Traffic injury prevention · 2018
Improvement of injury severity prediction (ISP) of AACN during on-site triage using vehicle deformation pattern for car-to-car (C2C) side impacts.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Pal Chinmoy, Hirayama Shigeru, Narahari Sangolla, Jeyabharath Manoharan, Prakash Gopinath, Kulothungan Vimalathithan, Combest John
Аннотация
OBJECTIVE: The Advanced Automatic Crash Notification (AACN) system needs to predict injury accurately, to provide appropriate treatment for seriously injured occupants involved in motor vehicle crashes. This study investigates the possibility of improving the accuracy of the AACN system, using vehicle deformation parameters in car-to-car (C2C) side impacts.
METHODS: This study was based on car-to-car (C2C) crash data from NASS-CDS, CY 2004-2014. Variables from Kononen's algorithm (published in 2011) were used to build a "base model" for this study. Two additional variables, intrusion magnitude and max deformation location, are added to Kononen's algorithm variables (age, belt usage, number of events, and delta-v) to build a "proposed model." This proposed model operates in two stages: In the first stage, the AACN system uses Kononen's variables and predicts injury severity, based on which emergency medical services (EMS) is dispatched; in the second stage, the EMS team conveys deformation-related information, for accurate prediction of serious injury.
RESULTS: Logistic regression analysis reveals that the vehicle deformation location and intrusion magnitude are significant parameters in predicting the level of injury. The percentage of serious injury decreases as the deformation location shifts away from the driver sitting position. The proposed model can improve the sensitivity (serious injury correctly predicted as serious) from 50% to 63%, and overall prediction accuracy increased from 83.5% to 85.9%.
CONCLUSION: The proposed method can improve the accuracy of injury prediction in side-impact collisions. Similar opportunities exist for other crash modes also.
METHODS: This study was based on car-to-car (C2C) crash data from NASS-CDS, CY 2004-2014. Variables from Kononen's algorithm (published in 2011) were used to build a "base model" for this study. Two additional variables, intrusion magnitude and max deformation location, are added to Kononen's algorithm variables (age, belt usage, number of events, and delta-v) to build a "proposed model." This proposed model operates in two stages: In the first stage, the AACN system uses Kononen's variables and predicts injury severity, based on which emergency medical services (EMS) is dispatched; in the second stage, the EMS team conveys deformation-related information, for accurate prediction of serious injury.
RESULTS: Logistic regression analysis reveals that the vehicle deformation location and intrusion magnitude are significant parameters in predicting the level of injury. The percentage of serious injury decreases as the deformation location shifts away from the driver sitting position. The proposed model can improve the sensitivity (serious injury correctly predicted as serious) from 50% to 63%, and overall prediction accuracy increased from 83.5% to 85.9%.
CONCLUSION: The proposed method can improve the accuracy of injury prediction in side-impact collisions. Similar opportunities exist for other crash modes also.