Layout of 2+1-Routes in Germany - New Findings: Mailto:irzik@bast - de
Layout of 2+1-Routes in Germany - New Findings: Mailto:irzik@bast - de
Marco Irzik
mailto:[email protected]
This paper includes altogether 10 pages. One page includes the title page and another one the abstract with 190
words. The body of paper with 4,559 words including the references, 1 table and 9 figures (= 10 x 250 = 2,500
words) is presented on 8 pages. The total count of words is 7,059.
Dr.-Ing. Marco Irzik
ABSTRACT
The aim of the presented dissertation (Irzik, 2009) was to determine an optimal length for two-lane sections
within 2+1-routes in dependence on various parameters. A correlation between traffic volume and share of
incomplete dissolving processes has been used in order to define the necessary length of a two-lane section. A
maximum length – depending on traffic volume and share of heavy goods vehicles – is also suggested. This
restriction shall prevent that the length of a two-lane section and with it the length of the one-lane section of the
opposite direction will increase excessively. Otherwise negative effects on platooning would have to be
expected. Finally a minimum length according to traffic safety is recommended for two-lane sections within
2+1-routes. This recommendation is based on the relation between the share of vehicles changing from the
passing to the right lane on the last 200 m of a two-lane section related to the total number of all passing
processes and the number of observed conflicts while lane changing on the last 200 m. The overall result of the
presented dissertation is a simplified method for determining the optimal length of two-lane sections within 2+1-
routes.
Dr.-Ing. Marco Irzik 1
1. INTRODUCTION
A 2+1-route is characterised by a single carriageway three-lane cross section. It consists of two-lane sections
with a passing lane and one-lane sections before (= feeder sections), behind and besides each passing section
(see figure 1). While the traffic flow in a passing section has the possibility to pass slow vehicles, the traffic flow
in the opposite direction is not allowed to pass. At the same time vehicle platoons are arising in one-lane
sections. Therefore, the passing lanes have to alternate in each direction along the whole 2+1-route. If vehicle
platoons are not dissolved at the end of the passing sections, the level of service impairs and negative impacts on
traffic safety will have to be expected.
In recent years a number of research projects dealt with traffic safety of and investigations into traffic flow on
2+1-routes (i.e. Weber, Löhe 2003; Brannolte, Baselau, Dong 2004; Gattis, Bhave, Duncan, 2006). Nevertheless
up to now there have been no comprehensive findings concerning passing process and dissolving of platoons in
dependence on marginal conditions relating to design and operation others than model based considerations.
Therefore the prime aim of the presented dissertation (Irzik, 2009) was to develop a procedure in order to
determine the optimal length of passing sections on 2+1-routes on the basis of empirical studies. The solution to
a conflict of objective hereby formed a special problem: On the one hand passing sections must be long enough
for all entering platoons to be dissolved at its end. On the other hand they must not be too long since this would
have an adverse effect on the length of the opposing one-lane section and thus on platooning. Apart from
determining the dimension of passing sections based on platoon dissolving, aspects of traffic safety were also
included in the investigation. The research project "Passing process on 2+1-routes" (Friedrich, Dammann, Irzik,
2005) commissioned by the Federal Highway Research Institute constitutes the basis of the presented
dissertation (Irzik, 2009).
2. METHODOLOGY
The summary of current scientific knowledge concerning this topic formed the initial step. After a description of
the choice of investigated sections, the relevant parameters of traffic flow and passing process as well as
platooning and dissolving of platoons were introduced. Following the description of the developed and applied
investigation methodology for the empirical investigations, the procedure of the analysis of empirically obtained
data as well as the used methods/techniques was listed. As a rule the obtained results were represented
statistically, and if possible compared to results of previous studies. At first the speed and passing process (here
especially the beginning and the end of passing) and platooning ahead of as well as within 2+1-routes were
analysed. Finally different approaches to the determination of the length of passing sections required for
dissolving platoons on 2+1-routes were tested.
