Occupant Kinematic and Injury Responses in Zero-Gravity Seat Under Low-, Medium-, and High-Speed Rear Impacts with Different Seat Belt Systems

Abstract

This study investigates occupant kinematic and injury responses in zero-gravity seats under rear impacts at 16 km/h, 40 km/h, and 56 km/h and evaluates the protective performance of a conventional three-point seat belt system and a four-point seat belt system. First, a THUMS (Total Human Model for Safety)-based finite element assembly consisting of a regular seat model and a conventional three-point seat belt system was verified by comparing the kinematic responses and time-history curves of head acceleration, head rotation, and the T1 acceleration of PMHS (Postmortem Human Subject) tests. Then, a THUMS-based finite element assembly in a zero-gravity seat with a three-point seat belt system was created, and computational biomechanical analyses revealed that at low-to-medium impact speeds (16 and 40 km/h), the occupant exhibited backward sliding in the zero-gravity seat along the seatback with lower limb rotation and did not experience head and neck injury. However, a 56 km/h impact induced an excessive seatback rotation and caused the head to become out of position. The neck collided with the upper part of the headrest and caused a surge in the contact force between the neck and the headrest. The head injury and neck injury were comprehensively analyzed via the head injury metrics and neck injury metrics, including cervical spine injury metrics and cervical ligament injury metrics. Further, a four-point seat belt system was adopted and demonstrated better and more balanced restraining effects by reducing the relative displacement between the occupant’s head and chest in the x- and y-directions by 26% and 84%, respectively. Therefore, the occupant’s head remains in position and the collision between the neck and the headrest can be avoided. Maximum reductions in the head and neck injury metrics reached 70% and 57%, respectively. The current study illustrates the disadvantages of the traditional three-point seat belt system in restraining the occupant in a zero-gravity seat under rear impact and shows the four-point seat belt to be a better alternative. This study sheds light on seat belt system design and optimization towards future zero-gravity seats under rear impact.

1. Introduction

With the continuous advancement of intelligent and connected automotive technologies, driving tasks have been significantly reduced and may no longer be required in the future. Car buyers often look for features that enhance ride comfort, especially given the increasing prevalence of traffic congestion and long commutes in major cities. Research into zero-gravity seats originates from NASA’s research in the 1980s on neutral body posture (NBP), a natural floating posture that astronauts experience in a zero-gravity space environment [1,2]. The NBP is characterized by a semi-crouched torso, flexed arms and legs, and bent-forward head and neck. Nissan Motor Company [3] first introduced the NBP design concept into automotive car seat design with a two-piece backrest configuration in 2005 and illustrated that zero-gravity seats can improve support for the spine and the region from the pelvis to the chest, enhancing blood circulation and significantly reducing fatigue during prolonged sitting.

As an innovative technological product, automotive zero-gravity seats are increasingly attracting widespread attention from both the consumer and automotive manufacturers, particularly in the highly competitive Chinese electric automotive market. Compared to conventional car seats that only offer seatback recline adjustment, zero-gravity seats additionally provide adjustments for the seat cushion angle, leg support, and armrests [4]. As of today, nearly 20 vehicle models equipped with zero-gravity seats have been introduced to the Chinese market. In 2022, the AITO M7 became the first vehicle manufactured in China to feature zero-gravity seats. The AITO M7’s second-row zero-gravity seats feature a range of adjustment modes designed to maximize passenger comfort. With the adjustable backrest, seat cushion, and footrest, the zero-gravity seats provide the passenger with a nearly fully reclined seating posture. Additionally, several mass-produced models, both domestic and international, such as the Kia Carnival, Hyundai Custo, and IM LS7, have also incorporated zero-gravity seats in the front passenger or second-row seating positions [5]. However, this type of seat is currently only designed for use in a parked state due to a lack of safety protection strategies; it is essential to investigate the safety features of such seats to expand their usage scenarios to regular driving scenarios.

