The knee is one of the regions of interest for pedestrian safety assessment. Past testing to study knee ligament injuries for pedestrian impact only included knees in full extension and mostly focused on global responses. As the knee flexion angle and the initial ligament laxity may affect the elongation at which ligaments fail, the objectives of this study were (1) to design an experimental protocol to assess the laxity of knee ligaments before measuring their elongation at failure, (2) to apply it in paired knee tests at two flexion angles (10 and 45 degrees). The laxity tests combined strain gauges to measure bone strains near insertions that would result from ligament forces and a custom machine to exercise the knee in all directions. Failure was assessed using a four-point bending setup with additional degrees of freedom on the axial rotation and displacement of the femur. A template was designed to ensure that the two setups used the exact same starting position. The protocol was applied to six pairs of knees which were tested until the failure of all ligaments. In the laxity tests, a higher compliance of the knee was observed at 45 degrees compared to 10 degrees. Minimum lengths associated with the beginning of bone loading were also successfully identified for the collateral ligaments, but the process was less successful for the cruciate ligaments. The failure tests suggested increased elongation and length at failure for the ligaments and their bundles at 45°. This could be consistent with the higher compliance in static test, but the minimum lengths identified on the collaterals did not explain this difference during failure. The results highlight the possible relationship between position, laxity and elongation at failure in a lateral loading and provide a dataset including 3D coordinates of insertions to continue the investigation using a modelling approach. Perspectives are also outlined to improve upon the laxity determination protocol.

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.

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.

Far side has been identified in the literature as a potential cause of numerous injuries and fatalities. Euro NCAP developed a far side test protocol to be performed to assess adult protection. A monitoring phase was undertaken between January 2018 and December 2019, and the far side assessment will become part of the rating for all vehicles launched in 2020 onward. A test buck was developed and 6 paired WorldSID / Post Mortem Human Subjects (PMHS) were subjected to the test protocol proposed by Euro NCAP to contribute to the development of limits. The buck consisted of a rigid seat and a rigid central console covered with 50 mm of Ethafoam TM 180 with a density of 16 kg/m3. The buck was mounted on the sled with an angle of 75° between the X axis of the vehicle and the X axis of the sled. The peak head excursion was compared between PMHS and the WorldSID dummy. It was found reasonably similar. However, the dummy repeatability was found to be poor. Out of 6 tests conducted on 6 PMHS, 2 specimens sustained AIS3 and, 3 specimens AIS2 cervical spine injuries, 3 specimens sustained AIS3, 1 AIS2 and 1 AIS1 thoracic injuries, and 2 specimens sustained AIS2 abdominal injuries. The peak values recorded on the dummy according to the Euro NCAP protocol were compared with the injury assessments of the PMHS tests. In the configuration used, which includes a central console, the hard thorax injury prediction was found to be excellent. For the neck injury prediction, the data were merged with similar results available in the literature and an Injury Risk Curve was proposed as a derivative from the curve published by Mertz et al. (2003) for neck extension.

Several studies, available in the literature, were conducted to establish the most relevant criterion for predicting the thoracic injury risk on the THOR dummy. The criteria, such as the maximum deflection or a combination of parameters including the difference between the chest right and left deflections, were all developed based on given samples of Post Mortem Human Subject (PMHS). However, they were not validated against independent data and they are not always consistent with the observations from field data analysis. For this reason, 8 additional PMHS and matching THOR tests were carried out to assess the ability of the criteria to predict risks. Accident investigations showed that a reduction of the belt loads reduces the risk of rib fractures. Two configurations with different levels of force limitation were therefore chosen. A configuration representing an average European vehicle was chosen as a reference. It consists of a 3-point belt with a 3.5 kN and then 2 kN digressive limiter, combined with a 54-liter airbag. For better reproducibility and durability, the tests were performed with a pre-inflated bag and a semi-rigid seat. In this first configuration, the THOR dummy had a maximum resulting deflection of 43 mm. To differentiate the criteria, the second configuration was chosen such that it resulted in about the same deflection on the THOR dummy, but with a 5 kN belt force limitation combined with a lower pressure airbag. To reach this target of 43 mm, the pulse severity was lowered. Some criteria were higher in this second configuration, which allows them to be distinguished from the maximum deflection criterion. Four tests on four PMHS were performed in each configuration. The injury assessments showed that the total number of fractures was almost the same in both configurations, but that the number of separated fractures was greater in the 5 kN configuration. 25% of the subjects sustained AIS >3 injuries related to the number of displaced fractures in the 3.5/2 kN load limitation configuration. The result increased to 75% in the 5kN configuration. In total, 8 PMHS and the matching THOR tests were performed and used to assess the ability of the thoracic criteria to predict rib fractures in 2 types of chest loading configurations. The test results did not allow to conclude on the relevance of the criteria measured on the THOR dummy for the total number of rib fractures identified at autopsy (NFR). However, clearly different assessments for separated rib fractures (NSFR), make it possible to differentiate the criteria. The maximum resultant deflection failed to properly predict separated rib fractures while other criteria that include the left-to-right rib deflection difference did.