3. ANALYSIS OF LITERATURE
All studies dealing with traffic safety on 2+1-routes during the past few years show that 2+1-routes are
characterised by a higher traffic safety, expressed in terms of accident cost rate, in comparison to other single
carriageway cross sections outside built-up areas. However, due to the comparatively high speed level the
average accident severity is slightly higher in comparison to narrower two-lane rural roads.
Weber, Löhe (2003) couldn’t conformed safety concerns regarding the opening of 2+1-routes for mixed traffic.
But it must be kept in mind that Weber, Löhe observed small numbers of agricultural vehicles respectively a
marginal share of bicycle traffic on these roads and resulting low accident figures only. Furthermore especially
the simulation results obtained in the development of a design procedure for the up-date of the German Highway
Capacity Manual (HBS, FGSV, 2001) proved an explicit negative influence of slow moving traffic on the quality
of traffic flow (see Baselau, 2006, Brannolte, Baselau, Dong, 2004).
The work of Weber, Löhe (2003) contains a hint at a minimum length of passing sections for reasons of safety.
In their study, Weber, Löhe (2003) recommend a minimum length of 1,000 m for passing lanes within 2+1-
routes with mixed traffic. The accident analysis within the frame of their investigation showed a
disproportionally high number of accidents on sections shorter than 1,000 m compared to sections with lengths
exceeding 1,000 m.
Dr.-Ing. Marco Irzik 2
Weber, Löhe (2003) also found out that an adapted alignment in combination with partially grade separated
junctions (see figure 2) does not only have a positive influence on a homogeneous speed level but also on a
lower accident cost rate. Palm, Schmidt (1999) already stated that – like for accidents on sections between
junctions – the accident rate and accident cost rate for accidents at junctions are lower on 2+1 routes than on
single carriageway two-lane cross sections. This can most likely be explained by the fact that the number of at-
grade junctions on 2+1-routes is smaller than on two-lane rural road cross sections. The 2+1-route on the Federal
Road B10 near Landau investigated by Kölle (1999) contained one partially grade separated and one partially at-
grade junction (see figure 2). Despite a speed limit of 70 km/h in the junction area of the partially at-grade
junction the accident cost rate for this junction area is more than twice that of the partially grade separated
junction. As considering an almost equal accident rate, this result points at a substantially higher accident
severity in the partially at-grade junction. This was due to the high speeds within the junction area, which were
observed in defiance of the speed limit. Hence the predominant number of accidents at the partially at-grade
junction occurred in the junction area of the major road B10. At the partially grade separated junction, however,
only few accidents had to be registered. These were mainly damage only accidents that were caused by lane
changes in the merging area with lane additions. Main accident locations were the at-grade parts of these
partially grade separated junctions at the minor road. Kölle formed so called "composite systems" in order to
determine the influence of junctions on the adjoining sections. A composite system consisted e.g. of the junction
itself and 500 m of the adjoining sections. Therefore composite system no. 1 includes the partially at-grade
junction, and no. 2 the partially at-grade junction. The comparison between these both systems showed that for
the latter, the accident cost rate of system no. 2 amounts to only about a fifth of the accident cost rate determined
for the system no. 1. Kölle's investigation included two further junctions, a partially grade separated junction and
an at-grade junction, on a 2+1 route (B49). The accident cost rate at this partially grade separated junction was
slightly higher than at the partially at-grade junction on the B10. But it was only half of the rate of the at-grade
junction along the same 2+1 route. At the partially grade separated junction on the B49 there were also only few
accidents caused by entering or exiting vehicles. Like the partially grade separated junction on the B10, that one
on the B49 is characterised by a high level of safety as compared to other partially grade separated junctions
investigated by Kölle.
