg); it is also quickly reduced to the width of the roadway in a typical cross section, in
actuality at the connection point with the alignment on the right.
For two-lane exits, the ls width is usually set at 7 m. When the road includes on its
main roadway only one lane per direction, the merge from two lanes down to one lane is done
in the tangent section, in accordance with standard practices and at a speed of 60 km/h.
The exit radius (Rs) should be greater than the inscribed radius of the roundabout (Ri ),
with a minimum of 15 m and a maximum of 30 m. There are some special configurations of
the legs that may justify greater exit radii. Furthermore, the placement, right after an exit, of a
counter-curve with a smaller radius than the exit radius (Rs), should be avoided due to safety
and fluidity concerns.
Table 1 -- Summary of the Construction Parameters of Entrance and Exit Lanes
Roundabout Radius Rg 12 m £ Rg £ 25 m Rg = 12 Rg = 15 Rg = 20 Rg = 25
* If these radii generate major modifications to the path of a secondary leg (which can occur when the
axes of two consecutive legs produce a sharp angle), they can be reduced down to the minimum listed
values (Re = 10 m and Rs = 156 m), or even lower. In this case, the turning maneuvers of trucks that
use the facility must be verified and, as the case may, specials arrangements must be made for them
(excess paved width area).
13
UVP: French acronym for "units of private vehicles"
14
In both cases, traffic is measured at peak hours.
83
Figure 4: Case of a four-lane road with two roadbeds: typical treatment of the exit, and
example of the restoration of the second lane after a merge.
2.5.1. FUNCTIONS
Splitter islands perform six major functions:
- they serve as a refuges for pedestrians, allowing them to cross a roadway in two stages;
- they prevent collisions between the two directions of travel (especially where the exiting radius is
small) by separating entry and exit movements;
- they increase capacity, by allowing drivers waiting at the entry yield line to ascertain earlier
which other vehicles are exiting and which ones they must yield to;
As is the case with other types of at-grade intersections, the size of the island
represents, by definition, the neutralized space located between the roadways of the lanes for
the various directions of travel (namely the "envelope" of the painted island).
The splitter island is generally triangular and tapered at the base; its shape is
generated by a so-called construction triangle.
84
The position of a construction triangle of a standard island is derived from the axis of
the leg (which indicates the direction that determines the height of the triangle) and the edge
of the circulating roadway (which determines the base of this height). For a roundabout
radius (Rg) greater than 15 m, the construction triangle is slightly offset to the left, in order to
allow the axis of the road to pass through the center of the island's nose.
Whenever possible, the height (H) of the triangle should be greater than 15 m. In
actuality, the construction triangle can be assigned a height equal to the radius of the
roundabout. An island width (li ) of 4 m is adequate for small roundabouts.
The lowest acceptable width of a splitter island is 2 m. In reality, the construction triangle can
have a base (B) equivalent to a quarter of the radius of the roundabout.
The splitter island should be surrounded by low edges, preferably retroreflective. The
curb markings on the entry and exit lanes follow the edges of the splitter island at a distance
of 0.50 m (5 u at least), including the marking of 3 u. It should not feature any objects (for
decoration, marking or superfluous signing) in the sight triangle area defined under 1.3.2.
The surfaces of the splitter islands should preferably be light in color, and create
enough of a contrast (by day as by night) against the road surface.
With extremely secondary legs, the splitter island can be implemented just through
markings.
Figure 5: Construction of the splitters islands on the legs of a roundabout with a radius (Rg)
≥ 15 m (extremely secondary legs excepted)
H=Rg
R=Re +le
*
0,50 m (ou 5 u )
Rg
r=Rg/50
B=Rg/4 Rr =4Rg
d=(r+0,5)/2 1,0 m r=Rg/50
r=Rg/50 Rr =4Rg
Rg/16
* l
R=Rs+ s
85
Figure 6: Construction of the splitter islands on legs of roundabouts with a radius ( Rg) < 15
m (extremely secondary legs excepted).
H=Rg
R=Re+ le
*
0,50 m (ou 5 u)
Rg
r=Rg /50 Rr=4Rg
B=Rg /4 1m r=Rg /50
r=Rg/50 Rr=4Rg
*
R=Rs +ls
* The curves that allow a tapering of the base of the island are respectively parallel to the
right edges of the entry and exit lanes.
Offset of the Island from the Axis d d = (0.5 + Rg/50)/2 or 0 0 0.40 0.45 0.50
In general, right-turn slip lanes should not be provided on roundabouts. They diminish
the drivers' understanding of the intersection and installing directional signs is more
complicated. They can also reduce safety for pedestrians and bicyclists. Slip lanes
alongside the circulating roadway should be excluded. But a right-turn slip lane can be
justified if there is a heavy right-turn demand, and if this movement could lead to saturation,
even if two lanes are provided (See 1.2.1.). In this case, the right-turn lane eliminates the
need to provide an extra entrance lane, but it is important to pay particular attention to the
spread of traffic movements at reverse flow times.
Traffic in the right-turn lane should yield to traffic in the leg which the slip lane merges.
The connection is generally achieved by a merge lane. However, if this axis includes two
lanes per direction in a typical cross section, the slip lane can be kept parallel to the
roundabout's exit, in order to create the rectilinear lane on the right, whereas the lane
86
emanating from the roundabout will be allocated to the left.
87
2.6.2. GEOMETRY
The lane should be 4 m wide, from the straight exit transition to the entrance of the
merge lane. It includes a 2.00 m paved shoulder on the right, and a 0.50 m paved shoulder
on the left.
The signs used for this situation should include modified D42b signs: type 1 or 2, as
seen in Figure 15. The choice of one of these two signs will depend on the geometry
assigned to the right-turn lane, and especially the length of the deceleration lane.
L i = 70 m 70 m
R ≥ 40 m
R.P.
L b = 80m
15
The condition that must be assured is Lcl £ 0.017xRxa, whereby Lcl is the length of the easement, R is the
radius of the circular arc, and a is the overall angle variation (in degrees).
16
It is not recommended to design trajectories that would be so easy and direct that they would encourage
excessive speeds by vehicles on the way out that should be attentive to the right of way of vehicles exiting the
roundabout.
88
2.7. SPECIAL CASES OF DIFFICULT APPROACHES
In general, the slight inflection to the right caused by the presence of the splitter island
is sufficient. In some difficult cases (subsequent to a long, tangent alignment or if visibility is
reduced), two trajectory inflections can be included, more to warn the user than to slow him
down (See Figure 15). However, when visibility during approach of the roundabout is
inadequate, lengthening the splitter island generally remains a preferable corrective solution
(See Figure 16).