Currently, research on zero-gravity seats primarily focuses on injury assessment and prevention strategies regarding the lumbar spine and submarining in frontal crashes [6,7,8]. However, studies of zero-gravity seats in rear-impact settings are very few and limited, although they are crucial for optimizing automotive seat and restraint system design to mitigate head or neck-related injuries, such as whiplash injuries and for enhancing the occupant’s safety by ensuring proper head and neck support, energy absorption, and controlled seatback motion during collisions [9,10]. The most relevant research into rear impacts with zero-gravity seats is the study of rear impacts with reclined seats, since a reclined seating position is an essential feature of zero-gravity seats. Hasija et al. [11] employed the Global Human Body Models Consortium (GHBMC) 50th percentile simplified (M50-OS) male finite element model and a Honda Accord seat model to evaluate occupant injury risks under low-speed (15 mph) and high-speed (35 mph) rear-impact conditions with varying seatback recline angles. The injury metrics included the Head Injury Criterion (HIC), Brain Injury Criterion (BrIC), and maximum chest deflection. The study found that increased seatback recline angles during rear impacts correlated with higher BrIC values, indicating elevated brain injury risks. However, the maximum initial occupant recline angle in this study was only 45°, significantly smaller than the typical angles of zero-gravity seats, which are larger than 50°. Additionally, the injury assessment focused solely on head and chest injuries, neglecting neck injury metrics such as N_ij or neck ligament strain, which are critical for evaluating occupant protection in highly reclined configurations. Anh Vu Ngo et al. [12] investigated the effects of hinge failure on occupant injuries for different reclined seatbacks under various rear-impact speeds using a production seat model and the THUMS (Total Human Model for Safety). The study showed that, under moderate- and high-speed rear impacts, the threshold of hinger failure was reached for a series production seat and the neck injuries were reduced; on the contrary, the head and chest injuries were exacerbated. However, the considered seatback recline angles were 22° and 45°, which were significantly smaller than the recline angle of zero-gravity seats. Additionally, the research did not address protective strategies for neck injuries, which is highly associated with reclined seating configurations. Wu et al. [13] investigated occupant injuries for medium- and high-impact speed cases with the consideration of a wide range of seatback recline angles, including 30°, 60°, and 90°. The study revealed that at seatback angles of 60° and 90°, the head displacement increased and the head detached from the headrest, while the neck experienced stretching, leading to injuries in the anterior longitudinal ligament of the neck. This phenomenon was attributed to two factors. One is that larger seatback angles reduced the headrest’s ability to stabilize head posture. Another one is that the conventional three-point seat belts failed to effectively restrain horizontal occupant motion, yielding excessive head displacement.

With respect to the widely used three-point seat belt restraint system, the four-point seat belt system seems a better way to mitigate head movement by restraining both shoulders instead of one shoulder. Two types of four-point seat belt systems [14], the X4 and V4 belts, have been utilized to reduce chest injuries during frontal crashes. Further, X-type four-point seat belts [15] have been utilized to restrain occupants for the comfortability study of highly reclined seats under autonomous driving scenarios. However, the protective performance of seat belts under rear impacts has never been evaluated.

A review of the literature indicates that the current research on occupant injuries in rear impacts primarily focuses on reclined seats, and there is a lack of comprehensive studies on zero-gravity seats. For zero-gravity seats, in addition to seatback recline angles, evaluations need to account for seat cushion angle adjustments, which may result in a larger recline angle for occupants and footrest configurations. Further, it is worth mentioning that the head’s out-of-position posture and neck injuries remain common issues in highly reclined seats during rear collisions, and currently there is a lack of effective protection strategies for these scenarios.

This study aims to investigate the kinematics responses and injury mechanisms of an occupant restrained in a zero-gravity seat under varying rear-impact speeds and evaluate the effectiveness of different restraint systems. Firstly, the THUMS-based finite element model assembled with a production seat and three-point seat belt system were validated by comparing kinematics and head accelerations of THUMS with PMHS (Postmortem Human Subject) test data from standard seating postures. Based on a validated model of the seat and the corresponding seat belt system, a zero-gravity seat model incorporating adjustable seat cushion angles and footrest configurations was created. When the occupant was restrained by the three-point seat belt system, the simulation results revealed significant increases in head and neck injuries when the impact speed increased to 56 km/h due to the head’s out-of-position posture and the following collision between the head and neck and the headrest. To mitigate these injury risks, a four-point seat belt system was proposed, demonstrating a significant improvement of head and neck injury protection by redistributing restraint forces on the occupant’s torso and shoulders. These findings provide critical insights for a better understanding of the occupant injury mechanism and restraint strategy design and optimization towards zero-gravity seat designs under various rear-collision speeds.

