Stapp car crash journal最新文献

{"title":"Driving Behavior during Right-Turn Maneuvers at Intersections on Left-Hand Traffic Roads.","authors":"Yasuhiro Matsui,&nbsp;Naruyuki Hosokawa,&nbsp;Shoko Oikawa","doi":"10.4271/2022-22-0008","DOIUrl":"10.4271/2022-22-0008","url":null,"abstract":"<p><p>In Japan, where vehicles drive on the left side of the road, pedestrian fatal accidents caused by vehicles traveling at speeds of less than or equal to 20 km/h, occur most frequently when a vehicle is turning right. The objective of the present study is to clarify the driving behavior in terms of eye glances and driver speeds when drivers of two different types of vehicles turn right at an intersection on a left-hand traffic road. We experimentally investigated the drivers' gaze, vehicle speed, and distance on the vehicle traveling trajectory from the vehicle to the pedestrian crossing line, using a sedan and a truck with a gross vehicle weight of < 7.5 tons (a light-duty truck) during right-turn maneuver. We considered four different conditions: no pedestrian dummy (No-P), right pedestrian dummy (R-P), left pedestrian dummy (L-P), and right and left pedestrian dummies (RL-P). Regarding the gazing characteristics, there was no significant difference in the average total gaze time at each AOI between the two vehicles under different conditions, which suggests that the total gaze time was not affected by the vehicle type. All participants gazed at the pedestrian dummies in R-P, L-P, and RL-P. However, the average total gaze time at the right pedestrian dummy (0.63-0.72 s) in R-P was significantly shorter than that at the left pedestrian dummy (1.46-1.57 s) in L- P for both vehicles. The average vehicle speed at the entrance line to the intersection (L1) of the light-duty truck (16.8-18.2 km/h) was lower than that of the sedan (18.8-19.7 km/h). The average vehicle speed at the pedestrian crossing line (L0) of the light-duty truck (15.5-16.0 km/h) was lower than that of the sedan (16.0-17.8 km/h). There was no significant difference in the average vehicle speeds at L1 and L0 between them under any two conditions. We investigated the estimated time to collision (TTC), calculated from the distance on the vehicle traveling trajectory from the vehicle to the pedestrian crossing line and the vehicle speed at the moment when the drivers first gazed at the pedestrian dummies. The average TTC of the right pedestrian dummy in R-P for the sedan (3.5 s) was significantly shorter than that for the light-duty truck (4.0 s). Similarly, the average TTC of the left pedestrian dummy in L-P for the sedan (3.7 s) was significantly shorter than that for the light-duty truck (4.8 s). The driving characteristics obtained in this study may contribute to the development of advanced driver support systems, particularly for vehicles turning right at intersections.</p>","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"66 ","pages":"217-238"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41150609","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Investigation of Potential Injury Patterns and Occupant Kinematicsin Frontal Impact with PMHS in Reclined Postures.","authors":"Pascal Baudrit,&nbsp;Jérôme Uriot,&nbsp;Olivier Richard,&nbsp;Matthieu Debray","doi":"10.4271/2022-22-0001","DOIUrl":"https://doi.org/10.4271/2022-22-0001","url":null,"abstract":"<p><p>The reality of the autonomous vehicle in a near future is growing and is expected to induce significant change inthe occupant posture with respect to a standard driving posture. The delegated driving would allow sleeping and/or resting in a seatwith a reclined posture. However, the data in the literature are rare on the body kinematics, human tolerance, and injury types insuch reclined postures. The current study aims at increasing the knowledge in the domain and providing useful data to assess therelevance of the standard injury assessment tools such as anthropomorphic test devices or finite element human body models. For that purpose, a test series of three male Post-Mortem Human Subjects (PMHS) were performed in frontal impact at a 13.4 m/sdelta V. The backseat inclination was 58 degrees with respect to the vertical axis. The semi-rigid seat developed by Uriot et al.