FIGURE 2 Selected junction types according to the new guidelines for the design of rural roads (RAL)
currently being developed (FGSV, 2007)
In terms of the existing findings regarding passing process and dissolving of platoons it had to be assessed that
they were lacking adequate empirical verification. So the procedure developed by Roos (1989) is based on
theoretical considerations. For this reason there were still uncertainties in the determination of optimal section
lengths in dependence of traffic parameters. Therefore it seemed urgently necessary to broaden the knowledge
about passing process and dissolving of platoons. It was found out that different definitions of platoons from
literature exist. In this investigation a vehicle was assigned to a platoon if the gross time gap to the vehicle ahead
or the platoon leader was below 3 s (following Roos and Brannolte, Baselau, Dong). In this case the platoon
leader's speed must not exceed 90 km/h. Further is agreed that the platoon leader himself does not belong to the
platoon. As a supplementary condition within this investigation it was agreed that compared to Roos und
Brannolte, Baselau, Dong the vehicle must pass the platoon leader in the passing section. Otherwise it was
assumed that vehicles which do not pass the platoon leader have already reached their desired speed in the feeder
section and are therefore not obstructed in their movement. Because of the chosen investigation methods, the
afore mentioned studies could not comply with this condition.
Figure 3 Start of the two-lane passing section with Figure 4 Start of the two-lane section with an
lane addition at partially grade separated junction uncritical transition
BP = bypass
adt [veh/24]
investigated
Length [m]
Transition
Lfeeder [m]
Road No.
dt = direction towards
Position
section
Number
Number
50% 20 50%
600
40% 40%
15
400 30% 30%
10
20% 20%
200 5
10% 10%
0 0% 0 0%
1 2 3 4 5 6 7 8 9 10 11 12 13
1
2
3
4
5
6
7
8
9
10
11
12
13
>13
Figure 5 Number of observed platoon lengths Figure 6 Number of 85 % platoon length (n=111)
(n=3.163)
1.400
platoon dissolving distance [m]
1.200
.
1.000
Weg zur Pulkauflösung [m]
800
600
400
200
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Pulklänge [Fz]
platoon lenght [veh]
Figure 7 Platoon dissolving distance in dependency of the platoon length (Box-Whisker-Plots) (n=1.384)
However, a procedure for the determination of the section length on the basis of empirically monitored platoon
dissolving distances (see figure 7) also shows some weak points. Such a procedure also developed in the scope
of this study, based in general on the theoretical deliberations of Roos (1989). It has the decisive disadvantage of
not taking into account the platoons that were not dissolved. Furthermore in this procedure it is necessary to
Dr.-Ing. Marco Irzik 5
consider influence factors which result from the design/lay-out of the transition at the beginning, the (absolute)
position or the length of the passing sections as well as the traffic volume, the share of HGVs and probably a
prohibition of overtaking for HGVs. All these factors make this procedure much more difficult.Detached from
Roos procedure a different approach was pursued, which, in addition, is much easier to handle. In this approach
the rate of platoons that were not dissolved are taken as a level of service in the determination of the passing
section length.
In the analysis of platoon dissolving there were some platoons on individual investigation sections that could not
be dissolved fully on the current lengths of these sections. Based on the total quantity of platoons during the
investigation period, the rate of not dissolved platoons can be obtained. This rate is especially dependant on the
traffic volume but also on the section length. So, if a level for the rate of not dissolved platoons is pre-set, the
required section length can be determined in dependency of traffic volume based on this correlation.
Influencing factors arising from the length or position of the overtaking sections are implied in this procedure.
Since all but one section, where platoons were not fully dissolved, were only found on investigated sections with
a lane addition, there is no need to explicitly provide for the way the platoon leader needs for merging in form of
a supplement to the length a platoon needs for dissolving. From the insights obtained into the platoon leaders'
lane changing behaviour it appears admissible to adapt the required section lengths for the dissolving process in
case of a beginning with an uncritical transition by a reduction. This is because on sections with lane addition the
platoon leaders generally merged after 125 m, while on those investigated sections that begin with an uncritical
transition the first vehicle to pass the platoon leader merged at the latest after 50 m in 85% of all cases. So the
reduction can be 75 m.