There are few experimental data in literature relating to femur cross-section moments in full-scale PMHS pedestrian tests. However, these moments are used as injury criteria to evaluate pedestrian protection. This study aimed to provide this type of data. In previous studies by the authors, 15 post-mortem human subject (PMHS) were impacted by four different generic simplified vehicle bucks (two sedans, a SUV and a van). Each PMHS was instrumented with three mono-axis strain gauges on its right femur (impacted before the left femur). To derive the moments in the femur from these strain measurements, three-point bending tests were performed on the femurs isolated from the corresponding PMHS tested. The moment-strain relationship for each strain measurement location was then established for each femur. By applying these relationships to the strains measured in the full-scale PMHS tests, the corresponding cross-section moments on the femur were derived. Finally, by scaling these moments, biofidelity corridors were established for the 50th percentile male pedestrian in terms of upper, middle and lower cross-section moments in the femur, and for four vehicle profiles. These corridors will allow assessing biofidelity of human body models, physical dummies and lower extremity impactors.

The EuroSID-2re (ES-2re) Anthropomorphic Test Device (ATD) commonly known as the crash test dummy is also used in the military domain to assess the risk of injury of armored vehicles occupants from lateral impact. The loading conditions range from low velocity - long duration impacts (4 m/s - 50 ms) similar to the automotive domain, to high velocity - short duration impacts (28 m/s - 3 ms) corresponding to cases where the panel deforms under an explosion. The human shoulder response to lateral impact was investigated at bounds of the loading condition spectrum previously mentioned, and also at intermediate conditions (14 m/s - 9 ms) in previous studies. The aim of the current study is to provide additional insight at the intermediate loading conditions which are not found in the literature. Eight pure lateral shoulder impact tests were performed on Post Mortem Human Subjects (PMHS) using an 8.1 kg rigid impactor at velocities ranging from 3.3 m/s to 8.8 m/s with the duration ranging from 25 ms to 35 ms. The PMHS were instrumented with accelerometers attached to the sternum, and the upper thoracic spine (T1 vertebra). Strain gages were glued onto the right and left clavicles and ribs 2 to 6. The shoulder force was measured at the interface with the impactor and the impact was filmed by high speed cameras (5000 fps) to track the YZ displacements of the impactor, T1 vertebra, and sternum in the laboratory frame. Three shoulders out of the eight sustained AIS 2 injuries which included a clavicle fracture. The impactor forces ranged from 1200 to 4600 N. The PMHS accelerations ranged from 44 to 163 g at the sternum, and from 17 to 60 g at the T1 vertebra. The analysis of the strain gage signals revealed that the clavicle fractures occurred at the beginning of the impact and coincided with a peak force. An estimate of the acromion-to-shoulder compression (Cmax) was computed. It ranged from 0% to 15% for the non-injured shoulders, and from 19% to 28% for the injured shoulders. This new PMHS test series will be used in a future work to develop a shoulder injury criterion for the ES-2re ATD that is relevant for the whole loading conditions spectrum of the military domain.

In the last decade, extensive efforts have been made to understand the physics of submarining and its consequences in terms of abdominal injuries. For that purpose, 27 Post Mortem Human Subject (PMHS) tests were performed in well controlled conditions on a sled and response corridors were provided to assess the biofidelity of dummies or human body models. All these efforts were based on the 50th percentile male. In parallel, efforts were initiated to transfer the understanding of submarining and the prediction criteria to the THOR dummies. Both the biofidelity targets and the criteria were scaled down from the 50th percentile male to the 5th percentile THOR female. The objective of this project was to run a set of reference PMHS tests in order to check the biofidelity of the THOR F05 in terms of submarining. Three series of tests were performed on nine PMHS, the first one was designed to avoid submarining, the second and third ones were designed to result in submarining. In the first configuration, no submarining was observed in 3 cases out of 4 and only one iliac wing fracture occurred in one subject. In the second and third configurations, all subjects but one sustained submarining. In addition, two subjects out of three in the third configuration sustained substantial iliac wing fractures. Nevertheless, all configurations can be represented by at least one or several cases without any pelvis fracture. Corridors were constructed for the external forces and the PMHS kinematics. They are provided in this paper as new experimental references to assess the biofidelity of small female human surrogates in different configurations where submarining did or did not occur.

