Investigating the impacts of crash prediction models on quantifying safety effectiveness of Adaptive Signal Control Systems

• Evaluates safety effects of Adaptive Signal Control System (ASCS) on crash frequency. Implements Empirical Bayes (EB) and Fully Bayesian (FB) frameworks. Finds ASCS reduces frequencies for most corridors intersections. Shows that vary with different intersection features. Introduction: can improve both operational benefits at signalized corridors. Methods: This paper develops a series models accounting model forms possible predictors implements these in frameworks evaluation studies. Different are validated terms the ability to reduce potential bias variance prediction effectiveness estimation accuracy using real-world data from non-ASCS sites. then six total 65 intersections same type ASCS, South Carolina. Results: Validation results show FB accounts traffic volume, roadway geometric features, year factor, spatial shows best performance among all models. The study findings reveal crash, fatal injury angle varies features (i.e., AADT major streets, number legs an intersection, through lanes access points minor speed limit streets). Conclusions: is associated reductions, its Practical Applications: this research encourage more deployments provide insights into selecting deployment sites reducing crashes considering variation ASCS.