Designing the entry lane in a reversing curve shape is therefore generally not
necessary; it even reduces the facility's safety if the roundabout's center is not located on the
axis of its roads.
Figure 85: Example of a difficult approach with landscaped central median and two
inflexions
30 m 60 m
3 m
6 m
R=100 m R=200 m
Figure 96: Example of a difficult approach treated with an elongated splitter island
~2 m
The installation of slowing devices (such as rumble strips) on the braking areas is
generally not recommended due to the loss of adhesion they may produce.
89
3. SPECIAL ACCOMMODATIONS
3.1.1. PEDESTRIANS
For the sake of pedestrian safety, wide and rapid entrances (and exits) should be
avoided. Furthermore, the intersection should be designed compact in size in order to reduce
out-of-direction travel.
When passing through a roundabout, two-wheel vehicles are clearly those exposed to
the highest accident risk (even though it is still lower than at other types of at-grade
intersections).
The positioning of bus stops must be consistent with the routes' requirements;
accesses must always be designed in a way that minimizes the number of crossings and the
lengths of the trajectories.
- on the entry lane (on the roadway), immediately upstream from the crosswalk, if traffic
is moderate and stops are short in duration. This design cannot be used if there are
two entry lanes (a vehicle cannot be allowed to overtake a stopped bus);
90
As a general rule, stops right in the circulating roadway should be avoided; however,
on the largest roundabouts in existence, a complete bus pullout can be considered in the
periphery of the circulating roadway, as long as no disruptions are generated (such as by
pedestrians crossing the circulating roadway).
Figure 10: Two possible bus stop locations
10 m 20 m 10 m 1 m 1 m 20 m
3 m
2,50 m
4 m 4 m
By their very definition, special convoys exceed the regulatory limits defined in the
Vehicular Code. Through a special exemption,17 convoys are authorized to travel over
predetermined itineraries, that can "handle" their special features.
17
Regulated under the terms of the Memorandum No. 97-48 of May 30, 1997.
91
Finally, for left-turns, when the convoys are escorted, traveling against the flow of
traffic can be considered; in this case, the convoys slightly overlap the traversable splitter
islands (which will therefore require removable amenities).
Due to safety considerations, long convoys should be kept away from large
roundabouts that do not provide trajectories that are any better than those of small
roundabouts unless they feature a large entrance radius and a wide circulating roadway--
these provisions would allow regular traffic to travel at high speeds.
The construction of a "pierced" roundabout cannot be excluded, closed off through the
use of removable equipment, which represents, in some instances, an attractive solution for
direct movements. In this case however, the passage should be (a) sufficiently slanted in
relation to the axis of the legs,18 in order to prevent a poor perception of the facility upon
approach, and (b) without any hazardous devices or obstacles on the central island (See
2.1.4.).
b) Method
There is no standard approach for taking into account the traffic constraints caused by
special convoys. The facilities have to be designed on a case by case basis, according to the
characteristics of the convoys, the turn movements, the configuration of the intersection, etc.
However, the following method can be applied:
1. chart a roundabout that meets the facility's requirements and conforms to standard
design practices, without concern for special convoys;
2. identify the technical features of the convoys that follow the particular itinerary, by
going beyond their administrative classification in order to inquire about their sizes,
their turn radii, etc, as well as the turns executed in the intersection;19
3. represent the trajectories of special convoys and the areas covered by the wheels and
the overlapping sections (turn movement template);
4. include traversable areas on the central island, the entrance and exit splitter islands,
and the edges;
5. identify areas without elevated equipment or removable components for the areas
covered by the overlapping sections.
c) Extra-wide traversable areas
The extra-wide (semi)-traversable areas of the central island and possibly at the
edges (entrance and exit of the circulating roadway) make up the basic facility aimed at
improving the turns of special convoys. However, these areas should be treated in such a
way as to dissuade light vehicles from driving over them. To make the most out of the
available space without jeopardizing the safety of other users, the following precautions must
be taken:
- on the central island, a slope tilted to the outside of the circulating roadway, at a grade
between 4% & 6%;
- a counter-slope, at the level of the traversable areas at the edges (as the case may
be);
- a raising of the extra-wide areas above the roadway (not to exceed 3 cm);
18
In this case, convoys pass on the left of the entrance splitter island and on the right of the exit splitter
island.
19
A useful resource is "Transports exceptionnels - Définition des convois-types et règles pour la vérification
des ouvrages d'art" (SETRA - DR; October 1983).
92
- a rough surface that provides a noticeable contrast, by day and by night, against the
circulating roadway;
- separation of these areas from the circulating roadway via regular, retroreflective
markings.
The traversable area of the central island need not be circular; lenticular shapes, for
example, can be considered.
Figure 11: Example of the treatment of the circular island and the edges in order to
facilitate the turns of special convoys moving directly or executing turns.
Traversable extra-wide
areas on the edges
Traversable section of
the central island Non-traversable area
of the central island
3.2. MAINTENANCE
Due to the fundamental role of maintenance (upkeep of the appurtenances,
enhancement of safety and comfort, preservation of the vegetation and planted areas, etc),
this aspect must already be included at the design stage.
Since the rules governing maintenance remain basically the same as those for the
rest of the path, the reader should refer to the ARP and to the specialized literature. However,
the central islands of roundabouts, which are often landscaped, do invite special comments.
Safety and accessibility problems are tied to the maintenance of these central islands
(in addition to the induced costs and limitations). Everything should therefore be done to
reduce and optimize the tasks involved (grass-cutting, watering, pruning, etc). In this regard,
plant species that are similar in appearance, slow growing, rustic and sober looking should be
chosen. Furthermore, one should bear in mind that the maintenance of a grassy area
requires regular cutting (in particular to meet the visibility requirements stated in 1.3.2.).
When the landscaping requires regular watering, automatic sprinklers should be included, as
well as a system to dispose of excess water. A treatment in stone of the central island can
be considered.
Furthermore, the durability and the sustainability of the landscaped areas must be
ensured, to prevent changes in their functions (safety, embellishment, etc) over time or at
certain periods of the year (in the winter).
93
Finally, during maintenance work, steps must be taken to ensure access to the central
island by maintenance vehicles and protect the safety of the workers. Potential stoppage
areas reserved on central islands reserved for maintenance vehicles must take into account
visibility at time of pull-out and should not be located facing entrances.
ø The type of pavement used for the circulating roadway and the entrance and exit legs
should be chosen according to the types of demands exerted upon it. The
juxtaposition of different types of pavement should be avoided.
ø The upper layer should be at least 4 cm thick: its formulation should be suited to the
demands: stability and tear resistance should be top priorities.
Note: Research laboratories in the area should be consulted in order to determine the best
technical solutions (base and wearing surface), based specifically on local conditions.
