The effect of seatback deformation on out-of-position front-seat occupants in severe rear impacts

Abstract Objective This study assesses the effects of seat deflection in severe oblique rear impacts with laterally out-of-position ATDs where the head is not supported by the head restraint. Method Six high-speed rear sled tests were conducted at 48 km/h with a 195 degree PDOF. A lap-shoulder belted 50th percentile Hybrid III ATD was leaned inboard and seated in six different front passenger seats (A-F); five of the seats were selected from mid-sized sedans and one was a non-production rigidified Seat Integrated Restraint (SIR) seat. FRED-III pull tests resulted in seat stiffnesses that varied from 73 to 172 N/mm. Seat F had the greatest stiffness. The seat and ATD responses were assessed. The biomechanical responses were evaluated and compared to relevant IARVs. Results In all tests the ATD moved rearward and twisted the seat. There was limited differential motion of the torso relative to the seatback. The ATD position and PDOF prevented head restraint engagement allowing head and neck extension over the seatback. The seatback angle was measured on the inboard side. At maximum yield, it was greatest with Seat E, followed by Seat A and Seat D, at 71, 67 and 62 degrees, respectively. The duration of rearward deformation was also greatest with Seat A, Seat D and Seat E providing longer ride-down. The head, chest and upper neck responses were below IARVs. Lower-neck extension moments were above injury threshold with Seat B, C and F. Seat F had the highest lower-neck moment. Conclusion Seats with greater deformation provided the greatest ride-down durations and the lowest overall biomechanical responses. The combination of high impact severity and lack of head support resulted in high lower-neck responses, highlighting the potential benefit of energy management from deforming seat structures.


Introduction
The topic of seat strength has been discussed over the last five decades. Seat strength has gradually increased with time (Viano et al. 2009;Viano and White 2016), due to design changes including seatback and head restraint height requirements to better support the head and upper torso. Most modern seats have a dual recliner design.
In 2004, NHTSA concluded that improving seating systems performance is more complex than simply increasing the strength of the seatback (NHTSA 2004). Various factors can influence occupant injury risk in rear impacts including crash severity, impact direction, vehicle type, initial posture, occupant age, sex, and medical history. NHTSA was more recently petitioned to increase rearward seatback loading requirements (ARCCA, 2015). Research has shown a concern for increasing seat strength and stiffness that may lead to increased injury risks, in particular for vulnerable occupants as report by (Viano et al. 2009), and when the head is unsupported in severe rear crashes report by (Viano et al. 2009). Seat Integrated Restraint (SIR) seats have the shoulder belt anchor mounted the seat structure itself, and not directly to the vehicle structure. These seats have been offered as potential design alternatives to conventional yielding seats with restraint systems that mount to vehicle structure. SIR seats tend to be stronger due to frontal impact loading requirements.
The association between being out-of-position (OOP) and the risk of extension loading of the spine has been investigated with rear impact tests (Benson et al. 1996;Viano et al. 2009;Viano et al. 2021;Parenteau et al. 2021). One of the most severe positions for neck extension loading consists of leaning inboard. This position can lead to the occupant head, neck, and upper torso to displace off the head restraint and upper seatback due to offset interaction (Viano 2011). This may lead to spinal fracture, dislocation and transection with disabling spinal cord injury (Viano et al. 2009). Rear oblique impacts may also result in offset loading and thus increase in spine loading (Croteau et al. 2022). Viano (2011) reported on rear impact field accidents where the front seat occupants sustained serious thoracic injury. In all cases, the delta V was between 40 and 56 km/h. The impacts were offset with a slight oblique PDOF. In some cases, the occupant head was displaced laterally. The seatback frame acted like a fulcrum as the unsupported head and upper body moved rearward relative to the vehicle interior causing extension of the spine.
The objective of this study was to assess the effect of seat deformation in severe oblique rear impacts when the head is unsupported.

