Injury Severity Prediction Algorithm Based on Select Vehicle Category for Advanced Automatic Collision Notification SUBARU CORPORATION

Format:

Book

Conference/Event

Author/Creator:

Ejima, Susumu, author.

Contributor:

Cunningham, Kristen

Goto, Tsukasa

Wang, Stewart

Zhang, Peng

Conference Name:

WCX SAE World Congress Experience (2022-04-05 : Detroit & Online, Michigan, United States)

Language:

English

Physical Description:

1 online resource cm

Place of Publication:

Warrendale, PA SAE International 2022

Summary:

With the evolution of telemetry technology in vehicles, Advanced Automatic Collision Notification (AACN), which detects occupants at risk of serious injury in the event of a crash and triages them to the trauma center quickly, may greatly improve their treatment.An Injury Severity Prediction (ISP) algorithm for AACN was developed using a logistic regression model to predict the probability of sustaining an Injury Severity Score (ISS) 15+ injury. National Automotive Sampling System Crashworthiness Data System (NASS-CDS: 1999-2015) and model year 2000 or later were filtered for new case selection criteria, based on vehicle body type, to match Subaru vehicle category. This new proposed algorithm uses crash direction, change in velocity, multiple impacts, seat belt use, vehicle type, presence of any older occupant, and presence of any female occupant. Moreover, presence of the right-front passenger and its interaction with crash direction were considered, which affected risk prediction significantly especially in the side-impact crashes. Variable selection techniques were used to construct the final ISP algorithm with relevant features. In this paper, we presented results of two type of injury prediction algorithms, which do not (ISP) or do (ISP-R) consider the effect of a right-front passenger were proposed.In order to evaluate model performance, five-fold cross-validation was performed within the training data (NASS-CDS 1999-2015). Additionally, the ISP algorithm for model was also externally validated using National Automotive Sampling System Crash Investigation Sampling System (NASS-CISS: 2017-2019). The area under the receiver operator characteristic curve (AUCs) was used as the metric to evaluate model performances, AUC was 0.854 with the ISP model, 0.862 with the ISP-R model for cross-validation and 0.817 with the ISP model, 0.828 with the ISP-R model for external validation. Delta-V, seat belt use, and crash direction were important predictors of serious injury, and moreover, the presence of right-front passenger was a significant injury risk modifier, especially for side impact crashes

Notes:

Vendor supplied data

Publisher Number:

2022-01-0834

Access Restriction:

Restricted for use by site license