Analysis of AEB False Positives in Urban Bus Operations

Sammanfattning

False positive (FP) activations of Automatic Emergency Braking (AEB) in urbanbuses can cause unnecessary harsh braking and reduce system acceptance for situationsinvolving both passengers and vulnerable road users. This report implementsa traceable, criteria-based pipeline to identify candidate FP events from raw ASAMMDF 4 (MF4) recordings, structured as file-level screening, qualified brake-eventdetection, and FP classification. The implementation uses core Controller AreaNetwork (CAN) signals to detect AEB-active braking events and applies collisionriskand driver-response criteria for classification, where time-to-collision (TTC) iscomputed from recorded signals and used as one decision input. The TTC computationmatches a ground-truth TTC reference provided by a colleague within theGroup Trucks Technology team on the evaluated data, and the end-to-end detectionpipeline runs on an internal dataset to automatically extract candidate FP eventsfor further attribution and performance analysis.