ø Plans should be made for the disposal of the water used on the plants of the central
island (as necessary).
ø Plans to limit tangential efforts should not lead to the increase of other parameters
(entry radius, inscribed radius of the central island, etc.); such measures would
reduce safety and could increase the cost of the facility. The aforementioned
precautions are generally adequate and do not produce negative effects. However, by
choosing a connecting geometry in accordance with standard designs (radial
alignment, no " reverse curves," etc), one can reduce some of demands made on the
facility. Moreover, the slope must be regular and moderate (1.5 to 2%).
20
These precautions should be modulated according to the amount of truck traffic.
21
This rule should be applied only if possible on existing roads. Construction work in [[text missing]...
doesn't always make it possible.
94
4. AMENITIES AND SIGNING
Regardless of whether the facility involves a new or an existing road, the study of the
project's design must anticipate, and as far upstream as possible, what measures will be
taken to manage the road (signing, protective barriers, equipment, etc), to ensure that the
facility's geometry will be compatible with the needs and performances of the amenities.22
The issues related to amenities and signing are listed below; for greater details
(nomenclature, usage conditions, etc), please refer to the published instructions, memoranda
and other technical guides.
4.1. SIGNING
Although not required by current regulations, the placement on each access lane of an
AB3a sign (YIELD) and the marking of its transverse line--T'2 type, 50 cm--is strongly
recommended. The AB3a signs are only repeated on the splitter islands if the entrances
include 2 lanes. We remind the readers that adding blinking lights to the AB3a signs is not
allowed.
On major traffic roads, the placement of an AB7 sign (END OF THROUGH ROAD)
upstream from the roundabout is unnecessary. However, this does not prevent one from
placing AB6 signs (THROUGH ROAD) downstream from the roundabout, on the legs leading
to the major transit route.
22
The strong interdependence between geometry and the proximate or interchange amenities often requires
special studies on signing (road signs, pavement markings and directional signs) and amenities, concurrently
with the general trajectory studies.
23
In accordance with the Ministerial Order of February 16, 1984.
95
On the central island, the only regulatory sign that must be placed in front of each
entrance is the B21-1 sign (MANDATORY DIRECTION), to the exclusion of any other sign
(J4, B1, B21f, etc). The B21-1 sign is placed on the axis of the direction of the entry lane at
approximately 20 m from the YIELD line (See Figure 19).
The B1 sign (WRONG WAY) is only useful on legs made up of one single entrance
(such as a roundabout as part of an interchange).
A25
B21-1
J5
As is the case for the A25 sign, it is recommended that the D42b sign be placed when
the work begins, as soon as the central island has been created and forces a trajectory
constraint. When the D42b sign is used twice (green and white color-coded), a placement
interval of at least 100 m is required.
E42
D21A (green)
D21A (white)
96
The diagram should attempt to provide the best possible representation of the layout
of the intersection. If there are more than four legs, or if they are very irregularly arranged
around the circulating roadway, the legs of the diagram should be separated by 45° angles or
by multiples of 45°.
Around the circulating roadway, exits should be announced by D21 signs. These
signs are placed on posts located on the splitter islands. The positions of connecting roads
should not be mentioned (indications such as "OTHER DIRECTIONS" are prohibited).
Furthermore, the position of direct right-turn lanes is indicated via a D31 sign.
4.1.4. DELINEATORS
The splitter islands are announced by J5 delineators, located at the center of the
island heads. The use of any other warning device (such as J4 delineators, delineations, J11
delineators, etc) could reduce the intersection's comprehensibility. These devices should not
be used on new facilities.
On the feeder lanes, the edge, approach and surrounding markings of the islands
must adhere to the "Instruction ministérielle sur la signalisation routière, Livre 1, 7ème partie:
marques sur chaussées." [National guidelines on road signs, Book 1, Part 7: Pavement
Markings.]
On the circulating roadway, lane markings are generally not necessary; they should
even be avoided when the width of the circulating roadway is under 9 m or when the
roundabout's radius (Rg) is under 20 m. Indeed, users do not follow the guided pathways,
especially truck drivers who are restricted by the turning capabilities of their vehicles. On the
other hand, axial markings on the circulating roadway (T3 - 2u type) can be useful on some
very large existing roundabouts (not recommended for new facilities) in order to provide
adequate guidance to users in merge areas.
The markings of the edges lines of the circulating roadway are continuous, with a
width of 3 u.
T’2 (0,50 m)
Continuous (3u) T3 (3 u)
L/2=58,50 m L = 117 m
97
4.1.6. TEMPORARY CONSTRUCTION SIGNS
When construction begins, and as soon as the central island is created and generates
trajectory constraints, as well as during the months following the creation of the roundabout,
we recommend the placement of KC1 type temporary signs bearing a message such as
"WARNING: MODIFIED INTERSECTION", in addition to the D42b and A25 signs.
The provisional use of temporary K8 type signs (red chevrons) can be useful to
reduce the risks of accidents associated with driver habits that persist after the
transformation of an at-grade intersection. This problem is mostly relevant to the legs that
had right of way prior to the intersection's transformation. After a few months however, the
signs and delineators must remain unostentatious in order to remain credible.
For projects involving existing roundabouts, the various modes of operation and their
related temporary signs are described in the document entitled "Manuel du chef de chantier:
Routes bidirectionnelles" [Two-way roads: A construction site manager's guide]. (SETRA;
1993).
Moreover, the design and construction of a roundabout should be done such in such a
way that, insofar as possible, one could forego the use of safety barriers--bearing in mind that
obstacles at a roundabout (or in the immediate vicinity) generally cannot be adequately
isolated. Eliminating obstacles, moving them back or making them fragile are all preferable
safety solutions that should always be considered before any decision to protect them with
barriers.
24
Safety Barrier is the term employed in French to designate all restraining devices.
25
See ARP, § 2.2.g.
98
4.3. ILLUMINATION
In general, roundabout intersections in rural areas do not require illumination (similarly
to other at-grade intersections). Indeed, except when illuminated areas are in the immediate
vicinity, illumination has not made any positive impact on the nighttime safety of roundabouts.
- the investment costs can be high (varying according to the chosen solution, the size
of the roundabout and the proximity of a power source);
- maintenance and power costs are high (over 15 years, these add up to between 1 and
1.5 times the initial investment costs);26
- poles are obstacles that are both hazardous and difficult to isolate (See 4.2.).
However, the roundabout must be illuminated if at least one of its legs is illuminated or
if a brightly lit area is near the roundabout.