In a previous study (Song et al. 2017), an adjustable generic simplified vehicle buck was developed; eleven PMHS were impacted by the buck representing a SUV, a van and a sedan successively; and biofidelity corridors were established. The objectives of the current study were 1) to develop the computational model of the buck, and 2) to simulate these PMHS tests with the buck model and to assess the biofidelity of the GHBMC 50th percentile male pedestrian simplified model (GHBMC M50-PS). First, coupon tensile tests and static and dynamic compression tests were performed on the steel tubes representing the bonnet leading edge (BLE), the bumper and the spoiler used in the above PMHS tests. Based on these tests, the computational models of the above components were then developed and validated. Next, the buck model was built with the component models, and used to simulate the PMHS tests with the GHBMC M50-PS model. These simulations allowed to evaluate the biofidelity of the GHBMC M50-PS model in terms of 1) impact forces between the pedestrian and the buck, 2) pedestrian kinematics, and 3) injury outcome resulted. The model well predicted the total longitudinal impact force between the pedestrian and the buck for all three vehicle types, with a total CORA score between 0.72 and 0.78. However, the force distribution across the BLE, bumper and spoiler showed some significant deviations. The kinematic response of the model was rated as fair with a total CORA score ranging between 0.52 and 0.58. It seems necessary to increase the compliance of the GHBMC M50-PS model and its energy dissipation capability in order to achieve a better correlation of its kinematic response. Finally, the model predicted more knee ligament ruptures than observed in the PMHS tests, but less bone fracture of the femur and the fibula.

This study aims to provide a set of reference post-mortem human subject tests which can be used, with easily reproducible test conditions, for developing and/or validating pedestrian dummies and computational human body models against a road vehicle. An adjustable generic buck was first developed to represent vehicle front-ends. It was composed of four components: two steel cylindrical tubes screwed on rigid supports in V-form represent the bumper and spoiler respectively, a quarter of a steel cylindrical tube represents the bonnet leading edge, and a steel plate represents the bonnet. These components were positioned differently to represent three types of vehicle profile: a sedan, a SUV and a van. Eleven post-mortem human subjects were then impacted laterally in a mid-gait stance by the bucks at 40 km/h: three tests with the sedan, five with the SUV, and three with the van. Kinematics of the subjects were recorded via high speed videos, impact forces between the subjects and the bucks were measured via load cells behind each tube, femur and tibia deformation and fractures were monitored via gauges on these bones. Based on these tests, biofidelity corridors were established in terms of: 1) displacement time history and trajectory of the head, shoulder, T1, T4, T12, sacrum, knee and ankle, 2) impact forces between the subjects and the buck. Injury outcome was established for each PMHS via autopsy. Simplicity of its geometry and use of standard steel tubes and plates for the buck will make it easy to perform future, new post-mortem human subject tests in the same conditions, or to assess dummies or computational human body models using these reference tests.

Sled tests focused on pelvis behavior and submarining can be found in the literature. However, they were performed either with rigid seats or with commercial seats. The objective of this study was to get reference tests to assess the submarining ability of dummies in more realistic conditions than on rigid seat, but still in a repeatable and reproducible setup. For this purpose, a semi-rigid seat was developed, which mimics the behavior of real seats, although it is made of rigid plates and springs that are easy to reproduce and simulate with an FE model. In total, eight PMHS sled tests were performed on this semirigid seat to get data in two different configurations: first in a front seat configuration that was designed to prevent submarining, then in a rear seat configuration with adjusted spring stiffness to generate submarining. All subjects sustained extensive rib fractures from the shoulder belt loading. No pelvis fractures and no submarining were observed in the front seat configuration, but two subjects sustained lumbar vertebrae fractures. In the rear seat configuration, all subjects sustained pelvic fractures and demonstrated submarining. Corridors were constructed for the external forces and the PMHS kinematics. They are provided in this paper as new reference tests to assess the biofidelity of human surrogates in different configurations that either result in submarining or do not. In future, it is intended to analyze further seat and restraint system configurations to be able to define a submarining predictor.

