A method for predicting crash configurations using counterfactual simulations and real-world data

Traffic safety technologies revolve around two principle ideas; crash avoidance and injury mitigation for inevitable crashes. The development of relevant vehicle mitigating should consider the interaction those technologies, ensuring that crashes can be adequately managed by occupant vulnerable road user (VRU) protection systems. A step towards is accurate description expected remaining when crash-avoiding are available in vehicles. With overall objective facilitating assessment future traffic safety, this study develops a method predicting configurations introducing countermeasures. predicted one important factor prioritizing evaluation VRU By using real-world accident data to form baseline performing counterfactual model-in-the-loop (MIL) pre-crash simulations, change situations (vehicle crashes) provided vehicles with predicted. built on novel configuration definition, which supports further analysis in-crash phase. clustering grouping crashes, limited number identified, still representing covering variation. developed was applied Swedish national- in-depth related urban intersections highway driving, conceptual Autonomous Emergency Braking system (AEB) computational model. Based national analysis, conflict Same-Direction rear-end frontal (SD-ref) 53 % vehicle-to-vehicle (v2v) Straight Crossing Path (SCP) 21 v2v intersection were selected study. Pre-crash baselines, SD-ref (n = 1010) SCP 4814), prepared based variations these. simulations identified not avoided AEB, these revealed 5 52 representative situations, respectively, used crashworthiness studies. results demonstrated feasible way identifying, predictive way, testing prevention capabilities.