No hints regarding e.g. a maximum platoon share which should be avoided according to aspects of safety can be
gained from a comparison of parameters of platooning and the safety relevant parameters "crossing ghost island
at the critical transition" respectively "sum of (weighted) conflicts on the last 200 m of a pssing section". With
regards to the determination of an optimal length of passing sections, the results received in the analysis of
platooning on feeder sections within 2+1 routes were used. It appeared that starting from a platoon share of
34.5 % (according to the definition used within the frame of this investigation) platoons with at least 7 vehicles
in any platoon must be expected. It was shown that 15 % of all platoons of 7 vehicles required more than 900 m
before they were dissolved. However, 95 % of the 1,384 platoons where the dissolving process was observed
until the end needed less than 900 m to dissolve. Furthermore the frequency distribution of the 85 %-platoon
lengths revealed that considering peculiarities of individual feeder sections a maximum 85 % platoon length of 7
vehicles can be assumed. As a result both aspects together are proposed to avoid platoons longer than 7 vehicles
respectively the relevant share of platoons by restricting the length of the feeder section and correspondingly the
length of the passing section in the opposite direction. Since a correlation between the share of platoons, the
traffic volume, HGV-share and the length of the feeder section could be proved, this specification serves to
determine a maximum length of a passing section. Due to the discovered effects an increasing HGV-rate leads to
a decreasing maximum length (see figure 8). This means that several shorter sections should be preferred to
fewer longer sections for higher HGV-rates. In principle this coincides with the recommendations regarding an
optimum length contained in the RAS-Q 96 (FGSV, 1996) derived from the results of the project group
"Intermediate cross sections" (Brannolte et. al., 1992).
6. RECOMMENDATIONS
6.1 General
Apart from recommendations for the determination of the optimal passing section length within a 2+1-route
gained from this investigation, hints will be provided for cross section design and choice of junction type which
are partially based on findings from previous studies but which have also been derived from available
investigations in the frame of the dissertation.
2+1-routes as defined by the RAS-Q 96 (FGSV, 1996). A traffic volume of 1,300 vehicles per hour and direction
must be considered as upper limit.
With a practicable procedure for the determination of the optimal passing section length in mind, the
specification of a quality standard for the number of not dissolved platoons is suggested. In figure 8 the
admissible respectively recommended passing section lengths are depicted depending on
• traffic volume,
• position within the 2+1-route (Pos.),
• HGV-rate and
• share of not dissolved platoons .
In order to limit the lengths of passing sections with regards to platooning in the one-lane sections of the
opposite direction, a platoon share of 34.5 % according to the platoon definition applied in this study is
recommended as "to be avoided" (see number 5). For such a platoon share, platoon lengths of more than 7
vehicles must be expected. The investigation did, however, show that platoons of such a length are quite rare but
often characterised by extremely long platoon dissolving distances as compared to the total of observed platoons.
For this reason it seams to be admissible to limit the length of passing sections according to the chosen criteria in
order to avoid oversizing.
1.800
0% 2% 4%
1.700
[m] section
max. L (HGV-rate=5%)
max. (SV-Anteil=5%)
Lmax = 1.600 m
.