(2015) was used with a stiffer seat ramp. The restraint was composed of a lap belt equipped with two 3.5 kN load limiters, and ofa shoulder belt equipped of a 4 kN load limiter on the upper anchorage placed in the vicinity of the shoulder. The belts, the semi-rigid seat, and the footrest were equipped with force sensors. The rotations of the seat pan and of the seat ramp were also measured. The PMHS were instrumented with multi-axis accelerometers and Y angular velocity sensors attached to the head, thorax (T1 andT12 vertebrae), and sacrum. Strain gauges were glued onto the anterior face of the L1 to L5 lumbar vertebrae and onto the anteriorface of the iliac wings. To estimate the pelvis kinematics, a rigid support equipped with targets was fixed onto the femur shaft. Prior to test, X-ray imagery was performed to exhibit the initial curvature of the lumbar spine. After the tests, an in-depth necropsywas done, with a specific attention to the lumbar spine. In the chosen test conditions, no lap-belt submarining was observed for the three PMHS. One PMHS sustained an AIS2 pelvic ringfracture and another one sustained an AIS4 injury with complete separation of the left and right sacroiliac joints. Lumbar discruptures and vertebral fractures were observed for the three PMHS (AIS 2 and AIS3 coding). The number of separated rib fractureswere very different from one PMHS to another (0, 6 and 33). Response corridors for the external forces and kinematics were builtand are presented in the paper. The results are discussed by comparing with existing data for which the backseat was in standardposture.</p>","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"66 ","pages":"1-30"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41168556","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effect of Tissue Erosion Modeling Techniques on Pedestrian Impact Kinematics.","authors":"Daniel Grindle,&nbsp;Costin Untaroiu","doi":"10.4271/2022-22-0007","DOIUrl":"10.4271/2022-22-0007","url":null,"abstract":"<p><p>The pedestrian is one of the most vulnerable road users and has experienced increased numbers of injuries and deaths caused by car-to-pedestrian collisions over the last decade. To curb this trend, finite element models of pedestrians have been developed to investigate pedestrian protection in vehicle impact simulations. While useful, modeling practices vary across research groups, especially when applying knee/ankle ligament and bone failure. To help better standardize modeling practices this study explored the effect of knee ligament and bone element elimination on pedestrian impact outcomes. A male 50th percentile model was impacted by three European generic vehicles at 30, 40, and 50 km/h. The pedestrian model was set to three element elimination settings: the \"Off-model\" didn't allow any element erosion, the \"Lig-model\" allowed lower-extremity ligament erosion, and the \"All-model\" allowed lower-extremity ligament and bone erosion. Failure toggling had a significant effect on impact outcomes (0 < p ≤ 0.03). The head impact time response was typically the smallest for the \"Off-model\" while the wrap around distance response was always largest for the All-model. Moderate differences in maximum vehicle-pedestrian contact forces across elimination techniques were reported in this study (0.1 - 1.7 kN). Future work will examine additional failure modelling approaches, model anthropometries and vehicles to expand this investigation.</p>","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"66 ","pages":"207-216"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41151113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Comparison of the Mid-Size Male THOR and Hybrid III ATDs in Vehicle Frontal Crash Tests.","authors":"Chris O'Connor,&nbsp;Agnes Kim,&nbsp;Tim Barrette,&nbsp;Jeff Dix","doi":"10.4271/2022-22-0005","DOIUrl":"10.4271/2022-22-0005","url":null,"abstract":"<p><p>In order to evaluate the THOR-50M as a front impact Anthropomorphic Test Device (ATD) for vehicle safety design, the ATD was compared to the H3-50M in matching vehicle crash tests for 20 unique vehicle models from 2 vehicle manufacturers. For the belted driver condition, a total of fifty-four crash tests were investigated in the 56.3 km/h (35 mph) front rigid barrier impact condition. Four more tests were compared for the unbelted driver and right front passenger at 40.2 km/h (25 mph) in the flat frontal and 30-degree right oblique rigid barrier impact conditions. The two ATDs were also evaluated for their ability to predict injury risk by comparing their fleet average injury risk to Crash Investigation Sampling System (CISS) accident data for similar conditions. The differences in seating position and their effect on ATD responses were also investigated. This study showed that the belted THOR-50M injury responses were higher than the H3-50M by 25%-180%, in all reported ATD responses, except chest acceleration. For one unbelted condition, the THOR-50M reported 200%-300% higher neck responses than the H3-50M, primarily due to head contact to the roof structure in a mid-sized sedan. The THOR-50M overpredicted the injury risk based on chest deflection compared to the CISS accident data by at least a factor of 4 times. The THOR-50M also overpredicted the injury risk based on BrIC by at least a factor of 10 times. Future work is needed to investigate these overpredictions with respect to ATD construction, injury risk curves, and seating procedures.