Safety of two-lane roundabouts, VSS 2010/301, FB 1541

Doerfel, Marion; Bättig, Daniel; Lindenmann, Hans Peter; Huber, Christian Ary; Berger, Nicolas (2015). Safety of two-lane roundabouts, VSS 2010/301, FB 1541 mobility Platform, Dokumenten-Nummer FB 1541, kostenpflichtig: UVEK

Full text not available from this repository. (Request a copy)

Summary This study provides information on the safety level of two-lane roundabouts in Switzerland, not only two-lane roundabouts generally but also different subtypes and differentiated by roundabout elements. Furthermore, the research describes the relation-ship between roundabout design and accidents. Finally it offers recommendations concerning the choice of geometric design features and the subtype of two-lane roundabouts. For over 30 years, roundabouts have been an increasingly favoured type of intersection in Switzerland. The transformation of existing junctions into compact single-lane round-abouts (type “Kleinkreisel”) has proved particularly effective in regard to road safety. Research activities and a wealth of experience resulted in recommendations for the design of single-lane roundabouts which were published in the Swiss standard SN 640 263 Knoten; Knoten mit Kreisverkehr [6] by the Swiss Association of Road and Traffic Professionals (VSS) in the late 1990s. In recent years, two-lane roundabouts have been built in order to increase capacity. Two-lane roundabouts have often been labelled as more dangerous than their single-lane counterparts. In 2009, in the context of research regarding the capacity of two-lane roundabouts [12] an initial study on safety of 15 roundabouts led to the conclusion that, as expected, the number of accidents on two-lane roundabouts is higher than on those with one lane only, but accident rates in terms of traffic volume were found to be similar. The study also showed that the greater geometric dimensions, higher speed level, and the lane changes usually found on two-lane roundabouts contribute to unfavourable abnormalities in accident occurrence in comparison with single-lane roundabouts. In [12] the accidents are assumed to be associated with the geometric design features. Therefore the present research aims to determine the safety level of two-lane round-abouts in Switzerland, and to investigate the influence of geometric parameters on accident occurrence on two-lane roundabouts. Roundabouts are classed as two-lane roundabouts if they feature two circulating lanes or one extra wide circulating lane and at least one leg with two entry lanes. The peculiarities and differences between the 2/2 type (at least one leg with two entry lanes / two circulating lanes) and the 2/1+ type (at least one leg with two entry lanes / one extra wide circulating lane) are additionally examined. The study is based on a sample of 97 two-lane roundabouts. The sample set includes type 2/2 and type 2/1+ roundabouts and a small number of special types which also matched the research criteria. Several hypotheses regarding the number of accidents caused by the geometry of round-abouts are formulated. To estimate the plausibility of the hypotheses, a homogenous Poisson process models the distribution of the number of accidents. This number is therefore a linear function of the observation period. The expected number of accidents per year μ is then estimated with the help of a generalised linear model as a function of the geometrical dimensions and the average daily traffic (ADT). The models have been checked by in-sample prediction. The findings of this research refute previous concerns that two-lane roundabouts would show a very unfavourable impact on road safety. The safety level, expressed by the accident rate, is equivalent to that of single-lane roundabouts. When a two-lane roundabout is deemed necessary for capacity reasons, two-lane round-abouts with two circulating lanes (type 2/2) should be selected as data on the differences in safety level and severity of accidents of the three examined types of roundabouts clearly show. The number of two-lane entries should be limited to the number established for capacity reasons, at least and preferably just one. Two-lane exits should be avoided wherever possible as these exits can rarely be justified. The results also indicate that roundabouts with four legs are safer than roundabouts with three legs. However, it should be noted that the three-leg roundabouts contained in the sample set have an unevenly distributed arrangement of legs (no deflection). The number of legs can hardly be influenced because it is a consequence of the traffic planning situation. Regarding the influence of other geometric parameters on the accident occurrence the following can be assumed: • The outside diameter, followed by the fleet angle, has the largest influence on the expected average number of accidents per year. Smaller outside diameters are safer than large ones. They should particularly be favoured at lower traffic volumes. The required minimum external diameter is 32 m. There are no grounds for external diameters of 50 m and more. The fleet angle should be as large as possible: an angle of at least 35 degrees should be ensured but angles of 45 degrees and more are recommended. • It may be deduced from these results that the average number of accidents per year is influenced by the width of both the circulating lane and the entry lanes. The number of accidents increases the wider the entry lane and decreases slightly with increasing width of the circulating lane. The entry lane should therefore be designed to be as narrow as possible (the minimum width is 6 m), even with small outside diameters. The width of the two-lane circulating lane is, taking account of tractrices, to be determined by the required lane width including curve widenings. Larger widths may be permitted. • Furthermore, statements can be made about the influence of central islands. Narrow central islands in entries/exits are more advantageous than wide ones. A sufficient physical separating effect must, however, be ensured. • Due to contradictory results regarding the width of the central island in the entry and the size of the entry angle, no recommendation concerning the dimension of entry angles can be given. • Accident situations (types of accidents) point to specific conclusions in regards to point of entry: the geometric design of entries seeks to use narrow lanes and channelling elements in an effort to reduce approach speed. This recommendation is merely based on indications provided by this research, but has also been described similarly in the literature. On the basis of the findings in this research it is recommended that a standard for the planning and design of two-lane roundabouts with two circulating lanes be developed and adopted as soon as possible. This will facilitate the choice of roundabout subtype and geometric design features. The safety of two-lane roundabouts designed as special shapes (turbo roundabout, etc.) should be investigated in greater depth immediately because the present study has indicated that they have an increased severity of accidents compared to the two other types. These research results can be used for the following purposes: • Directly, as a general quantitative guide during the planning of new two-lane round-abouts, until the corresponding standard for the design of two-lane roundabouts has been adopted. This concerns findings about the geometric dimensions, as far as these are relevant. • The findings on the operating design of two-lane roundabouts can be used for the assessment of traffic planning issues from now on. • The results on accident rates, accident severity and distribution of accident types can serve as a basis for examining the general safety level of two-lane roundabouts. These research results are unsuitable for the following purposes: • The results are too unreliable for the comprehensive detailed design of roundabout access, entry and exit. • The conclusions are not to be used directly for the renovation of present roundabouts, since each roundabout has a unique geometry and design which lead to a corresponding extent of accident situations. For existing roundabouts, measures for improving road safety are in any case to be made on the basis of an accident analysis. The findings of this research provide advice on the selection of measures. • The model used to examine the influence of geometric and operational factors on accidents is only applicable to a limited extent for the estimation of future accident frequency. This is not least because the traffic volume is, to some extent, an estimate. Additionally, not all geometric parameters are included in the model.

Item Type:

Report (Report)


School of Architecture, Wood and Civil Engineering > Institute for Urban Development and Infrastructure
School of Architecture, Wood and Civil Engineering > Institute for Urban Development and Infrastructure > Transport Infrastructure


Doerfel, Marion;
Bättig, Daniel;
Lindenmann, Hans Peter;
Huber, Christian Ary and
Berger, Nicolas




[UNSPECIFIED] Bundesamt für Strassen ASTRA


Marion Doerfel

Date Deposited:

03 May 2021 14:58

Last Modified:

03 May 2021 14:58


Actions (login required)

View Item View Item
Provide Feedback