Method
Six sled tests were conducted at the Ford Safety Center with the seats mounted on flatbed fixture, angled with a 195 degree PDOF (Principal Direction of Force), and subjected to a delta V of 48 km/h. Figure 1 shows the overall sled test setup. Appendix A (see online supplements) provides additional information on set-up.
Seats: Five seats were chosen from a peer group of modern mid-sized sedans with electric recliners and one modified Seat Integrated Restraint (SIR) seat. The SIR seat was modified to be a dual recliner systems from a single recliner system for symmetrical seat stiffness. Seat deflection was assessed in the sled tests and in pull-tests.
The seats were quasi-statically pull tested utilizing the FRED-III methodology to obtain force deflection and stiffness characteristics. FRED-III methodology is the same as FRED-II (Viano 2019). FRED-III updates the FRED-II in construction to aid in testing. Figure A1 (see online supplements) shows the overall FRED-III setup.
The seatback was positioned with the upper seatback at approximately 20 degrees rearward of vertical. The seats were set to their mid-track locations with the vertical lifts in a full down position consistent with the FMVSS 208. All seats were tested with the same restraint system except Seat F which used its integrated system.
Pulse: Figure A3 (see online supplements) shows an example of the 48 km/h pulse used in all 6 tests. It consists of a 15 g peak acceleration at about 90 ms. The pulse duration is approximately 165 ms.
ATD: All six tests were conducted with an instrumented 50th percentile Hybrid III ATD (anthropometric test device). The ATDs were lap-shoulder belted using a belt system and configuration representative of a mid-sized sedan. The torso was leaned inboard so the head-torso angle was about 20 degrees along its longitudinal axis to simulate a lateral lean where the head would not be supported by the head restraint.
The ATDs were instrumented with pelvis, chest, head triaxial accelerometers and angular rate sensors, and with upper and lower-neck load cells. The data was captured and processed according to the SAE J211 standard.
Analysis: The peak biomechanical responses of the Hybrid III were normalized to their percent of IARV (Injury Assessment Reference Value) by thresholds published by Mertz et al. (2016). The percent injury risk associated with each ATD biomechanical response was estimated utilizing Sigmoidal functions fitted AIS 3þ injury risk functions as described in Parenteau et al. (2021): The seatback and ATD kinematics responses were determined from video analysis using the Kinova TM software to track targets on the ATD and seat.

Seat characteristics
The FRED-III quasi-static seat pull tests resulted in stiffnesses of 73 to 172 N/mm in the pull distance region of 2.2-8.9 kN of load. The modified SIR seat, Seat F, had the highest stiffness and highest ultimate load. Seat C had the lowest stiffness value. Table A1 (see online supplements) shows the stiffness results, head restraint gap, model, and other details. Figure 2A (see online supplements) shows the force displacement plots for all six seats.

ATD kinematics in the sled tests
The ATD moved rearward relative to the sled buck and loaded the seat in the rear impact sled tests. The seatbacks rotated rearward to maximum yield and twisted inboard due to offset occupant loading. As the ATD continued to move rearward the pelvis rose relative to the seat cushion surface mainly due to seatback rotation and geometry. There was limited differential motion of the torso relative to the seatback. The ATD position and PDOF prevented head restraint engagement. The head and neck extended as the upper torso hinged along the inboard upper corner of the seatback, and the head continued to move rearward and down. The ATD then rebounded forward and loaded the seatbelts. Figure A4 (see online supplements) shows the occupant kinematics at 50, 150, and 250 ms. Figure 2 shows the ATD kinematics at maximum seatback deflection and corresponding time. The inboard angle was greatest with Seat E, followed by Seat A and Seat D, at 71, 67 and 62 degrees, respectively. The duration was also greatest with Seat A, Seat D and Seat E providing longer ride-down. Table 1 summarizes selected seatback responses including the inboard angle at maximum deflection. Table A2 (see online supplements) provides additional seatback angles from video analysis, including inboard and outboard angle pre and post-test, and at maximum deflection. The pretest seatback angle varied from 19 to 24 degrees on the inboard side. The inboard and outboard angle data was used to determine seatback twist and elastic and plastic deformation. Seatback twist at maximum deformation was greatest with Seat D, at 29 degrees. The post-test seatback angle on the inboard side was lowest with Seat F at 24 degrees. Plastic deformation was largest in Seat A, D and E; Seat D had the largest elastic deformation.

Occupant biomechanical responses
The peak biomechanical responses were normalized by available corresponding IARVs. Table A3 (see online supplements) and Figure A6 (see online supplements) show selected normalized responses. The head, chest and upper neck responses were below IARVs. Lower neck extension moments were above injury thresholds with Seat B, C and F. Table 1 also summarizes normalized chest 3 ms responses. The chest responses were well below IARVs and the corresponding injury risk was low, with less than 0.5%. The chest percent IARV was highest with Seat F, at 54%. Figure A8 (see online supplements) shows the normalized lower-neck extension, -My, as a function of plastic seatback deformation. A linear trendline was fitted through the results. The results show that lower-neck response decreased with seatback deformation (R 2 ¼ 0.8028). Figure A5 (see online supplements) shows the head trajectories obtained from video analysis. Table 2 summarizes the peak head displacement along the x and z-axis. Head displacement increased with an increase in seatback deformation. The head rearward displacement was lowest with Seat F and highest with Seat A. Seat E had the greatest downward displacement. Table 2 also summarizes normalized head resultant acceleration and HIC 15 responses and corresponding injury risk. The highest HIC 15 responses were for Seat C and F with an injury risk of 1.01% and 1.11%, respectively. Figure 3 shows the lower neck moment along the lateral axis. The peak extension moment occurred at about 132 ms  Figure 2. ATD kinematics at maximum seatback rearward deformation.
with Seat F and at about 210 ms with Seat A (Table A4, see online supplements). Table 2 also summarizes selected normalized lower neck responses -My and associated injury risk. Three of the extension responses were above IARVs (Seat B, C and F). The corresponding injury risk varied from 8.3% with Seat D to 46.8% with Seat F.