</p>","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"66 ","pages":"143-173"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41142738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Lower Extremity Validation of a Human Body Model for High Rate Axial Loading in the Underbody Blast Environment.","authors":"Zachary S Hostetler,&nbsp;Juliette Caffrey,&nbsp;Jazmine Aira,&nbsp;F Scott Gayzik","doi":"10.4271/2022-22-0004","DOIUrl":"https://doi.org/10.4271/2022-22-0004","url":null,"abstract":"<p><p>While the use of Human Body Models (HBMs) in the underbody blast (UBB) environment has increased and shown positive results, the potential of these models has not been fully explored. Obtaining accurate kinematic and kinetic response are necessary to better understand the injury mechanisms for military safety applications. The objective of this study was to validate the Global Human Body Models Consortium (GHBMC) M50 lower extremity using a combined objective rating scheme in vertical and horizontal high-rate axial loading. The model's lower extremity biomechanical response was compared to Post Mortem Human Subjects (PMHS) subjects for vertically and horizontally-applied high rate axial loading. Two distinct experimental setups were used for model validation, comprising a total of 33 distinct end points for validation. A combined Correlation and Analysis (CORA) score that incorporates CORA, time-to-peak (TTP) and peak magnitude of the experimental signals and ISO TS 18571 was used to evaluate the model response. For the horizontal impacts, the combined CORA scores were 0.80, 0.84, and 0.81 for compression, force, and strain respectively. For the vertical impacts combined CORA scores for the knee Z force, compression and heel Z displacement ranged from 0.70-0.81, 0.87-0.91, and 0.82-0.99 respectively. The GHBMC lower extremity model showed good agreement with PMHS experimental data in the horizontal and vertical loading environment in 33 unique tests. The accuracy is demonstrated by using the ISO TS 18571 standard and a combined CORA score that takes into consideration the peak and time to peak of the signal. The results of this study show that GHBMC v 6.0 HBM lower extremity can be used for kinetic and kinematic predictions in the UBB environment.</p>","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"66 ","pages":"99-142"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41136963","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Understanding Head Injury Risks During Car-to-Pedestrian Collisions Using Realistic Vehicle and Detailed Human Body Models.","authors":"Kalish Gunasekaran,&nbsp;Sakib Ul Islam,&nbsp;Haojie Mao","doi":"10.4271/2022-22-0006","DOIUrl":"10.4271/2022-22-0006","url":null,"abstract":"<p><p>Traumatic brain injury (TBI) is the leading cause of death and long-term disability in road traffic accidents (RTAs). Researchers have examined the effect of vehicle front shape and pedestrian body size on the risk of pedestrian head injury. On the other hand, the relationship between vehicle front shape parameters and pedestrian TBI risks involving a diverse population with varying body sizes has yet to be investigated. Thus, the purpose of this study was to comprehensively study the effect of vehicle front shape parameters and various pedestrian bodies ranging from 95th percentile male (AM95) to 6 years old (YO) child on the dynamic response of the head and the risk of TBIs during primary (vehicle) impact. At three different collision speeds (30, 40, and 50 km/h), a total of 36 car-to-pedestrian collisions (CPCs) were reconstructed using three different vehicle types (Subcompact passenger sedan, mid-sedan, and sports utility vehicle (SUV)) and four distinct THUMS pedestrian finite element (FE) models (AM50, AM95, AF05, and 6YO). We assessed skull stress and brain strains besides head linear and rotational kinematics. Our findings indicate that vehicle shape parameters especially bonnet leading edge height (BLEH), when being divided by the height of the Center of Gravity of the human body, correlated positively to head kinematics. The data from this study using realistic vehicle structures and detailed human body models showed that smaller BLEH/CG ratios reduced head injury criteria (HIC) and brain injury criteria (BrIC) values for the car center to mid-stance walking pedestrian impacts but with low-to-moderate R squared values between 0.2 to 0.5. Smaller BLEH/CG reduced head lateral bending velocities with R squared values of 0.57 to 0.63 for all impact velocities, and reduced HIC with R squared value of 0.62 for 50 km/h cases. In the future, simulations with realistic car structures and detailed human body models will be further used to simulate impacts at different locations and with various body shapes/postures.