2. Materials and Methods

2.1. Validation of the Finite Element Model of Occupant Restraint Systems with a Regular Seat Model and a Three-Point Seat Belt System

The validity of the finite element model of occupant restraint systems directly affects the prediction accuracy of occupant kinematics and subsequent injury risk assessment during collision events. The Total Human Model for Safety (THUMS ver6.1) developed by Toyota was employed as the simulation model, and the seat model was derived from the 2012 Toyota Camry full-vehicle model [16], a publicly available vehicle model widely utilized by the global safety and crash research community. The THUMS incorporates detailed anatomical structures, including skeletal components, muscles, internal organs, ligaments, and soft tissues, which enable the comprehensive simulation of biomechanical behaviors across human tissues. As an initiative study to understand an occupant’s kinematic and injury responses under different rear-impact speeds, only the 50th adult male human model was considered since it represents one of the largest population groups. During this study, both the seat and biomechanical models were validated by Postmortem Human Subject (PMHS) experiments from Kang et al. [17,18], and PMHS2 was used as the reference anatomy. The H-point position of the THUMS was adjusted to an 80 mm vertical distance from the head center of gravity to the head restraint upper surface with a 55 mm minimum horizontal gap between the head and head restraint surface, ensuring alignment with the PMHS test configurations. The arms of the THUMS were positioned to naturally hang alongside the thighs through a postural adjustment process. The parameters of the seat belt system are shown in

Table 1. Seat belt system parameters.

Seat Belt Thickness (mm)	Seat Belt Width (mm)	Retractor to Seatback (mm)	Shoulder Guide Loop to Centerline (mm)
1.2	48	72	183

. A three-point seat belt system without retractors or pretensioners constrained the human body model in the seat. Standard adult seat belts were utilized with 48 mm width and 1.2 mm thickness. The retractor was positioned 72 mm inward from the seatback structural frame. A shoulder guide loop was installed 183 mm lateral from the seat centerline at the headrest’s right side, positioned slightly above the backrest upper edge. The guide loop and buckle components were rigidly connected to the seat frame structures. The seat model and human body were positioned and matched in Hyperworks, and the completed finite element model is shown in Figure 1a. The prescribed acceleration pulse [17] in Figure 1b was applied to the vehicle floor. Acceleration and displacement history curves of the head center of gravity and cervical vertebrae were recorded.

After the creation of the finite element simulation model, the model was analyzed in LS-DYNA, and the corresponding animation result files and time-history data were obtained. As shown in Figure 2, the simulated animation results postprocessed in Hyperview were directly compared with the video footage taken by the high-speed camera at 0, 60, 80, 100, 120, 140, and 160 ms in the PMHS tests [18]. In the PMHS tests, from 0 to 60 ms, the occupant’s chest was pushed forward by the seatback, but the occupant’s head was delayed due to the inertia effect and remained in the initial position. In the THUMS analysis, during that period, the occupant’s chest moved forward while the head remained stationary, illustrating a similar kinematic response with the PMHS tests. From 60 to 120 ms, the PMHS occupant’s upper body moved further forward, and the occupant’s head began to contact the headrest around 60 ms. After that, the occupant’s head remained in contact with the headrest. During that period in the THUMS simulation, the occupant’s head also moved backward relative to the chest, began to contact the headrest at 61 ms, and remained in contact with headrest. From 120 to 160 ms, as the PMHS occupant’s head continued moving backward to the maximum position relative to the chest, it rebounded forward and gradually separated from the headrest. During that period, the THUMS simulation showed a similar movement, and the occupant’s head rebounded forward and separated from the headrest at 133 ms. Therefore, the kinematic responses of the THUMS were in good agreement with the PMHS tests during the whole collision period.

In terms of the time-history data, the head accelerations, head rotation angle, and thoracic spine vertebral accelerations were processed in Hypergraph, and they were compared with the PHMS test data recorded by sensors in the study conducted by Kang et al. [17]. As shown in Figure 3, the predicted head acceleration in the x-direction, the head rotation about the y-axis, and the first thoracic spine vertebral (T1) acceleration in the x-direction all have a good agreement with the PMHS test, and the errors between the peak values are 5.5%, 0.9% and 11.6%, respectively. Further, a Correlation Analysis (CORA) [19] was conducted by Hypergraph to quantitatively evaluate the time histories between and simulation results and the PMHS test. The phase, amplitude, and slope of the time-history curves of the PMHS test and predicted responses were compared, and then the correlation scores were computed to evaluate the similarity of the two curves. The relative score ranges from 0 to 1, with 1 indicating a perfect match.

Table 2. Correlation Analysis for the model validation.

Response	Corridor Score	Cross-Correlation Scores	CORA Score
Head acceleration	0.72	0.74	0.73
T1 acceleration	0.71	0.88	0.79
Head rotation about y	0.67	0.81	0.74

lists the corridor scores, cross-correlation scores, and CORA scores of the time-history curves of the head acceleration in the x-direction, the head rotation angle about the y-axis, and T1 acceleration in the x-direction, and specifically, the corresponding CORA scores are 0.73, 0.74, and 0.79, respectively. Overall, the simulation results have good agreement with the PMHS test data. Therefore, the effectiveness of the THUMS-based finite element model of occupant restraint systems with a 3-point seat belt system and a conventional seat was sufficiently validated. Accordingly, the THUMS-based finite element model assembly was able to provide good accuracy for the subsequent computational biomechanics analysis of occupants in zero-gravity seats with different restraint systems.