1.600 max.
max. L (HGV-rate=10%)
(SV-Anteil=10%)
min. L (Pos. 2)
1.300
min. L (Pos. 3)
1.200
Lmin = 1.070 m
Länge eines
1.100
Length (L)
1.000
900
800
400 500 600 700 800 900 1000 1100 1200 1300
qmax [Kfz/h]
max. traffic vol. [veh per h and direction]
Figure 8 Determination of lengths of passing sections which begin with a lane addition at partially grade
separated junctions
A rate of 32.5 % of vehicles, changing from the passing to the right lane at the end of a passing section, is seen to
be critical and therefore must "be avoided" in the determination of the minimal length of passing sections
derived from safety considerations. This specification is motivated by the fact that even though only for rates
above 45 % an increasing number of (weighted) conflicts per vehicle had to be observed the range between 20 %
and 45 % was not covered by the investigation. At a rate of lane changes at the end of 32.5 % a (weighted)
number of 0.036 conflicts within 30 minutes per vehicle must be expected theoretically. This limit is considered
to be admissible because in approx. one third of the 21 observed 30 minute intervals less (weighted) conflicts
within 30 minutes per vehicle occurred.
For sections beginning with an uncritical transition a reduction of about 75 m of the required section length for
complete platoon dissolving determined according to figure 8 appears to be admissible due to the results
obtained in the analysis of platoon leaders and the resulting differences with regards to the start of platoon
dissolving (see number 5). In these sections, also the maximum length of a passing section can be 200 m longer
because only the ghost islands located at the transition areas must be added to the length of a passing section to
Dr.-Ing. Marco Irzik 7
reach the length of the one-lane section of the opposite direction. For sections that begin with a lane addition at a
partially grade separated junction the junction area has to be considered in contrast to the comparatively short
ghost island at an uncritical transition. This area was calculated with approx. 200 m.
Figure 8 shows as an example the determination of the required length (Lreq) for a passing section under
consideration of the maximum length (Lmax) and the recommended minimum length (Lmin) for passing sections at
position 1 with a maximum traffic volume of 700 vehicles per hour and direction and a share of HGV of 10 %.
Figure 9 RQ 15,5 according to new guidelines for the design of rural roads currently being developed
(RAL, FGSV, 2007)
6.4 Junctions
Since among the routes investigated by Weber, Löhe (2003) those with a relatively straight alignment and
partially grade separated junctions displayed the most favourable accident cost rates and an even speed level, a
high design standard should be intended for 2+1-routes wherever possible. It can be deduced from this demand
that, if possible, uncritical transitions should be positioned in such a way that there will be a lane addition in
junction areas for the entering traffic. There were no noteworthy disruptions of traffic flow and traffic safety
observed by Weber, Löhe for this junction layout.
Work performed by Kölle (1999) supports the construction of partially grade separated junctions along 2+1-
routes recommended by Weber, Löhe. Kölle recommends implementation of partially grade separated junctions
wherever high travel speeds should be achieved for simulaneous short junctions distances and high junction
Dr.-Ing. Marco Irzik 8
volume. Since 2+1-routes in general belong to road category LS I (see FGSV, 2007) high travel speeds should be
guaranteed. Simultaneously high junction volumes prevail on 2+1-routes.
The raised concerns that those "high" speeds will be "carried" from the links into the partially grade separated
junction cannot be supported by the analyses of speed behaviour which were carried out within the frame of this
investigation. First there must be stated that traffic reaches the (partially grade separated) junction only following
a one-lane section, not following a passing section. For one-lane sections (inside and outside 2+1-routes) the
results showed that excessive speeds had to be observed in only a minor number of cases and the legal speed
limit was only violated in a range similar to that one to be found on "ordinary" two-lane rural road cross sections.
Furthermore the results derived from the investigation by Kölle (see number 3) showed that partially grade
separated junctions on 2+1-routes hold a much higher safety level compared to the safety level of other junction
types on 2+1-routes but also compared to other partially grade separated junctions on two-lane rural roads.