</p>","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"66 ","pages":"175-205"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41158181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Obese Occupant Response in Reclined and Upright Seated Postures in Frontal Impacts.","authors":"Karthik Somasundaram,&nbsp;John R Humm,&nbsp;Narayan Yoganandan,&nbsp;Hans Hauschild,&nbsp;Klaus Driesslein,&nbsp;Frank A Pintar","doi":"10.4271/2022-22-0002","DOIUrl":"https://doi.org/10.4271/2022-22-0002","url":null,"abstract":"<p><p>The American population is getting heavier and automated vehicles will accommodate unconventional postures. While studies replicating mid-size and upright fore-aft seated occupants are numerous, experiments with post-mortem human subjects (PMHS) with obese and reclined occupants are sparse. The objective of this study was to compare the kinematics of the head-neck, torso and pelvis, and document injuries and injury patterns in frontal impacts. Six PMHS with a mean body mass index of 38.2 ± 5.3 kg/m2 were equally divided between upright and reclined groups (seatback: 23°, 45°), restrained by a three-point integrated belt, positioned on a semi-rigid seat, and exposed to low and moderate velocities (15, 32 km/h). Data included belt loads, spinal accelerations, kinematics, and injuries from x-rays, computed tomography, and necropsy. At 15 km/h speed, no significant difference in the occupant kinematics and evidence of orthopedic failure was observed. At 32 km/h speed, the primary difference between the cohorts was significantly larger Z displacements in the reclined occupant at the head (190 ± 32 mm, vs. 105 ± 33 mm p < 0.05) and femur (52 ± 18 mm vs. 30 ± 10 mm, p < 0.05). All the moderate-speed tests produced at least one thorax injury. Rib fractures were scattered around the circumference of the rib-cage in the upright, while they were primarily concentrated on the anterior aspect of the rib-cage in two reclined specimens. Although MAIS was the same in both groups, the reclined specimens had more bi-cortical rib fractures, suggesting the potential for pneumothorax. While not statistical, these results suggest enhanced injuries with reclined obese occupants. These results could serve as a data set for validating the response of restrained obese anthropometric test device (ATDs) and computational human body models.</p>","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"66 ","pages":"31-68"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"41131425","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"THOR-05F Response in Sled Tests Inducing Submarining and Comparison with PMHS Response Corridors.","authors":"Olivier Richard, M. Lebarbé, J. Uriot, X. Trosseille, P. Petit, Z. J. Wang, Ellen Lee","doi":"10.4271/2021-22-0005","DOIUrl":"https://doi.org/10.4271/2021-22-0005","url":null,"abstract":"The Test Device for Human Occupant Restraint (THOR) is an advanced crash test dummy designed for frontal impact. Originally released in a 50th percentile male version (THOR-50M), a female 5th version (THOR-05F) was prototyped in 2017 (Wang et al., 2017) and compared with biofidelity sub-system tests (Wang et al., 2018). The same year, Trosseille et al. (2018) published response corridors using nine 5th percentile female Post Mortem Human Subjects (PMHS) tested in three sled configurations, including both submarining and non-submarining cases. The goal of this paper is to provide an initial evaluation of the THOR-05F biofidelity in a full-scale sled test, by comparing its response with the PMHS corridors published by Trosseille et al. (2018). Significant similarities between PMHS and THOR-05F were observed: as in Trosseille et al. (2018), the THOR-05F did not submarine in configuration 1, and submarined in configurations 2 and 3. The lap belt tension and seat forces were similar in magnitude. For configurations 2 and 3, the pelvis excursions were of the same order of magnitude between both human surrogates. However, significant differences were also observed: compared to the PMHS, the THOR-05F showed shoulder belt forces that were 1.6 to 2.1 times higher in magnitude, and lap belt force time histories that were delayed by 10 to 20 ms. In configuration 1, the chest and pelvis resultant accelerations of the dummy were delayed as well, and the pelvis excursion and rotation more than doubled that of the PMHS.","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"65 1","pages":"91-138"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"49353241","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Quantifying the Effect of Pelvis Fracture on Lumbar Spine Compression during High-rate Vertical