2.2. Creation of the Finite Element Model of the Zero-Gravity Seat

The finite element model of the zero-gravity seat was constructed based on the previously validated Toyota Camry front passenger seat model. To eliminate uncertified material failure modes, particularly given production seat structures’ limited capacity to withstand high-speed rear impact, the seat frame components were reinforced as rigid components. However, to capture the classical rearward rotation phenomenon of the seatback of production seats [20], a flexible hinge between the seatback and seat pan structure was implemented into the current zero-gravity seat by referencing a validated finite element model from a 2014 Honda Accord seat [21]. In the seat model, rotational joints and linear beam elements were integrated at the backrest hinge to simulate backward tilting during collision events, and backrest rotation was governed solely by the axial rotational stiffness of the beam element.

The automotive industry witnessed a milestone in seat technology development when Nissan’s 2013 Altima became the first production vehicle to implement zero-gravity seat systems derived from NASA research [3]. The seatback and seat pan angles were adjusted to achieve a thigh–torso angle that was close to the NASA-defined neutral posture of 128° (±7°). Very recently, this microgravity-inspired technology has been rapidly adopted by various seat suppliers or Original Equipment Manufacturers (OEMs), particularly in the Chinese market [5]. However, it is worth mentioning that the above angular parameters vary to some degree for different designs and applications from different automotive companies. For instance, AITO M9’s zero-gravity seats offer a seat pan to seatback angle of 121° [22]. To create a zero-gravity seat for the current study, the backrest angle of the validated seat model in Section 2.1 was adjusted to 55° by rotating the backseat 30°, and the seat cushion inclination was set to 25° by raising the seat pan 15°. This seating configuration, illustrated in Figure 4, enables occupants to achieve a 120° seat pan to seatback angle, replicating the neutral body posture observed in microgravity environments.

2.3. Finite Element Model Assembly of Occupant Restraint Systems in Zero-Gravity Seats with Different Seat Belt Systems

The THUMS was integrated with the zero-gravity seat model through posture adjustment of the THUMS and the specification of contact in terms of Hypermesh. The lumbar spine of the THUMS was stretched to be parallel to the seatback, and the femur was revolved around the femoral acetabular joint to fit the seat cushion. In addition, the knee joint was stretched to be parallel with the footrest. The surface-to-surface contacts were defined between the THUMS and the seat, and the nodes-to-surface contact was defined between the THUMS and the seat belts. Subsequently, gravity was applied to the finite element model assembly, and the THUMS posture was adjusted until the back and legs were fully fitted to the seat, with the arms hanging naturally downward on both sides.

After the integration of the THUMS and the zero-gravity seat, conventional integrated three-point seat belts with a retractor, load limiter, and pretensioner were created, as shown in Figure 5a. Then, the four-point seat belt system was developed by augmenting the existing three-point configuration with the addition of a shoulder belt on the seat’s left side without altering any seat belt parameters. The anchor point configuration of the abdominal belt remained consistent with the three-point seat belt design. The finalized finite element model of the THUMS and the zero-gravity seat with four-point seat belt systems is presented in Figure 5b.

2.4. Rear-Impact Loading Conditions and Injury Measurement

In this work, a computational biomechanics analysis was performed to simulate rear-end collisions with three impact speeds. A rear-end collision at 16 km/h was considered as a low-speed impact, which represents collision events under urban congested road conditions, such as collision events in parking lots or in front of traffic lights. This impact speed is currently included in Euro-NCAP whiplash testing [23]. A rear-end collision at 40 km/h is classified as a medium-speed collision, according to FMVSS 301 [24]. The vehicle speed is used to simulate a moderate rear-end collision when the vehicle drives on a city’s main road. A speed of 56 km/h is specified in the frontal crash test during the UN R94 frontal-crash test [25], and it represents a high-speed rear-collision scenario in this study. Therefore, these three speeds cover a complete set of rear-end collision scenarios. The low-, medium-, and high-impact scenarios were modeled through three sled tests by applying corresponding acceleration pulses on the vehicle floor along the x-direction, as shown in Figure 6. The acceleration pulse of 16 km/h was derived from the Euro-NCAP low-speed whiplash test [23]. The pulse of 40 km/h was referenced from the car-crash test in a NHTSA report following FMVSS 301 [26]. Albert et al. [27] reported an acceleration pulse for a 56 km/h collision speed, and this load represented an extreme rear-impact scenario. In this study, six simulation cases were performed with the combination of three different rear-impact speeds (16 km/h, 40 km/h, and 56 km/h) and two restraint systems (three point and four point). The responses of seatback rotation and head movements were recorded via the specification of nodal history curves for corresponding nodes in the finite element model assembly. Meanwhile, a cross-section was created between the first (C1) and second (C2) cervical vertebra to obtain the neck shear forces and moments of the THUMS, which were utilized to calculate the neck injury.

