论文信息 - Evaluation of safety critical event triggers in the UDrive data

Evaluation of safety critical event triggers in the UDrive data

Abstract The understanding of the root causes of traffic collisions has taken a new turn with naturalistic driving studies. Naturalistic driving Data bring insight into normal behavior on the road and help characterize safety related and safety critical events (SCEs) through kinematic triggers and video analysis. The UDRIVE Database includes two million kilometers of naturalistic driving and was recorded in five European countries between 2015 and 2017. Given the large volume of raw data considered, the performance of these kinematic triggers in detecting SCEs should be carefully evaluated to optimize the human resources required to discard false positive events. The aim of the present article is twofold: (i) to describe the triggers used to detect safety critical events from the UDRIVE Database and (ii) to discuss the relevance of SCE triggers performance evaluation used in the literature.

Erwan Lecuyer | Stéphane Buffat | Laurette Guyonvarch

[1] Marco Dozza,et al. Recognising safety critical events: can automatic video processing improve naturalistic data analyses? , 2013, Accident; analysis and prevention.

[2] Nicole van Nes,et al. The study design of UDRIVE: the naturalistic driving study across Europe for cars, trucks and scooters , 2016 .

[3] Robert J Carroll,et al. The development of a naturalistic data collection system to perform critical incident analysis: an investigation of safety and fatigue issues in long-haul trucking. , 2006, Accident; analysis and prevention.

[4] Jonathan M Hankey,et al. Performance of basic kinematic thresholds in the identification of crash and near-crash events within naturalistic driving data. , 2017, Accident; analysis and prevention.

[5] Takaya Saito,et al. The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced Datasets , 2015, PloS one.