Impact of Ball Bank Indicator on Predicting Rural Curve Crashes

Abstract

The ball bank indicator (BBI) measures the lateral forces on a vehicle. It is used to establish the advisory speed limit as outlined in the American Association of State Highway and Transportation Officials (AASHTO)’s Green Book. BBI values respond to roadway geometry and driver behavior. Currently, BBI data are available from various curves across the U.S. However, the relationship between BBI and curve lane departures is unknown. Therefore, the objective of this paper is to assess the impact of BBI as an explanatory variable for curve lane departures within a safety performance function (SPF) (i.e., a crash prediction model). To accomplish this objective, a study is conducted on rural curves in Districts 1, 2, and 6 of Georgia Department of Transportation in the U.S. BBI is integrated into a negative binomial model alongside other common explanatory variables used in the Highway Safety Manual. This SPF, with BBI incorporated, is compared with a baseline SPF without the BBI. The results show BBI is a statistically significant variable under a 99.9% threshold. Additionally, it was found that the model with BBI has 2.78% and 2.83% less mean absolute error and route mean squared error, respectively. Though the improvement in the model is minor, this finding is notable because BBI data may already be available for a transportation agency to leverage to assess risk on curves. Furthermore, this data could be even more beneficial if it were crowdsourced to gauge real-world behaviors.