If the goal is to improve the roundabout's nighttime visibility, one can consider a
"staging" of the facility (indirect lighting of the approaches or, most often, of the central
island).27
In any event, lighting poles located on the central islands of roundabout intersections
should be avoided (just like any other hazardous obstacle or device). If, however, for some
special reason, the lighting cannot be placed on the outside of the circulating roadway, a
central pole may be considered,28 as long as the radius of the central island (Ri ) is at least 10
m.29 This arrangement is also discouraged for Ri values greater than 20 m (poles too high,
requiring excessive illumination power). Furthermore, luminaires should never be placed on
the edges of the central island or on the splitter islands.
26
For costs, please refer to "The Illumination of Roundabout Intersections" SETRA - CETUR; 1991
27
In case of doubt, a wise and safe thing to do is to make room, at the construction stage, for conduits to
accommodate possible inside lighting from the central island.
28
Sometimes, the placement of a central pole causes fewer problems than several luminaires around the
circulating roadway, so close to the road that they would have to be protected by safety barriers (See 4.2.).
29
This distance should not lead to an oversized facility.
99
GLOSSARY
This glossary provides simplified definitions of the main technical terms used in this
guide. It is not intended to be exhaustive and, for the sake of clarity, it has been limited to the
specific meanings of the terms found in this publication. These definitions are aimed mostly
at eliminating all ambiguities due to terminology, and pertain essentially to terms related to the
design of intersections.
Acceleration lane A supplementary lateral lane allowing vehicles joining the road
to accelerate into the main flow of traffic.
At-grade intersection An intersection were all traffic exchanges occur at the same
level.
Capacity The maximum flow that can go through a traffic corridor, a lane,
a leg, etc, considering its characteristics and the intersecting
directions of traffic with the right of way.1
Channelization island An island that separates two lanes used by merging or exiting
vehicles, generally triangular in shape with a concave curb.
Channelization (divergence)The splitting of two flows coming from the same direction into
two separate directions.
Criss-crossing accident A collision between a vehicle on the right of way axis and a
vehicle obligated to yield that is crossing or making a left turn.
Crossing sight distance The sight distance needed by a vehicle forced to yield to safely
cross an intersection, given the speeds driven on the main
axis.
Curb A concrete section that is elevated above the roadway and
1
This is therefore not an inherent feature of the leg.
2
Within an interurban at-grade intersection, a traffic corridor generally only includes one traffic lane and
therefore becomes undistinguishable from it.
100
separates it from the sidewalk or surrounds the non-traversable
islands.
Deceleration lane A supplementary lateral lane allowing vehicles exiting the road
to slow down away from the main axis.
Direct right-turn lane On some special roundabouts, a transition lane allowing the
vehicles of a given leg to avoid having to enter the circulating
roadway to make a right turn.
Entering and exiting traffic All the directions of the flow of traffic between two axes
of an intersection.
Entry approach sight distance The sight distance needed by drivers to see the facility
they are approaching (splitter island of the main or secondary
lane). It should be at least equal to the stopping distance for an
obstacle.
Entry radius (Re) The radius of the arc of the circle connecting the outside curbs
of the entry/exit lane of a secondary leg and the roadway of the
main axis.3
101
Exit lane The extremity of a leg allowing vehicles to exit an intersection in
a specific direction.
Four-way intersection An at-grade intersection with four legs, aligned 2X2 (or almost).
Grade-separated intersection
or interchange An intersection where the exchanges are separated
from each other and processed away from the main axes (in
order to reduce conflicts along intersecting paths).
Graded shoulder A section on the roadway's edge, lacking any obstacle, with
pavement markings, that includes the extra width of the
roadway and possibly an additional section that may or not be
paved.
Graded shoulder (right side) A graded shoulder to the right of the roadway.
Graded shoulder (left side) A graded shoulder to the left of one-way roadways.
Incidence angle (written as Q) The angle created by the axis of a secondary leg and the
perpendicular of the main leg. It is also the complementary
angle of the angle created by the two axes. In particular, the
incidence angle is zero when the two axes are perpendicular.
Inscribed radius (Rg) The radius at the curb of the outside edge of the
circulating roadway (i.e., the radius at the inside curb of the
circulating roadway).
Intersection or crossing angle The angle created by the two directions of traffic.
Island tip A rounded area made up of the ends of the projecting splitter
islands or the tips of the channelization islands.
102
Island tip A special top of an island implementing the channelization of
two flows of traffic (directional islands) or the entrance into the
intersection (splitter island of two opposite flows of traffic).
Left or right slip lane A lane at the entrance of an intersection, reserved for vehicles
turning to the left or right.
Legibility The ability of a road and its environment to provide to any user,
through all its components, an image that is accurate, easily
and readily understandable of the nature of the facility and its
surroundings, of its uses, of the likely or possible movements of
other users, as well as indicate to that user what he is
supposed to do. (Definition taken from "Sécurité des routes et
des rues" [Road and Street Safety].
Main road 1. A road that supports and feeds into a national road system or
regional road systems. They generally carry a daily traffic flow
in excess of 1,500 vehicles (According to the "ARP").
103
Peak thirtieth hour traffic The thirtieth highest hourly flow observed over a period of one
year. This flow is oftentimes used to calculate the size needed
for an intersection's components.
Reserved capacity The difference between the traffic "offer" (capacity) and
traffic "demand" at the entry leg under consideration. It is
expressed in uvp/h. It is often (wrongly) associated with the
coefficient of reserved capacity, which is the ratio of this
differential demand.
Roadway width The portion of the roadway between the edge markings.
Secondary road 1. A road with mostly local traffic, with a low traffic volume
(generally under 1,500 vehicles per day), with all types of users, and
without the right of way at intersections.
2. At a standard at-grade intersection, a road that does not
follow the main direction of traffic (assuming it is possible to
rank the directions hierarchically). At an intersection with a
STOP or YIELD sign, this road usually yields to the main axis.
Splitter island An island that splits the lanes used by cars going in
opposite directions.
104
Staggered T intersection An at-grade intersection with four legs, but whose two
secondary legs are offset (but still close enough to operate as a single
facility) in such a way that the direct movements of the secondary road
are transformed into two consecutive, reverse turns (to the right and to
the left, or the opposite).
Traffic delay Delay caused by the lack of right of way and the
interactions between vehicles. It is often associated with the
wait time at the rear and front of a queue.
105
Vehicle off-tracking The geometric swept path of a vehicle during a change
of direction.
Y-shaped intersection A standard intersection with three legs, with a single secondary
leg at an oblique angle of incidence in relation to the axis
(greater than 20° from the perpendicular).
106
107
BIBLIOGRAPHY
Ñ GENERAL DESIGN AND GEOMETRY
C.1. Dossier pilote "Carrefours sur routes interurbaines". Carrefours dénivelés.