Head and neck injuries are critical injury types in different collision scenarios, such as motorcyclist collisions with safety barriers and vehicle rear impacts. For instance, with respect to motorcyclist collisions with safety barriers, the head injury metric (HIC30) and neck injury metrics (F_x, F_y, F_z, Moc_x, Moc_y, and M_z) were utilized to evaluate the protective effects of roadside safety equipment for motorcyclists [28]. Similarly, head and neck injuries of occupants are also the dominant injury types in vehicle rear impacts. The injury indices considered in this study include the Brain Injury Criteria (BrIC) [29], the Cumulative Strain Damage Measure (CSDM) [30], neck injury metrics (N_ij and N_km) [31,32], and the Neck Injury Criterion (NIC) [33]. BrIC is calculated based on the angular velocity of the head’s center of gravity. The CSDM is calculated based on strain distribution in brain tissue. For instance, the CSDM (0.25) has been widely utilized in the auto industry, and it reflects the volume percentage of brain tissue, where the maximum principal strain is beyond 0.25. A higher value of the CSDM (0.25) reflects the higher possibility of experiencing diffuse axonal injury. The BrIC and CSDM (0.25) are not yet standardized and are only mentioned in regulations as potential assessment tools [34,35]. Th neck injury metric (N_ij) is derived from the axial force and bending moment at the upper neck and is used to assess neck injuries in frontal impacts. In contrast, the neck injury metric (N_km) is calculated based on the shear force and bending moment at the upper neck, and it better reflects whiplash injuries during rear-end collisions. The Neck Injury Criterion (NIC) is calculated based on the relative acceleration difference between the head’s center of gravity and the first thoracic vertebra (T1) and is associated with the S-shaped deformation of the neck. N_ij, N_km, and NIC have been incorporated into standards. For instance, N_km and NIC are included in the Euro-NCAP whiplash test regulations [23], and N_ij is included in FMVSS 208 [36].

In addition to the above injury indices, the neck ligament injury is a type of complicated and important injury. This type of neck injury can be calibrated via the THUMS l but not via the anthropomorphic test device (ATD) model. The THUMS has a complete neck ligament structure, and the stress and strain statuses can be evaluated at the tissue level, which intuitively reflects the relationship between head and neck injuries and the flexion and stretching of the neck. During our study, the maximum principal strains of the anterior longitudinal ligament (ALL), posterior longitudinal ligament (PLL), capsular ligament (CL), ligamenta flava (LF), and interspinous ligaments (ISLs) were also recorded in the THUMS. After the simulation, the neck ligament injury was evaluated according to corresponding injury thresholds measured by Yoganandan et al. [37] via quasi-static tensile tests (10 mm/s) of cadaveric cervical ligaments. Although the study did not consider the effect of the loading rate, which may be important for impact scenarios, the study utilized a large sample size, up to 25 human cadavers for the in situ biomechanical properties, and systematically characterized the ligament length and cross-section in terms of a heavy-duty cryomicrotome device. Those threshold values have been widely utilized to evaluate ligament injury in the safety community. For instance, Jin et al. [38] cited these thresholds when the THUMS was used to study the protective effect of active rotating seats on occupants. Wu et al. [13] cited these thresholds when the THUMS was used to study occupant injuries in different postures during rear-end collisions. Meanwhile, the mechanical properties of human neck ligaments obtained in the study of Yoganandan et al. [37] have been widely used in the development of finite element models of occupant heads and necks. Correia et al. [39] referenced the ligament mechanical properties from the study of Yoganandan et al. [37] during the development of muscle activation of the head and neck in the Global Human Body Models Consortium (GHBMC) human body models.

3. Results

3.1. Kinematic Responses of Occupants in Zero-Gravity Seat with Three-Point Seat Belts at Different Impact Speeds

Figure 7 presents the time-history curves of the seatback rotation angle for different rear-impact speeds. In general, the seatback first rotates to a peak value and then rebounds back. As the impact velocity increases, the rotation angle also increases. The maximal rotation angles are 5.1°, 13.1°, and 18.6° at 16 km/h, 40 km/h, and 56 km/h impact speeds, respectively.