Discussion
Seats designs have changed over time. The risk of serious injury in rear crashes has decreased (Viano and Parenteau 2016) and continues to be lower than other crash modes ). There are suggestions for stiffer seat designs (ARCCA 2015). New injury patterns have emerged with modern seats (Viano 2011;Viano et al. 2019), in particular at high crash severity and offset and/or oblique loading. In this study, the combined effect of oblique loading and inboard lean was simulated with various modern seats. The PDOF was 195 degrees and the ATD was leaned inboard 20 degrees. The head of the ATD did not engage the head restraint in the tests. Some biomechanical responses exceeded the IARVs because of the lack of head and neck support in the seats that allowed for less plastic deformation. The injury risks were highest with the rigid seat (Seat F), in particular for lower neck extension. Seatback rearward deformation provided greater ride-down distance and yielding, lowering occupant loading and biomechanical responses. The results highlighted a potential benefit of seat deformation, in particular when the head and upper torso are unsupported. Parenteau et al. (2021) showed that more rigid seats have an increased overall risk of injury across the spectrum of low to severely high-speed rear impacts. Parenteau et al. (2022aParenteau et al. ( , 2022b discussed the sequential occupant loading of the pelvis, chest and head into the seat in rear impacts with a nominally seated ATD at 40 km/h. They reported that the peak head, chest and pelvis accelerations along the longitudinal axis, and upper and lower neck extension moments occurred prior to maximum dynamic seatback deflection. A similar analysis was conducted in this study. Using Seat A as an example, Figure 4 shows the resultant head, chest and pelvis acceleration as a function of time, and corresponding seatback rotation. Resultant accelerations were assessed due to the multi-directional loading. The peaks were 22 g for head, 19 g for chest and 32 g for pelvis, and occurred at 151 ms, 119 ms and 113 ms, respectively. Maximum change in seatback rotation (elastic deformation) was about 48 deg at about 228 ms. These results indicate that peak occupant acceleration occurred prior to maximum yield. Table A4 (see online supplements) provides additional information on timing for all six seats. Figure A7 (see online supplements) shows the upper and lower neck flexion/extension moment as a function of time. Due to the lack of head restraint engagement, the upper and lower neck extension moments occurred after maximum seatback deflection and coincided with the maximum head vertical displacement.
Head restraint interaction as the seatback yields rearward is an important factor in controlling kinematics. Lack of head support can lead to extension loading in the spine. Viano et al. (2009Viano et al. ( , 2021 conducted a series of rear sled tests with a 95th percentile ATD initially leaned inboard. The authors discussed how the pelvis was supported on the seatback while the head was not, causing the upper body to wrap around the inboard side of the seatback frame. Viano (2011) reported on field accidents where the front seat occupant was injured as the head and shoulders displaced inboard off the seatback. The thoracic spine extended around the upright seat due to the offset loading and experienced a fracture-dislocation. There are situations when the head can become unsupported in a rear impact including an oblique impact direction (Croteau et al. 2022), seatback support limiting rotation (Parenteau et al. 2020), and/or when the occupant is leaned inboard or outboard (Viano et al. 2009).
An increase in ride-down distance and duration are key factors for energy management. Energy is absorbed with seat   deformation or yield in rear crashes. Limiting yield and/or seatback rotation can lead to increased forces imparted onto an occupant. Occupant biomechanical responses have been shown to increase with seat strength, irrespective of crash severity Parenteau et al. 2021). In this study, the seat stiffness of the six different seats was determined in in-line quasi-static seat pull tests. The results indicated that seat stiffness did not totally predict seat and occupant responses, thought the stiffest seat had the highest overall ATD responses. The test conditions used in this study are considered severe; the combination of crash severity, impact direction, and initial occupant position adds to the complexity and severity. A 48þ km/h delta V represents < 1% of all tow away crashes (Viano and Parenteau 2022). The ATD was initially leaned inboard. This position is considered OOP. Non-nominal postures can occur, particularly for the rightfront passenger (Bohman et al. 2020;Reed et al. 2020). Reed et al. (2020) documented the frequency of right-front passenger postures and activities in a naturalistic behavior study. The authors observed that the head was often rotated or tilted downward It is predicted that occurrences of out of position occupants will increase as advanced vehicles relieve the driver of driving tasks.
The results of this study add to the body of literature investigating the effect of seatback deflection on occupant responses. While limiting seatback deflection can reduce front occupant rearward displacement, it can increase occupant accelerations and corresponding forces. The reduction in accelerations and forces can be particularly important in situations where the occupant is not nominally positioned and may not benefit from full engagement of the head restraint. Future research in quantifying occupant out of position exposure both in frequency and type of position would help seat designers in defining appropriate seat testing and refining seat design for the future. These tests were conducted with a 50th Hybrid III ATD which has been utilized in high-speed rear impacts for many decades. Responses of new ATD designs in high-speed rear impacts is a topic of interest in present and future research. This study is limited to the six studied seats, a large sample group may give further insights to trends of seat characteristics and occupant responses in high-speed OOP rear impacts.

Funding
The author(s) reported there is no funding associated with the work featured in this article.