Three images that the differences between standard arrow signs and fish-hook arrow signs, both used in roundabouts

Reducing Driver Errors at Two-Lane Roundabouts

Researchers studied driving behavior at four multilane roundabouts to better understand the relationship between traffic control designs and driver errors. Data collected showed that certain traffic control changes decreased turn violations but failed to eliminate yield violations. Researchers were unable to identify long-term solutions for improving roundabout design and signage, and recommended further research to improve the overall safety and mobility of multilane roundabouts.

“Even though the study did not provide a silver bullet on how to prevent crashes at multilane roundabouts, it did create an efficient tool to analyze and quantify driving behavior data,” said Joe Gustafson, Traffic Engineer, Washington County Public Works.

“This study has advanced our knowledge in multilane roundabout safety and is one step closer to providing much needed improvements to roundabout design guidance,” said John Hourdos, Director, Minnesota Traffic Observatory, University of Minnesota.

What Was the Need?

Roundabouts have been shown to improve overall in-tersection safety compared to traditional traffic signals. However, noninjury crashes are sometimes more frequent on multilane roundabouts than on single-lane roundabouts due in part to driver confusion. Common driver errors such as failing to yield and turning violations on multilane roundabouts have contributed to an increase in noninjury crashes.

Given the benefits of improved mobility, traffic throughput and injury reduction of multilane roundabouts, reducing the noninjury crash rate at multilane roundabouts is important to facilitating their use by Minnesota cities and counties. Identifying solutions to reduce common driving violations would be more sustainable than the current practice of converting multilane roundabouts back to single-lane roundabouts.

In a previous study on a two-lane roundabout in Richfield, Minnesota, researchers demonstrated that traffic control  changes could reduce some of these driver errors. However, more data was needed to support study results. Understanding driver behavior and improving traffic control devices are key factors in designing safer multilane roundabouts.

What Was Our Goal?

With limited research on modern multilane roundabouts, the Minnesota Traffic Observatory sought to collect more data to evaluate the correlation between traffic control design features and collisions. Instead of conducting manual observations, researchers used an innovative video analysis tool to collect and process recorded videos of driving behaviors at test sites. Based on the analysis, they attempted to identify driver behaviors and error rates to help reduce noninjury crashes at multilane roundabouts.

What Did We Do?

The research team selected four multilane roundabouts in Minnesota — two in Mankato, one in Lakeville and one in St. Cloud — to observe undesirable driving maneuvers. At each roundabout site, researchers mounted video cameras at key locations to record one to two weeks of driving behavior. Only one roundabout could be observed at a time because only one set of specialized video equipment was available.

The raw videos were processed to produce a data set for analysis. Researchers used TrafficIntelligence, an open-source computer vision program, to automate extraction of vehicle trajectories from the raw footages. They used the same software to correct any errors to improve data reliability. The resulting clean data from the recorded videos were supplemented with historical crash frequency data reports obtained from the Minnesota Department of Public Safety. Collectively, data from both sources allowed researchers to thoroughly investigate the frequency and crash types from the four roundabouts. A statistical analysis of the data revealed that turn violations and yield violations were among the most common driving errors.

Researchers also looked at how violation rates vary with the roundabout’s location and relevant design features. Based on these findings, researchers proposed signage and striping changes to reduce driver errors at the two Mankato test sites. After the changes were implemented, they collected additional video data.

What Did We Learn?

This study provided one of the most comprehensive analyses to date of driving behavior at multilane roundabouts. Researchers were successful in finding solutions for reducing turn violations, but they were unable to identify solutions for yield violations despite the vast amount of data.

Minor differences in the design at each roundabout presented specific challenges. The analysis focused on how each varying design feature impacted driving behavior. Proposed traffic control changes such as extending solid lines between entrance lanes, adjusting the position of yield signs and adding one-way signs successfully decreased turn violations. However, data from before and after traffic control changes showed an insignificant impact on decreasing yield violations. Importantly, the study produced a list of ineffective traffic control methods that can be eliminated from future studies, saving engineers time and money.

The TrafficIntelligence tool was crucial in efficiently processing and cleaning large amounts of raw video. With improvements made to the software program, the tool should be an asset to future research on roundabouts and to other studies requiring observations of driving behavior.

What’s Next?

The traffic control changes that were successful at reducing crashes at two-lane roundabouts should be implemented by traffic engineers. In particular, large overhead directional signs or extended solid lines between entrance lanes should be installed to help reduce turning violations. The analysis method used in this study could also be used for a robust before-and-after evaluation of future modifications to traffic control devices.

Additional research could further scrutinize the data already collected, and researchers recommend that more data be collected to examine additional traffic control methods and other intersection design elements to improve the overall safety and mobility of two-lane roundabouts. This research could also serve as an impetus for the study of numerous roundabouts in a pooled fund effort involving several states.


This post pertains to the LRRB-produced Report 2017-30, “Evaluation of Safety and Mobility of Two-Lane Roundabouts,” published July 2017. A webinar recording of the report is also available.

Leave a Comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s