The kinematic responses of the reclined occupant restrained in the zero-gravity seat are presented in Figure 8, Figure 9 and Figure 10 for low-, moderate-, and high-speed rear impacts in terms of Hyperview. The simulation results demonstrated significantly different biomechanical responses for different impact speeds. As shown in Figure 8, when the collision speed was 16 km/h, the seatback slightly rotated backward, and the maximal value was 5.1°. The pelvis moved forward due to the rebound of the seatback foam. However, its movement relative to the seat was limited. During the initial stage of collision, the contact force between the head and the headrest reached the first peak value, 1366 N, at 25 ms due to the pretension of the seat belts. Then, the seatback continued to rotate backward due to the push of the torso. The head slightly detached from the headrest at 33 ms then fell back on the headrest at 58 ms. The contact force between the head and headrest reached a peak value of 1125 N at 88 ms. The head completely detached from the headrest at 131 ms, and the neck started presenting a forward flexion state.

As shown in Figure 9, the kinematic responses of the occupant under 40 km/h impact were similar to the responses under the low-speed impact during the 0–100 ms period. However, with respect to the medium-speed impact, the backseat continued to rotate backward until 13.1°, and this resulted in a decrease in the restraining effect of the backseat on the pelvis. At 105 ms, the pelvis showed a strong tendency to move upward along the seatback. Meanwhile, the lower limbs began to significantly flip around the pelvis after 100 ms. In contrast to the low-speed impact case, the kinematics of the lower limbs maintained a similar posture during the whole collision period. With respect to the head contact force, the first peak forces induced by the seat belt pretension were the same for both impact speeds. However, the second contact peak force caused by the relative movement between the head and headrest was higher, reaching 1912 N, and the contact duration time was also longer. In addition, during the later stage of the collision, the neck had a similar forward flexion state as the low-speed impact scenario.

As shown in Figure 10, when the collision speed increased to 56 km/h, the maximal seatback rotation angle increased to 18.6°, and the restraining effect of the backseat continued to decrease. The buttocks completely separated from the seat cushion at 55 ms. Meanwhile, the lower limbs flipped more. In contrast with the low- and medium-speed impact, the kinematic responses of the head and neck were significantly different. At 70 ms, the head center of gravity passed the headrest and continued moving backward until 103 ms, and the neck tension force reached a maximal value of 1582 N. After that, the head continued flipping backward and severely collided with the upper edge of the headrest at 100 ms. And the contact force between head and headrest increased up to 7645 N, which was about four times the maximal contact force in the medium-speed impact. After the collision, the head flexed forward under the rebound of the headrest, accompanied by a certain degree of lateral flexion and rotation due to the unbalanced restrain forces of the three-point seat belt.

3.2. Injury Response of Occupants in Zero-Gravity Seats with 3-Point Seat Belt

Table 3. Summary of head and neck injury metrics for the zero-gravity seat.

Speed (km/h)	BrIC	CSDM (0.25) (vol%)	Nij	Nkm	NIC (m2/s2)
16	0.14	0.6	0.16	0.14	6.48
40	0.46	21	0.23	0.57	14.9
56	1.16	58	0.49	3.08	72.7

illustrates the head and neck injury metrics of occupants under different impact speeds. In general, the head and neck injury metrics were below their thresholds for both low- and medium-impact speeds. For instance, when the impact speed was 40 km/h, the BrIC, N_ij, and N_km were 0.46, 0.23, and 0.57, respectively, and all of them were smaller than 1. The value of CSDM (0.25) was 21%, which corresponds to the 7% probability of diffuse axonal injury occurrence. The NIC metric was 14.9 m²/s², which was close to the threshold of 15 m²/s² specified by the Euro-NCAP. However, when the impact speed increased to 56 km/h, the values of BrIC, N_ij, and N_km increased to 1.16, 0.49, and 3.08, respectively. The values of BrIC and N_km surpassed their critical thresholds. The value of CSDM (0.25) increased to 58%, and the probability of diffuse axonal injury was greater than 50%. It is worth mentioning that neck injury mechanisms evolved as the impact speed increased. When the impact speed was 16 km/h, the value of N_ij, governed by axial force and bending moment at the upper neck, and the value of N_km, governed by shear force and bending moment at the upper neck, were close. When the impact speed increased to 40 km/h, N_km became the dominant injury metric because of the head–neck forward flexion caused by the head rebound between 160 and 180 ms, as shown in Figure 9. When the impact speed became 56 km/h, as illustrated in Figure 10, N_km surged to 3.08 due to the collision between the neck and the edge of the headrest.