SETRA, octobre 1976.
(Pilot case "Intersections on Interurban Roads." Grade-separated intersections. SETRA, October
1976.)
C.2. I.C.T.A.A.L : Instruction sur les Conditions Techniques d'Aménagement des autoroutes de
Liaison . SETRA., octobre 1985.
(I.C.T.A.A.L.: Instruction on the Technical Standards for the Development of Connecting Highways.
SETRA, October 1985).
C.4. Carrefours giratoires : evolution des caractéristiques géométriques. Note d'information , série
"circulation-sécurité-exploitation", n° 60. SETRA, mai 1988.
(Roundabout Intersections: Evolution of their Geometric Features. Memorandum in the series "Traffic-
Safety-Operation," No. 60 SETRA, May 1988.)
C.5. Traitement des tourne-à-gauche; les aménagements à faible coût. Note d'information , série
"circulation-sécurité-exploitation", n° 70. SETRA, novembre 1989.
(The treatment of left turns: Low-cost facilities. Memorandum in the series "Traffic-Safety-Operation,"
No. 70 SETRA, November 1989.)
C.6. Déports en carrefours plans sur routes interurbaines – évolution des caractéristiques
géométriques. Note d'information , série "circulation-sécurité-exploitation", n° 71. SETRA,
novembre 1989.
(Displacements on at-grade intersection on interurban roads - evolution of their geometric features.
Memorandum in the series "Traffic-Safety-Operation," No. 71 SETRA, November 1989.)
C.8. A.R.P.: Recommandations techniques pour l'Aménagement des Routes Principales. SETRA.
Août 1994
(A.R.P.: Technical recommendations for the Development of Main Roads. SETRA, August 1994.)
C.9. GIRATION. Définition, calcul, dessin d'épures de gyration [logiciel]. CERTU- SETRA.
(GIRATION. Definition, calculation of turn movement templates (software). CERTU.)
Ñ INVESTMENT EVALUATIONS
1.1. Instruction relative aux méthodes d'évaluation des investissements routiers en rase campagne et
en milieu urbain. D.R., juillet 1995.
(Guideline regarding the methods of evaluation of roadwork investments in rural and in urban areas.
D.R., July 1995.)
108
Ñ EQUIPMENT AND SIGNING
E.1. Instruction interministérielle sur la signalisation routière. Arrêté du 24 novembre 1967, modifié,
parties 1 à 8, edition 1987, et supplements.
(Governmental guideline on road signs. Guideline of November 24 1967, amended, Parts 1 to 8, Edition
of 1987, with Appendices.)
E.2. Instruction interministérielle 82-31 du 22 mars 1982 relative à la signalisation de direction.
(Governmental guideline on directional signs, Mo. 82-31 of March 22 1982.)
E.3. Les dispositifs de retenue- où les mettre? Note d'information série " circulation-sécurité-
exploitation" n° 04. SETRA, février 1986.
(Restraining devices - Where to put them? Memorandum in the series "Traffic-Safety-Operation," No. 04
SETRA, February 1986.)
E.4. Instruction relative à l'agrément et aux conditions d'emploi des dispositifs de retenue des
véhicules contre les sorties accidentelles de chaussée. D.S.C.R. ; mai 1988
(Guideline regarding the upgrade and conditions of use of safety barriers to protect vehicles against
accidental exiting of the roadway. D.S.C.R. May 1988.)
E.5. L'éclairage des carrefours à sens giratoire . Guide technique . SETRA - CETUR; 1991.
(The Illumination of Roundabout Intersections. A Technical Guide. SETRA - CETUR, 1991.)
u ROADWAY
Ch.1. Nf P 98-302. Chaussées. bordures et caniveaux préfabriqués en béton. AFNOR, juin 1992.
(NF P 98-302. Roadways. Curbs and prefabricated concrete gutters. AFNOR, June 1982.)
U.1. I.C.T.A.V.R.U.: Instruction sur les conditions Techniques d'Aménagement des Voies Rapides
Urbaines. CETUR, 1990.
(I.C.T.A.V.R.U.: Instructions on the Technical Conditions for the Construction or Urban Thoroughfares.
CETUR, 1990.)
U.2. Ville plus sûre, quartier sans accidents. Savoir-faire et techniques. CETUR, 1990.
(A safer city, accident-free neighborhoods. Know-how and technologies. CETUR, 1990.)
u CITY BYPASSES
D.1. Conception des déviations d'agglomération : prise en compte de la sécurité. SETRA., juillet
1986. (The design of city bypasses: safety issues. SETRA, July 1986.)
u TRAFFIC
T.1. Temps d'attente et longueur de queues en carrefour interurbain sans feux ( le logiciel OCTAVE°.
Note d'information série "circulation-sécurité-exploitation" n° 44. SETRA, septembre 1986.
(Wait times and queue lengths at interurban non-signalized intersections (with the OCTAVE Software).
Memorandum in the series "Traffic-Safety-Operation," No. 21 SETRA, September 1986.)
T.2. Capacité des carrefours giratoires interurbains, premier resultants. Note d'information série
"circulation-sécurité-exploitation" n° 44. SETRA., août 1987.
(The capacity of interurban roundabouts, preliminary data. Memorandum in the series "Traffic-Safety-
109
Operation," No. 44 SETRA, August 1987.)
T.3. Circulaire relative au recensement de la circulation routière : Annexe technique. DSCR, éd.
annuelle.
(Bulletin on the collection of traffic data: Technical appendix. DSCR. Annual edition.)
T.4. Guide des etudes de traffic interurbain. Guide méthodologique. SETRA, mai 1992
(Guide to interurban traffic studies. Methodological guide. SETRA, May 1992.)
T.5. Les capteurs de traffic routier. Guide technique. SETRA, décembre 1995.
(Traffic sensors. Technical guide. SETRA. December 1995.)
u SAFETY
S.1. Sécurité des routes et des rues. SETRA- CETUR, septembre 1992.
(The safety of roads and streets. SETRA-CETUR, September 1992.)
S.2. Etudes préalables à des interventions sur l'infrastructure. Guide méthodologique, SETRA,
septembre 1992.
( Preliminary studies for infrastructural projects. Methodological guide. SETRA, September 1992.)
S.3. T. Brenac. Accidents en carrefour sur routes nationales. Modélisation du nombre d'accidents
prédictibles sur un carrefour et exemples d'application. Rapport INRETS n° 185. INRETS, août
1994.
(T. Brenac. Accidents at intersections on national roads. Modelling of the predictable number of
accidents and examples of applications. INRETS Report No. 185. INRETS, August 1994.)