Nine Post Mortem Human Subject (PMHS) tests were published by Luet et al. (2012) as a reference for submarining assessment. These tests suffered limitations, however, due to the rigid characteristics of the seat. The objective of this study is to assess the submarining ability of existing dummies in more realistic conditions, i.e. on a real seat. In total, nine PMHS sled tests were performed on a real seat to procure data in 3 different configurations. The tests were also duplicated with HII, HIII and THOR Mod Kit dummies to evaluate their behaviour in comparison to corridors built with 3 PMHS each. The external forces, the pelvis kinematics and the injury assessments are provided here. The dummy responses are compared to the PMHS corridors. No submarining was observed in the standard configuration. Iliac crest fractures and submarining were observed in the slouched position intended to generate submarining. Finally, for two subjects out of three, no iliac wing fracture and no submarining were observed in the slouched position with a lap‐belt load limitation at 4 kN and a pretension system.

A specific test setup was designed to recreate in a laboratory the features of a deforming armoured sidewall of a military veh icle submitted to an Improvised Explosive Device (IED) blast. A free moving impactor of 7 kg weight and 10 cm diameter contact face was propelled at 27 m/s and decelerated over 4 cm using copper tubes in order to control the penetration. Four tests on ES-2re 50 th percentile dummy were performed to check on the similarity of the impact with an IED blast case. The Y shoulder force sensor time history signals were similar to ones recorded in full-scale IED blast tests. Six tests on Post Mortem Human Subjects were performed at different violence levels. The PMHS maximum deflection s (from shoulder edge to sternum) and reaction force ranged from 3.2 cm to 4.7 cm, and from 5.7 kN to 11.7 kN, respectively. The three PMHS tests on which the highest forces were recorded presented moderate injury (AIS2) on the 2005 Abbreviated Injury Scale. Injuries were mainly humerus fractures (shaft and head). One subjec t sustained multiple fractures of the scapula. These results presented different features from those of previous studies that used lower impact rates (below 7 m/s). These results could be used to develop an AIS2+ shoulder injury criterion specific to these kinds of military impacts, using the ES-2re dummy.

This study focuses on the phenomenon of lap belt slip on the iliac spines of the pelvis, commonly named "submarining". The first objective was to compare the interaction between the pelvis and the lap belt for both dummies and Post Mortem Human Subjects (PMHS). The second objective was to identify parameters influencing the lap belt hooking by the pelvis. For that purpose, a hydraulic test device was developed in order to impose the tension and kinematics of the lap belt such that they mimic what occurs in frontal car crashes. The pelvis was firmly fixed on the frame of this sub-system test-rig, while the belt anchorages were mobile. Fourteen tests on four Post-Mortem Human Subjects (PMHS) and fifteen tests on the THOR NT, Hybrid III 50th and Hybrid III 95th percentile dummies were carried out. The belt tension was kept constant while a dynamic rotation was imposed on the belt anchorages. The test results show that, in the tests where submarining occurred, the belt angle relative to the pelvis was systematically greater for the pelves of dummies than for those of PMHS. An increase in the belt tension resulted in a greater submarining angle for both dummies and PMHS. The pelvis and abdomen stiffnesses were observed to be similar for the PMHS and the THOR NT dummy while significantly lower for the Hybrid III dummies.

The improvement of neck protection in rear impact requires a better understanding of the interactions between occupant, seatback and headrest. In this study a mathematical approach was developed to analyse the interactions, and to quantify the influences of different design parameters on neck responses. The first phase of this development consisted of neck modelling. For the second phase of the development, the interactions of the thorax and the pelvis with the seatback were modelized, as was that of the head and the headrest. Two types of seatback model were constructed: a global seatback modelling, and a more detailed approach. Finally, the influences of the following four design parameters on neck responses were analysed: head to headrest distance, seatback joint stiffness, and upper and lower seatback stiffnesses. Special attention was paid to the interactions between these parameters. The results indicate that softening of the upper seatback allows reduction of all neck injury risk indicators and enhances the headrest performance. Softening of the lower seatback increases the moment force at the C7/T1 joint and the head extension angle. Stiffening of the seatback joint aggravates, for a classical upper seatback structure, the moment loading of the neck and the head extension angle. Only the moment force at the C7/T1 joint is significantly affected by all parameter changes. For the covering abstract of the conference see IRRD E201282.

The objective of this study is to evaluate two applications of a multi- accelerometric method, already developed for dummies, to car safety tests. In the first case, the contribution of the angular acceleration to the analysis of two types of head impacts is evaluated. The study shows that it is possible to refine the head impact analysis. The second application consists of the calculation of steering column rearward displacement in a frontal crash test. This application shows that it is possible to approximate the kinematics of the steering wheel without having to analyze a film. The displacements obtained from the accelerometric approach are compared with those provided by the convential film analysis. This comparison shows acceptable convergence during the whole dynamic motion of the steering wheel column. The paper briefly describes how the steering wheel column rearward displacement analysis can be optimized. For the covering abstract of the conference see IRRD 864606.