Table 4. Summary of cervical ligament injury in rear impact for the zero-gravity seat.

Speed (km/h)	ALL	PLL	CL	LF	ISL
Thresholds [37]	0.35	0.34	1.48	0.88	0.68
16	0.12	0.17	0.29	0.12	0.52
40	0.25	0.29	0.51	0.28	0.53
56	0.69	1.59	2.64	0.48	1.46

illustrates the biomechanical evaluation of the cervical ligaments under low-, medium-, and high-impact speeds. With respect to the low- and medium-impact speeds, the headrest was able to provide good support to the head, and the maximum principle strains in five types of cervical ligaments were below the threshold values [37], which were 0.35, 0.34, 1.48, 0.88, and 0.68 for the anterior longitudinal ligament (ALL), posterior longitudinal ligament (PLL), capsular ligament (CL), ligamenta flava (LF), and interspinous ligaments (ISLs), respectively [37]. When the impact speed increased to 56 km/h, all ligament strains were above their thresholds, except for the LF. As illustrated in Figure 10, the large strains were induced by the collision between the head and neck and the headrest at 100 ms. The above comprehensive biomechanical analysis demonstrated that the conventional three-point seat belt system could not effectively protect the occupant under high-speed impact. Both the occupant’s head and neck experienced severe injuries, evidenced by the high values of the head and neck injury metrics.

3.3. Occupant Responses in Zero-Gravity Seat with Four-Point Seat Belt at High-Speed Rear Impact

Figure 11 shows the kinematic responses of the THUMS restrained in the zero-gravity seat by the three-point seat belt system and four-point seat belt system for the 56 km/h high-speed rear impact. In terms of Hyperview, the occupant’s kinematic images were extracted at four different moments, and then those images were merged with varying transparencies. Compared with the three-point seat belt system, the four-point seat belt system provided stronger and more balanced restraining effects for both shoulders of the occupant. In terms of Hypergraph, Figure 12 compares the displacement trajectory of the occupant’s head relative to the chest when the occupant was restrained via the three-point seat belt system and the four-point seat belt system. When the occupant was restrained by the three-point seat belt system, the maximum relative displacements were 116 mm and 110 mm along the x- and y-direction, respectively. In contrast, when the occupant was restrained by the four-point seat belt system, the maximum relative displacements were 85 mm and 16 mm along the x- and y-direction, respectively, which were 26% and 84% lower than that of the occupant restrained by the three-point seat belt system. The four-point seat belt system had an additional shoulder strap, and accordingly, more balanced forces were applied to constrain the occupant’s torso through the double shoulder straps. Therefore, the occupant’s head and neck were stabilized along both the x- and y-direction.

Meanwhile, the head and neck injury metrics are summarized in

Table 5. Comparison of head and neck injury metrics for zero-gravity seats with different seat belt systems.

Belt Type	BrIC	CSDM (0.25) (vol%)	Nij	Nkm	NIC (m2/s2)
3-Point	1.16	58	0.49	3.08	72.7
4-Point	0.39	17	0.38	1.46	44.1

, and the cervical ligament injury metrics are summarized in

Table 6. Comparison of cervical ligament injury metrics in rear impact for zero-gravity seats with different seat belt systems.

Belt Type	ALL	PLL	CL	LF	ISL
Thresholds [37]	0.35	0.34	1.48	0.88	0.68
3-Point	0.69	1.59	2.64	0.48	1.46
4-Point	0.43	0.76	1.12	0.37	0.79

. Compared with the three-point seat belt system, the head and neck injury and the cervical ligament injury were significantly reduced. Specifically, the head injury indices BrIC and CSDM (0.25) were reduced by 66% and 70%, respectively, and the neck injury indices N_ij, N_km, and NIC were reduced by 22%, 52%, and 39%, respectively. The ALL, PLL, CL, LF, and ISL were reduced by 37%, 52%, 57%, 22%, and 46%, respectively. It is worth mentioning that regarding the head injury indices and cervical ligament injury metrics, the CL and LF were below the injury thresholds under the protection of the four-point seat belt system. However, regarding the neck injury metrics N_km and NIC and the ligament injury metrics, the ALL, PLL, and ISLs were still beyond the injury thresholds, and other protection strategies still need to be developed to further reduce cervical spine and ligament injuries.