S.4. Accidents en carrefours: utilisation des modèles donnant le nombre moyen d'accidents
prévisible. Note d'information série "circulation-sécurité-exploitation" n° 116. SETRA, mars 1998.
(Accidents at intersections: the use of models to predict the average number of accidents.
Memorandum in the series "Traffic-Safety-Operation," No. 113 SETRA, March 1998.)
S.6. La sécurité des giratoires en rase campagne. Club d'échanges d'expérience sur les routes
départementales – région Normandie, décembre 1995.
(The safety of rural roundabouts. Discussion club on departmental roads - Normandy Region, December
1995.)
S.7. Aménagement des carrefours en rase campagne et sécurité. DDE de Seine-Maritiime – CETE
110
de Normandie-Centre – SETRA, décembre 1996.
(The construction and safety or rural intersections. DDE de Seine-maritime - CETE de Normandie -
Centre - SETRA, December 1996.)
111
APPENDICES
112
113
APPENDIX 1: THE SAFETY OF AT-GRADE
INTERSECTIONS
Outside of urban areas, approximately one fifth of all accidents occur at at-grade
intersections, mostly on standard (non-roundabout) intersections. The levels of risk vary
greatly according to types of intersections (for identical traffic configurations), due in part to
their differences in operating principles. Furthermore, the lowest levels of risk must be
associated with the highest exposures.
Table 1. The number and seriousness of accidents at at-grade intersections on main roads outside of
urban areas.
2
Number of injury accidents
where: where:
Application: at-grade intersection equipped with STOP or YIELD signs, Application: roundabout intersection where the total incoming traffic
where the main traffic(*) ranges between 3,000 and 25,000 v/d, and ranges between 3,200 and 40,000 v/d.
the secondary traffic (*) ranges between 500 and 8,000 v/d.3
The values of the annual ratios for the calculation of Fc are as follows (in accidents/108 v.km):
1991: 18.8; 1992:17.9; 1993:17.0; 1994: 15.9; 1995: 15.0; 1996: 14.1
2
See the SETRA Memorandum (Traffic/Safety/Operations Series) "Accidents at intersections: the use of
models to predict average accidents rates" (to be published in 1998), whose results are excerpted from the
INRETS report No. 185 (T. Brenac; 1994).
3
In the absence of any other formula, its use can be admitted (as a first approximation) for a wider range:
between 2,000 and 40,000 v/d on the main axis, and 0 to 13,000 on the secondary axis.
4
There is no adjustment coefficient available in relation to a studied time frame specific to a roundabout.
The coefficient used for standard at-grade intersections can however be used: this will result in an excessive
approximation of the number of predictable accidents.
114
Seriousness of Accidents
Serious injuries: 45 per 100 accidents Serious injuries: 33 per 100 accidents
Light injuries: 126 per 100 accidents Light injuries: 106 per 100 accidents
The traveling speed of the direction with the right of way is one of the main reasons
behind the inherent lack of safety of these types of intersections. Configurations that
accommodate high speeds on the legs with the right of way are those that include more than
one direct lane per direction or special right-turn lanes.
Intersections that are complex, atypical, ambiguous (left turn via the right, for
example), or that offer excessively fluid trajectories to vehicles obligated to yield (Y-shaped
intersections, for example), are not very safe. In these instances, lack of visibility, usually
associated with lateral screens or with objects on the path (convex horizontal alignment), is
often a cause of accidents.
Criss-crossings
Criss-crossings involve collisions between vehicles on the lane with the right of way
and vehicles on the secondary lane in the process of crossing or making a left turn. There
are several phenomena that can lead to these types of accidents: inaccurate perception of
the intersection or of the need to yield the right of way on the secondary road, poor visibility,
misunderstanding of the configuration and the workings of a complex intersection, the large
width to be crossed.
The incorporation of channelization islands on the secondary legs will help drivers
perceive the intersection, understand that they will have to yield and reduce their speeds.
Channelization islands also reduce the risk of criss-crossing accidents by 30 to 50%.
These accidents involve vehicles performing a left turn on the main lane that are
impacted by a vehicle moving straight through the same road. Rear collisions occur if the two
vehicles are moving in the same direction, and frontal or comparable collisions if they are
traveling in opposite directions.
115
A left turn refuge (possibly very short in length) is a very effective solution to reduce
rear collisions. In the case of T-shaped intersections or their accesses, the shoulder on the
opposite side of the secondary leg can be used by the vehicle coming from the rear, that was
surprised to see the vehicle making the left turn, to make an emergency avoidance maneuver
by shifting to the right.
However, this should only be seen as a partial solution, less effective than a
dedicated left-turn slip lane.
Raised islands (with traversable curbs, painted in white and featuring the J5 marker at
the front, but usually unlit) provide greater safety than merely painted islands, by offering an
improved perception of the intersection and a real protection to left-turning vehicles. These
types of accidents can thus be reduced by 50 to 80%. Collisions with raised islands are still
very rare.
Ñ 3. ROUNDABOUT INTERSECTIONS
In general, roundabouts are quite safe (See 1.). However, some configurations can markedly
reduce their safety levels. This is the case for oblong roundabouts, or pseudo-roundabouts
and, to a lesser degree, for roundabouts made more complicated by the addition of transition
lanes. Furthermore, large roundabouts are not as safe as smaller roundabouts.
The major types of accidents that occur on roundabouts involve entering vehicles
losing control, landing on the central island, collisions at the entries and, to a lesser extent,
losses of control on the circulating roadway.
Losses of control
Loss of control upon entering a roundabout is the most common and by far the
most lethal type of accident occurring on roundabouts outside of urban areas. Among the
contributing factors to these accidents we often find a poor perception of the approach to the
facility (strong reversing curve, for example), high approach speeds, as well as other factors,
such as the absence of a clear break in the outline of the path of the road (i.e., an impression
of continuity produced by trees or luminaires aligned on one the roundabout's legs). Most of
these losses of control only cause property damage, but the consequences are always more
serious if the central island features hazardous obstacles or devices.
Most roundabouts experience these types of accidents. Risks are increased in rural
areas, at night--although there apparently aren't more nighttime accidents on unlit
roundabouts than on illuminated roundabouts--and especially during the first few months after
the commissioning of a new facility.
5
This constraint must be located at the leg's termination onto the circulating roadway, and not upstream:
otherwise, arrangements such as "reversing curves" will often reduce safety.
116
Losses of control on a roundabout's circulating roadway are often associated
with excessively large entry radii, wide entries (2 lanes or more), or with non-circular central
islands. Many of these accidents involve overturned trucks.
When motorcycle riders lose control of their vehicles, poor road maintenance (loose
gravel, oil and gasoline spots, etc) should be added to the aforementioned causes.