7. CONCLUSION
As a major contribution to the advancement of knowledge a practical procedure for the determination of the
optimal length of a passing section on 2+1-routes could be developed within the frame of the dissertation
presented in this paper. In contrast to the procedure for a determination of a necessary length with regards to the
dissolving of platoons as known from literature (see Roos, 1989), it also contains recommendations concerning a
minimum length derived from safety considerations as well as an upper limit in order to avoid excessive
platooning in the opposite direction. The newly developed procedure is not based on theoretical deliberations but
on comprehensive empirical studies as well as correlation and regression analyses. Together with the procedure
for the assessment of traffic quality on roads with 2+1-layout developed by Brannolte, Baselau, Dong (2004)
respectively Baselau (2006), traffic planners have two helpful instruments for planning 2+1-routes. While the
procedure developed within the dissertation that forms the basis of this paper serves to determine the section
lengths in the (pre-) planning stage, the other procedure by Brannolte, Baselau, Dong respectively Baselau can
be used to assess the level of service on 2+1-routes.
REFERENCES
BASELAU, C. (2006): Entwicklung eines Verfahrens zur Beurteilung der Verkehrsqualität auf Straßen mit 2+1-
Verkehrsführung. Heft 03, Schriftenreihe Professur Verkehrsplanung und Verkehrstechnik, Bauhaus-Universität
Weimar
BRANNOLTE, U, BASELAU, C., DONG, P. (2004). Zusammenhänge zwischen Verkehrsstärke und
Verkehrsablauf auf neuen Querschnitten nach RAS-Q 96: Untersuchung des Verkehrsablaufs auf dem
Straßentyp RQ 15,5. Heft 899 der Schriftenreihe „Forschung Straßenbau und Straßenverkehrstechnik“,
herausgegeben vom Bundesminister für Verkehr, Bau- und Wohnungswesen, Bonn
FRIEDRICH, B.; DAMMANN, W.; IRZIK, M. (2005) Ausbaustandard und Überholverhalten auf 2+1-
Strecken. Schlussbericht zum Forschungsprojekt im Auftrag der Bundesanstalt für Straßenwesen, Institut für
Verkehrswirtschaft, Straßenwesen und Städtebau, Universität Hannover
FORSCHUNGSGESELLSCHAFT FÜR STRAßEN- UND VERKEHRSWESEN (2007). Richtlinien für die
Anlage von Landstraßen (RAL) – Entwurf. Stand: 07.03.2007. Arbeitsgruppe Straßenentwurf, Köln
GATTIS, J.L., BHAVE, R., DUNCAN, L.K. (2006): Alternating Passing Lane Lengths. Transportation
Research Record: Journal of the TRB, No. 1961, Washington D.C.
IRZIK, M. (2009): Überholverhalten auf 2+1-Strecken – Ein Beitrag zur Gestaltung von dreistreifigen
Landstraßen. Dissertation an der Fakultät Architektur, Bauingenieurwesen und Umweltwissenschaften der
Technischen Universität Carolo-Wilhelmina zu Braunschweig, http://www.digibib.tu-bs.de/?docid=00027821
KÖLLE, M. (1999): Sicherheitseigenschaften außerörtlicher Knotenpunkte. Heft V 67 der Berichte der
Bundesanstalt für Straßenwesen, Bergisch Gladbach
PALM, I.; SCHMIDT, G. (1999): Querschnittsbreiten einbahniger Außerortsstraßen und Verkehrssicherheit
und Sonderuntersuchung zum Querschnittstyp b2+1. Heft V 64 der Berichte der Bundesanstalt für
Straßenwesen, Bergisch Gladbach
PRIEMER, C. (2004): Überholverhalten auf 2+1-Strecken. Diplomarbeit am Institut für Verkehrswirtschaft,
Straßenwesen und Städtebau der Universität Hannover, 2004
ROOS, R. (1989). Pulkbildung und Pulkauflösung als Kriterien zur Bemessung dreistreifiger Außerortsstraßen
mit der Betriebsform b2+1. Dissertation an der TH Darmstadt
WEBER, R., LÖHE, U. (2003). Verkehrssicherheit und Verkehrsablauf auf b2+1-Strecken mit allgemeinem
Verkehr. Heft V 109 der Berichte der Bundesanstalt für Straßenwesen, Bergisch Gladbach