New Survey Method for Calibration of Roundabout Capacity Models

A new survey method has been developed under funding by SIDRA SOLUTIONS to estimate roundabout capacity by utilizing detailed vehicle trajectory data. This method uses gap acceptance cycles based on a signal analogy concept that was used for developing the gap acceptance capacity and performance models employed in the SIDRA INTERSECTION software. The method has been applied to two approaches of a single-lane roundabout with moderate traffic conditions in Raleigh, NC, USA, using vehicle trajectory data extracted from drone footage. The survey method defines reliable gap acceptance cycles for measuring follow-up headway, critical headway and cycle-by-cycle capacity simultaneously. This is done by observing the departure of queued vehicles in saturated, unblocked periods. Gap acceptance capacity estimates have been assessed for the Highway Capacity Manual (HCM), Siegloch and Akçelik models through direct observations using the new survey method, anchored regressions and unconstrained regressions. The anchored regressions mitigate the risk of unreliable capacity estimations stemming from data scarcity under low traffic conditions. Various aggregation methods were investigated for their effect on the fitness of the capacity models. The capacity model assessments presented in the paper show the validity of the survey method, supporting the novel signal analogy concept with real-world data. Future research is recommended on various aspects of the survey and on the application of the survey to gap acceptance processes at different intersection types.