Loss of control accidents upon exiting roundabouts remain very rare. They are
most often caused by insufficient exit radii. Once in a while, these radii are responsible for
frontal collisions with approaching vehicles (especially when the intersection has no splitter
islands).
Collisions
Nowadays, accidents involving conflicts over the right of way remain common in spite
of a generalization of the so-called "yield to the circulating roadway" rule, in effect for quite
some time now. The major causes of collisions are high entry speeds and poor visibility in
the area before the transverse YIELD line (such as obscuring vegetation on the splitter
islands). Entries with wide radii or that are excessively tangential are often blamed. In the
case of multi-lane entries, there are some configurations of vehicle positions that can create
moving masks.
- criss-crossing accidents at the exits, mostly on 2-lane exits where the amount of
traffic does not justify such a wide size;
- frontal collisions on the circulating roadway (with a vehicle traveling the wrong way).
Although deliberate violations are hard to prevent, these types of accidents are often
caused by the misunderstandings generated by highly complex facilities (inclusion of
special lanes, for example).
Two-wheel vehicles are especially vulnerable to refusals to yield at the entries and, to
a lesser extent, to other types of collisions. All the measures aimed at increasing a facility's
capacity (and accessorily its speed) are harmful to them. In fact, no modification specifically
intended for them is truly effective.
Pedestrians are not especially affected by roundabouts; they are, however, particularly
disadvantaged on large roundabouts. Most pedestrian accidents occur during crossings of
wide (multi-lane) and fast moving entries or exits.
117
APPENDIX 2: CAPACITY AND WAIT TIMES
Ñ 1. OVERVIEW
The calculation of the capacity (as well as the wait times6) is generally based on rush
hour traffic. Regular rush hour peaks should preferably be used over exceptional peaks.
Indeed, choosing exceptional peaks may lead (for the sake of just a few hours of saturation
per year) to solutions (capacity enhancements, improved roundabout features) that will
markedly increase the facility's cost and footprint, and could adversely affect its safety. It is
also useful to factor in reverse flow peaks in the case of some configurations that feature
swinging hourly peaks and, as the case may be, extreme seasonal variations.
The regular traffic of the so-called "peak thirtieth hour" is sometimes used to
determine the size of an intersection. But applying this method requires prior knowledge of
the distribution of hourly flows over a period of one year. The axes involved must therefore be
fitted with permanent recording stations (which are generally only located on major roads). In
fact, the selection of the reference period will depend on the developer's design policy and on
the level of the service he wants to offer the user.
Regardless of the method and the type of intersection, the calculation of the capacity
and of the wait times must take into account the main and secondary traffic flows: it is
therefore important to know the traffic flow distribution translated in the form of an "origin-
destination" matrix.
Other arrangements aimed at improving an intersection's capacity will also reduce its
overall level of safety. For example, any arrangement encouraging the side-by-side storage of
several vehicles coming out of secondary legs (via widening of the entry lanes or their
destinations through increased entry radii) should be excluded due to the mutual hindrances it
would produce.
On a "R" type single-lane road, the location of a dedicated left turn lane to exit the main
lane is the only arrangement that can be recommended in order to decongest the main axis
when the flows are heavy enough (100 v/d) for a T-shaped intersection, 200 v/d for a four-way
intersection). This will not have a major impact on the capacity of secondary legs.
6
Out of concern for creating a facility that is suited to the existing traffic, and not as part of a cost evaluation,
which is supposed to the add up all the lost times in all directions over an entire year.
118
• 2.1. THE "MANUAL" METHOD
In the absence of an updated manual method of estimation of the capacity and wait
times at a standard at-grade intersection, one can use, to make a preliminary estimate, the
old guidelines developed in the late 60s. By applying probabilities, these guidelines come in
the form of a series of formulas and nomographs. As a reference on this subject, we
particularly recommend the document entitled "Carrefours sur routes interurbaines, 1ère
partie: carrefours dénivelés" (SETRA; 1976) [Intersections on interurban roads, Part 1: grade-
separated intersections].
When a more accurate estimate is required, especially when a delicate balance with
another type of intersection must be achieved, we recommend using the OCTAVE software.
The OCTAVE software (SETRA; 1998) is a simulation program dealing with the
capacity of unsignalized intersections. The simulation is based on a discrete and per-event
microscopic model, displaying the behaviors of the vehicles in the various queues that may
develop. The shape of the intersection and the traffic rules can be specified with great
accuracy.
Thanks to OCTAVE, one can estimate the capacity, the wait times and the length of
the queues of the various directions of traffic. However, one must bear in mind that the
estimates of the wait times and the lengths of queues are highly assumptions-dependent and
should only be seen in terms of orders of magnitude.
Ñ 3. ROUNDABOUT INTERSECTIONS
The geometric design of a roundabout will have an impact on the capacity of its
various legs. Some geometric features can be mutually combined, positively or negatively.
These are the most important features:
- the width of the entry (le) (and not the number of entry lanes); the impact of a widening
of the entrance is however restricted by the width of the circulating roadway (la);
- the width of the circulating roadway (la), that must be restricted out of safety concerns;
- the width of the splitter island (li ), which influences the friction caused by the outgoing
traffic from the leg with the island;
- the radius of the central island (Ri ), which can impact capacity positively or negatively,
particularly when the values are small.
119
The impact on capacity produced by the angle and the radius of the entry (Re) is
generally negligible if one remains within acceptable safety limits. Moreover, the size of a
roundabout only has a minimal impact on the capacity of its entries.
This method can be applied to roundabouts in rural or suburban areas with central
islands with a radius (Ri ) equal to or greater than 15 m, and with single-lane entries. Its
results are quite accurate (it is actually quite conservative because the measurements that
were used to establish it are now quite old).
1. Determine the entering flows Qe, the exiting flows Qs, and the turning flows Qt, based
on the origin-destination matrix (in uvp/h, applying a coefficient of 2 for trucks, and of
0.5 for two-wheel vehicles);
2. Determine the corresponding exiting flow Qs' according to the width of the splitter
island:
3. Determine the Qs obstructing traffic, based on Qt, Qs' and the width of the circulating
roadway (la);
(C - Qe) / Qe
The GIRABASE software takes into account all the factors that have an impact on
capacity. Its algorithm was developed after numerous observations of existing roundabout
intersections. It can come up with results that are markedly different from those produced by
the aforementioned "quick" method, especially when it is applied to roundabouts whose sizing
deviates from old, and rather broad, standards.
120
121
APPENDIX 3: GEOMETRIC DELAY AT
INTERSECTIONS
Table 2 -- Estimate of the geometric delay according to the type of intersection and maneuvers,
light vehicles.
