Swiss drivers change lanes extremely early in weaving sections. All lane changes are concentrated in the merge area. Such driving behavior results in queues and reduction of speeds. Guidelines and norms for Swiss weaving sections will be made. The goal is to improve the operation of Swiss freeway weaving sections.
Weaving sections, i.e. freeway portions where an off-ramp closely follows an on-ramp (top image and figure 1) are often bottlenecks and cause congestion. The merging and diverging maneuvers from the on-ramp and off-ramp interact with each other and form a weaving pattern on the freeway section, hence the name.
Top Image and Figure 1: Example (Hagnau-Gellert) and illustration of a freeway weaving section
The interaction of the conflicting traffic streams from the merging and diverging flows cause problems. In particular, the lane-changing behavior affects the capacity of the weaving section and the average speed of the vehicles. To better understand these lane-changing maneuvers and their effects on traffic, we have analyzed empirical data collected on five weaving areas in Switzerland (Figure 2).
Figure 2: Locations of weaving sections
The analysis aims at answering these questions:
- How do Swiss drivers behave in weaving sections?
- Are Swiss driving habits in weaving sections good or bad for traffic?
- How do weaving bottlenecks form? How is capacity reduced?
- How can we better manage traffic in weaving sections to improve their operation?
- In the future, with autonomous vehicle technology, how can cars cooperate more efficiently in weaving sections?
The final goal is to comprehensively understand the operation of Swiss freeway weaving sections. Guidelines and norms for Swiss weaving sections will also be made.
There are two types of lane changes involved in a weaving section. The ‘weaving lane changes’ refer to those into and out of the ramp lane. The ‘non-weaving lane changes’ refer to those into and out of the over-taking lane. Results show that most lane changes at the weaving section are weaving lane changes. This could mean that the majority of the non-weaving lane changes are already finished before the vehicles enter the weaving section.
Most importantly, most lane changes happen at the merge, during all time periods, under all traffic conditions. This is consistent across days and sites. For example for the pilot site, 80% of the total lane changes happen within the first 20% (100m) of the weaving section (Figure 3). Our results show that such high concentration of lane changes at the merge induces bottlenecks, reducing both the capacity and the speeds. During rush hours, queues often form at the merge location and propagate a long distance upstream. This is detrimental for the highway operation. Hence, the drivers should be encouraged to use the whole weaving section to make lane changes.
Figure 3: Distribution of lane changes at weaving area Hagnau-Gellert
Haitao He is scientific collaborator and Monica Menendez director of the Traffic Engineering research group within the Institute for Transport Planning and Systems (IVT).