A study on the cyclist head kinematic responses in electric-bicycle-to-car accidents using decision-tree model. (September 2021)

• Record Type: Journal Article
• Authors: Gao, Wenrui; Bai, Zhonghao; Zhu, Feng; Chou, Clifford C.; Jiang, Binhui
• Is Part Of: Accident analysis and prevention. Volume 160(2022)
• Publication Date: 2021-09
• Subjects: Data mining -- Electric-bicycle -- Cyclist safety -- Head injury -- Head kinematic responses; Accidents -- Prevention -- Periodicals; Accident Prevention -- Periodicals; Accidents -- Prévention -- Périodiques; 363.106
• DOI: 10.1016/j.aap.2021.106305
• Languages: English
• ISSNs: 0001-4575

Abstract

Highlights: Emerging decision tree along with simulation model to E-bike-car accident analysis. Cyclist's head kinematic responses were collected from simulations to form dataset. C4.5 decision tree algorithm was used to mine and classify cyclist's head dataset. Decision tree models of cyclist's head kinematic responses were developed. Initial impact velocity of car was a key factor affecting cyclist's head responses. Abstract: Due to the high frequent traffic accidents involving electric bicycles (E-bike), it urgently needs improved protection of cyclists, especially their heads. In this study, by adjusting the initial impact velocities of E-bike and car, initial impact angle between E-bike and car, initial E-bike impact location, and body size of cyclist, 1512 different accident conditions were constructed and simulated using a verified E-bike-to-car impact multi-body model. The cyclist's head kinematic responses including the head relative impact velocity, WAD (Wrap around distance) of head impact location and HIC15 (15 ms Head Injury Criterion) were collected from simulation results to make up a dataset for data mining. The decision tree models of cyclist's head kinematic responses were then created from this dataset and verified accordingly. Based on simulated results obtained from decision tree models, it can be found as follows. 1. In the E-bike-to-car accidents, the average head impact relative velocity and WAD of head impact location are higher than those in the car-to-pedestrian accidents. 2. Increasing the initial impact velocity of car can increase the cyclist's head relative impact velocity, WAD of head impact location, and HIC15 . 3. The WAD of cyclist's head impact location is also significantly affected by the initial impact angle between E-bike and car and body size of cyclist: the WAD of head impact location becomes higher with increasing initial impact angle between E-bike and car and body size of cyclist. 4. The effects of initial E-bike impact location on the WAD of cyclist's head impact location is not significant when initial E-bike impact location is concentrated in the region of 0.25 m around the centerline of the car.