Geometric delays for trucks should be increased by 75% compared to the figures for
light vehicles.
To calculate the aggregate geometric delay (RGC) over a period of one year, the
following simplified formulas can be applied:
where:
- traffic flows are expressed in u.v.p. with a coefficient of 1.75 for a truck;
- Qech is the exchange TMJA between the two axes, defined as the sum of the turning
traffic directions;
122
123
APPENDIX 4: ESTIMATE OF THE V85 (85
PERCENTILE) SPEED
"In order to take into account the actual traveling speeds of users, and
in accordance with international standards, the V85 speed is generally applied,
namely the speed under which 85% of all users drive, under fluid (unimpeded)
traffic conditions. This speed can be estimated on the basis on the functions
or the nomographs below, which translate the results of studies on the relations
between geometry and speed. The "DIAVI" software can also be used to
estimate the traveling speeds at each position of a facility."
40
10
20 5
R(m)
0 200 400 600 800 1000
Figure 2 -- The V85 speed as a function of the ramp (> 250 m).
V(km/h) V(m/s)
140
35
120
V85 = 120 - 0,31p²
30
100
V85 = 102 - 0,31p² 2x2v
25
3v et 2v
80 (6 et 7m)
V85 = 92 - 0,31p² 20
2v (5m)
60
15
40
10
20 5
P(%)
0 2 4 6 8 10
124
125
APPENDIX 5: MEASUREMENT OF THE VISIBILITY
CONDITIONS BY THE "STOPWATCH
METHOD"
This method can be applied quite easily--all it requires is an operator and a stopwatch-
-and be completed quite rapidly (seldom more than 3 to 15 minutes per intersection, if the
traffic on the main axis is low).
Stopwatch in hand, the operator directs his eyesight to the edge of the right shoulder
of the secondary road (such as the position of a driver in a vehicle parked on the shoulder), at
a height of 1 m, and with a 4 m setback from the STOP bar, or 15 m behind the YIELD line.
The operator measures the elapsed time between the moment he sees a vehicle on
the main road and the moment it reaches his position, for 12 unimpeded light vehicles
(moving straight through).
The times are ranked in increasing order; the third time (t3) is then compared to the
recommended crossing time (t) for the configuration involved (the values of t are listed in
Table No.1 of Chapter 2).
Ñ PRACTICAL CONDITIONS
While observing, the operator can refrain from measuring times if the conditions of
visibility are clearly adequate (visibility of more than 300 m, for example) or inadequate
(visibility of less than 100 m).
Whenever t3 is less than t for one of the sight triangles, the intersection can be found
not to meet the visibility requirement; the operator can therefore limit himself to the most
unfavorable measurement if he is able to measure it a priori. Moreover, from a practical
standpoint, the measurement of the visibility times can be stopped after the third insufficient
time.
If the visibility is satisfactory at the point of observation (at 4 or 15 m), one must
ensure that, as one approaches the line, no visual obstacle will markedly disrupt the sight
triangle that was created.
126
127
APPENDIX 6: DESIGN OF THE CURBS
Ñ OVERVIEW
The curbs mostly include the raised splitter islands and, in the case of a roundabout,
the central island and (as the case may be), the edges of the entries and exits of the
circulating roadway.
Furthermore, all special operating conditions that can occur must be taken into
account, especially maintenance work (accessibility to the central island, for example),
wintertime practicability (high curbs could impede snowplows) and drainage requirements.
Note: The design of the islands per se is covered in the main section of this document, in
Chapters 2 (§ 2.5.6. & 3.2.3.) for the splitter islands, and 3 (§ 3.1.4. & 3.2.) for the
central islands of roundabouts.
Furthermore, a good level of visibility must be provided at the fronts of islands, as well
as sufficiently large (in height and width) advance directional markings. To improve nighttime
visibility8, the extending sections should be painted in white or dyed concrete curbs should be
used.
Figure 3 -- Detail of the construction of the curbs of splitter islands (embedded curbs)
0,07 0,07
h v =0,03 h v ≤≤0,06
7
For a small island on a very secondary leg (See Chapter 2.3.4.), this height should not exceed 3 cm.
8
In rural areas, intersections generally do not require illumination.
128
Embedded curbs (Type 12 or 14, for example) are preferred on splitter islands,
because they are more resistant to the possible demands made by turning trucks. Grouted
or extruded curbs could be used, as long as they are not too high to the eye (hv) (See above).
Ñ EDGE CURBS
For standard at-grade intersections, the inclusion of edge markings is generally
unnecessary (except when the continuity of a sidewalk must be ensured); stabilized lateral
stripes which may be coated over a width of 1 m9 are enough. The placement of edge
curbs therefore pertains mostly to rural roundabouts. Although not always necessary, edge
curbs are generally recommended. One should be sure to:
- maintain, as the case may be, the continuity of crosswalks by implementing the
sidewalk via T type curbs (which are normally reserved for these types of cases).
They should be lowered at a perpendicular angle if there are marked crosswalks on
the road;
- maintain the integrity of the shoulder10 (to prevent the digging of ruts by trucks) by
including, if there is no curb, a lowered coated strip 2 m wide, in connection with a
widening of the structure at the entrances and exits. When the curbs are (semi-
)traversable, these arrangements can also be useful;
- avoid interrupting the curb around the circulating roadway, due to the short distance
between one entry and the exit of the next leg (for standard sized roundabouts);
- limit their height to a maximum of 14 cm, although less hazardous shoulder curbs, of
which only 6 cm are visible, are preferable;
- take into account water drainage, by creating a gutter or, if there is no curb, a swale
between the pavement and the shoulder;
- take into account (in parts of the country where this may apply) constraints connected
with wintertime practicability, by limiting the height of the curbs, to avoid impeding
snowplows.
At the entries of the roundabout, it is recommended to have the edge curbs begin (as
applicable):
- in normal cases: at the H distance (height of the construction triangle of the splitter
island = Rg) of the YIELD line;
- in the case of a leg with 2X2 lanes, at the front of the splitter island (widening of public
transit lines and end of the railings);
- when the entry includes inflected trajectories (See Chapter 3.2.7.): at the end of the
alignment to the right.
9
Up to 2 m, for some entry and exit lanes of divided intersections on 2X2 lane roads.
10
Experience shows that in some areas, curbs feature tire marks or signs of unintentional crossings.
129
Furthermore, the starting point of the curbs at the entries should be lowered or offset.
Edge curbs at the exits are generally interrupted at the same point as the entry edge
curbs, but it is unnecessary to extend them beyond H.
130
131
NOTES
The document "Interurban junction design – grade intersection" is a
technical guide for designing grade intersections on main roads outside urban
areas.