Integrating safety into the fundamental relations of freeway traffic flows: A conflict-based safety assessment framework

Numerous statistical and data-driven modeling frameworks have estimated rear-end crashes crash-prone events from macroscopic traffic states which are at the heart of flow modelling control. However, existing focus on critical exclude a vast majority safer interactions, essential information with respect to identifying trade-offs between congestion management crash prevention. This study proposes flexible conflict-based framework extract safety freeway state variables (i.e., speed density) by utilizing all underlying car-following interactions. Time spent in conflict (TSC) is introduced as total time vehicles conflicts based given measure threshold be determined flexibly. Using NGSIM vehicle trajectory dataset, we show that proportion stopping distance (PSD) more desirable than several event-based measures (e.g., collision) for describing TSC variables. Besides, it shown PSD provides explicit about entire travel each because applies paper hybrid methodological combining probabilistic machine learning models develop relationships within assessment framework. At first, Machine separately developed estimate PSD-based using only stte Each approach evaluated comprehensively against empirical observations dataset. While has better predictive accuracy fixed PSDcr), explaining capability captures thresholds. Utilizing advantages these two approaches, proposed satisfactorily predicts corresponding PSD<PSDcr wide range thresholds an endogenous dimension fundamental relations flows can utilized balance efficiency safety. For instance, control studies utilize minimize while also minimizing situations such shockwaves oscillations.