Loading.","authors":"D. Barnes, N. Yoganandan, Jason Moore, J. Humm, F. Pintar, K. Loftis","doi":"10.4271/2021-22-0008","DOIUrl":"https://doi.org/10.4271/2021-22-0008","url":null,"abstract":"Fracture to the lumbo-pelvis region is prevalent in warfighters seated in military vehicles exposed to under-body blast (UBB). Previous high-rate vertical loading experimentation using whole body post-mortem human surrogates (PMHS) indicated that pelvis fracture tends to occur earlier in events and under higher magnitude seat input conditions compared to lumbar spine fracture. The current study hypothesizes that fracture of the pelvis under high-rate vertical loading reduces load transfer to the lumbar spine, thus reducing the potential for spine fracture. PMHS lumbo-pelvis components (L4-pelvis) were tested under high-rate vertical loading and force and acceleration metrics were measured both inferior-to and superior-to the specimen. The ratio of inferior-tosuperior responses was significantly reduced by unstable pelvis fracture for all metrics and a trend of reduced ratio was observed with increased pelvis AIS severity. This study has established that pelvis fracture reduces compression forces at the lumbar spine during high-rate vertical loading, thus reducing the potential for fracture to the lumbar spine. Therefore, pelvis injury potential should be considered when implementing lumbar injury criteria specific to UBB.","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"65 1","pages":"189-216"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"43560080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Instantaneous Brain Strain Estimation for Automotive Head Impacts via Deep Learning.","authors":"Shaoju Wu, Wei Zhao, S. Barbat, J. Ruan, Songbai Ji","doi":"10.4271/2021-22-0006","DOIUrl":"https://doi.org/10.4271/2021-22-0006","url":null,"abstract":"Efficient brain strain estimation is critical for routine application of a head injury model. Lately, a convolutional neural network (CNN) has been successfully developed to estimate spatially detailed brain strains instantly and accurately in contact sports. Here, we extend its application to automotive head impacts, where impact profiles are typically more complex with longer durations. Head impact kinematics (N=458) from two public databases were used to generate augmented impacts (N=2694). They were simulated using the anisotropic Worcester Head Injury Model (WHIM) V1.0, which provided baseline elementwise peak maximum principal strain (MPS). For each augmented impact, rotational velocity (vrot) and the corresponding rotational acceleration (arot) profiles were concatenated as static images to serve as CNN input. Three training strategies were evaluated: 1) \"baseline\", using random initial weights; 2) \"transfer learning\", using weight transfer from a previous CNN model trained on head impacts drawn from contact sports; and 3) \"combined training\", combining previous training data from contact sports (N=5661) for training. The combined training achieved the best performances. For peak MPS, the CNN achieved a coefficient of determination (R2) of 0.932 and root mean squared error (RMSE) of 0.031 for the real-world testing dataset. It also achieved a success rate of 60.5% and 94.8% for elementwise MPS, where the linear regression slope, k, and correlation coefficient, r, between estimated and simulated MPS did not deviate from 1.0 (when identical) by more than 0.1 and 0.2, respectively. Cumulative strain damage measure (CSDM) from the CNN estimation was also highly accurate compared to those from direct simulation across a range of thresholds (R2 of 0.899-0.943 with RMSE of 0.054-0.069). Finally, the CNN achieved an average k and r of 0.98±0.12 and 0.90±0.07, respectively, for six reconstructed car crash impacts drawn from two other sources independent of the training dataset. Importantly, the CNN is able to efficiently estimate elementwise MPS with sufficient accuracy while conventional kinematic injury metrics cannot. Therefore, the CNN has the potential to supersede current kinematic injury metrics that can only approximate a global peak MPS or CSDM. The CNN technique developed here may offer enhanced utility in the design and development of head protective countermeasures, including in the automotive industry. This is the first study aimed at instantly estimating spatially detailed brain strains for automotive head impacts, which employs >8.8 thousand impact simulations generated from ~1.5 years of nonstop computations on a high-performance computing platform.","PeriodicalId":35289,"journal":{"name":"Stapp car crash journal","volume":"65 1","pages":"139-162"},"PeriodicalIF":0.0,"publicationDate":"2021-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"47352355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}