Impact of Geometric Factors on the Capacity of Single-Lane Roundabouts

by Mark T. Johnson, P.E., and Ting-Li Lin, Ph.D.

Mark Johnson is the Principal Engineer for MTJ Roundabout Engineering. Mark founded MTJ in 2005, after 12 years of public- and private-sector transportation engineering positions, including four years for the City of Loveland, Colorado, and four years at the Wisconsin DOT where he was a key contributor to WisDOT roundabout program. MTJ's work is focused on roundabout solutions to complex traffic problems. Mark works with agencies and consultancies on roundabout feasibility analysis, design, review, and public outreach/education. Mark is co-author of the 2010 FHWA Roundabout Guide (NCHRP 672), is an authorized FHWA Peer-to-Peer Roundabout Reviewer, and an active member of the ITE and TRB Roundabout Committees.

Ting-Li Lin has a Master's degree and a PhD degree in statistics from the University of Wisconsin-Madison. He has 12 years of experience in statistical consulting, working with researchers and scientists from various fields, including engineering, biomedical science, agricultural science, social science, and the pharmaceutical industry. Ting-Li recently joined the Frontier Science and Technology Research Foundation as a research scientist.

Submitted to TRB for 2018 TRB Annual Meeting Presentation

Roundabout capacity is primarily estimated by gap-acceptance or by geometric models. The 2010 Highway Capacity Manual (HCM 10) uses a gap-acceptance model developed by Siegloch (1973) with empirically derived values of critical gap and headway for single-lane and two-lane entries. A geometric capacity model was developed by Kimber and Hollis (1980) that diverged from gap-based models. The capacity of up to four lane entries was empirically derived from six geometric parameters.

In 2012 capacity data was collected as part of the FHWA project, Assessment of Roundabout Capacity Models for the Highway Capacity Manual. This data was used to produce HCM6, an update of the HCM 2010. In HCM6, a capacity curve was fitted through all the capacity data for single-lane roundabouts. The large scatter of data about the mean capacity line suggests that the single-lane roundabouts may be separated into different geometric types to improve accuracy.

To investigate this hypothesis, the capacity data was separated into two geometrically distinct types of single-lane roundabouts: (1) smaller, compact and (2) larger, curvilinear. The range of data for the disaggregated and aggregated data was compared. Also, a capacity line was derived for each type using the geometric capacity model and compared to the disaggregated data with the HCM6 method.

The results demonstrate that differences in geometry, absent in HCM6, explain the wide data range. This was further confirmed by the geometric model that gave a good fit to both sets of data. These results indicate that the accuracy of capacity prediction is improved by including geometric variation.

Read the full paper here.