The neck injury indices N_ij and N_km are directly affected by the cervical axial force F_z and cervical shear forces F_x of the upper neck, respectively. Figure 13a presents the force components that are associated with the corresponding local coordinate system, and the corresponding upper-body structure was prepared in Hypermesh by the variation in the transparency of different components of the human body model. F_z with a positive sign represents compression, and F_z with a negative sign represents tension, as shown in Figure 13b. When the occupant was restrained by the three-point seat belt system, the head and neck collided with the headrest at 110 ms. The tensile cervical axial force F_z reached the maximal value, 2170 N, and the corresponding N_ij was 0.49. The maximal shear forces reached 2575 N, causing a significantly high value of N_km, 3.08. In contrast, when the occupant was restrained by the four-point seat belt system, those loadings were reduced significantly because the head’s out-of-position posture was avoided, and the following collision between the neck and the headrest was eliminated. The maximal cervical axial force F_z was reduced by 27%, and the corresponding N_ij was reduced by 22%. The maximal shear force, F_x, was reduced by 56%, and the corresponding N_km was reduced by 52%.

4. Discussion

The occupant’s head and neck experienced severe injury in the zero-gravity seat under high-speed rear impact with the three-point seat belt system. In terms of the four-point seat belt system, the head injury metrics can be reduced below the threshold, and the neck and its ligament injury metrics can also be significantly reduced. However, the neck injury metric N_km and the cervical ligament injury metrics ALL, PLL, and ISLs were still beyond their thresholds. In the future, further protection strategies or impact energy absorption devices need to be introduced to protect the occupant’s neck. For instance, one way is to enhance the seat belt protection performance by optimizing the seat belt webbing, retractor, and the force limiter. Meanwhile, active headrests can actively adjust the height and angle of headrests in the event of a collision to prevent the occupant’s head and neck from overextension. In addition, collapsible seat rails can extend the collision time and reduce the peak tension of the seat belt through preset controllable collapse. In addition, we choose the 50th percentile male model in this study since the model represents one of the largest population groups. In our future study, we plan to consider the effects of the occupant’s body size and different configurations. The family of THUMSs has a variety of human sizes, from 5th-percentile adult female (AF05) to 95th-percentile adult male (AM95). Further, in terms of the posture adjustment technique, different occupant postures can be realized.

The results of this study further expanded the research results of Hasija et al. [11] and Anh Vu Ngo et al. [12], filling the gap in the research of zero-gravity posture occupants in rear-end collisions. When the collision speed was 56 km/h and the occupant was restrained by a three-point seat belt, the occupant in the current zero-gravity seat and the occupant in the reclined seat [11,12] showed similar head and neck movement postures, and their neck collided with the headrest due to excessive extension. However, in the current study, the collision between the neck and headrest could be avoided when the four-point seat belt system was utilized to restrain the occupant. Further, the head injury metrics BrIC and CSDM (0.25) reported by Anh Vu Ngo et al. [12] were 1.56 and 62%, respectively. And they were reduced to 0.39 and 17% when the occupant was restrained via the four-point seat belt system in the current zero-gravity seat.

5. Conclusions

This study comprehensively evaluates the occupant kinematics and injury responses in a zero-gravity seat under low-, medium-, and high-impact speeds with different seat belt systems. The THUMS-based finite element model assembly of occupant restraint systems consisted of a conventional three-point seat belt and a seat model in the regular position. The simulation model was verified through the PMHS test, and good agreements were obtained for both the kinematic postures at different collision times and time-history curves of head acceleration, head rotation, and T1 acceleration. As the impact speed increased, the occupant in the zero-gravity seat with the three-point seat belt showed different kinematic and injury responses. When the impact speed was at a low or medium level, the occupant could be restrained by the seatback and the three-point seat belt system. The head injury metrics, including the BrIC and CSDM, neck injury metrics, including the cervical spine injury metric, N_ij, N_km, and NIC, and the cervical ligament injury metrics ALL, PLL, CL, LF, and ISLs were all below their thresholds. As the speed increased to 56 km/h, the maximal seatback rotation angle increased to 13.1°, and the occupant’s head moved out of the headrest and collided with the headrest at 100 ms. The collision led to a surge in the contact force between the head and the headrest and of all the head injury metrics and neck injuries, most of which surpassed their corresponding thresholds. In terms of the four-point seat belt, the out-of-position posture of the occupant’s head can be avoided, as well as the following collision between the neck and the headrest. The head injury indices BrIC and CSDM (0.25) were reduced by 66% and 70%, respectively, and the neck injury indices N_ij, N_km, and NIC were reduced by 22%, 52%, and 39% respectively. The cervical ligament injury metrics ALL, PLL, CL, LF, and ISL were reduced by 37%, 52%, 57%, 22%, and 46%, respectively.