Occupant safety effectiveness of proactive safety seat in autonomous emergency braking

The results of the present study showed that PSS is effective as an active safety system for ensuring occupant safety by counteracting OOSP before AEB operation. In near future, occupants are likely to adopt various postures for comfort before partially or fully autonomous driving, but there have been insufficient data regarding the safety of occupants in such postures. Several experimental and/or finite element model studies were performed to examine the responses of the occupant in unintended seat positions14,25,26. These studies verified the jeopardy of undesigned seat positions, but they did not propose means of counteracting these perils for occupant safety. In the development of an active safety system, it is important to ensure occupant safety before braking and/or collision in haphazard seat positions. In addition, information regarding the characteristics of human responses during the operation of the safety system will allow safety improvement by controlling the interactions of active and passive safety systems. This study confirmed the effectiveness of PSS operation to improve the interactions of active/passive safety systems and determined the characteristics of human responses during the operation of the safety system. Therefore, this study provides valuable insights to ameliorate risk to the occupant in accidental seat positions before braking and/or collision.

In the two-stage braking maneuver for AEB operation, to operate a maximum braking acceleration of 1 g under 50 km/h, AEB decelerates the vehicle’s speed gradually from TTC 1.6 s11,27. Previous studies reported that AEB at TTC 1.6–0.6 s starts a deceleration of 0.2–0.6 g; this phase is defined as the partial braking phase28. Referring to the above AEB operating mechanism, the tests in this study performed the partial braking phase exhibiting acceleration to 0.4 g for 1,000 s to assess the effects of PSS operation. Whereas we assumed that continued adjustment of the seat position during AEB operation would lead to unpredictable conditions and increase endangerment to the occupant because of incomplete PSS operation. For this reason, PSS operation was performed in TTC 2.6 to 1.6 s, which is before the partial braking phase of AEB.

However, TTC is sole a theoretical concept and it can differ from real-world conditions; thus, there are enough likelihoods that occurrence of a collision before the maximum braking operation of AEB. To alter from the fully reclined seat to NSP, PSS required at least double the 1 s that of TTC 2.6 to 1.6 s. Because of the designed time algorithm for completing PSS operation within 1 s and verification of the test environment only in the partial braking phase, the present test results may have limits for fully considering PSS that any OOSP be adjusted to NSP. Withal, to our knowledge, this study represents the first attempt to determine the effectiveness of the PSS for counteracting risk to the occupant in OOSP. Thus, the limits of the PSS operation process in the tests could be regarded to consideration of that the occupant’s safety against unpredictable perils during incomplete PSS operation. Furthermore, the characteristics of occupant responses in the partial braking phase could be utilized to gain fundamental insights to facilitate the improvement of active safety systems, such as PSS strategies. From this viewpoint, the experimental parameters for the PSS were sufficiently covered, and our results can be considered both valid and reliable.

The understanding and utilization of the PSS operating mechanism can be beneficial to ensure occupant safety, adjusting both the seat and occupant’s posture to the intended design position before the time of a collision. The protective operation of restraint systems is greatest when occupants and their ambient environment are in the intended design position (i.e., NSP) at the start of the collision11,12,13,14. However, real-world accident records indicate that, in many instances, both the seats and the occupants are not in the designed position at the time of a collision. Therefore, the functional mechanism of the PSS to adjust the seat position of the occupant to NSP before a collision is important to counterbalance the jeopardy by OOSP. The results of the present study showed that, regardless of whether the seat pan shifted forward or rearward, the occupant’s frontal motion with PSS operation was generally decreased and the motion characteristics showed a small standard error range, compared with PSS_Off conditions. In particular, PSS_On27R and PSS_On27F showed a decrease of the characteristic motions related to submarining (e.g., upper body hovering and falling, as well as pelvic slipping)14. These results indicated that the PSS had a protective effect when the occupant’s seat position was adjusted closer to NSP. However, the head excursion and neck rotation were 0.8 and 0.7-fold greater in PSS_On27F than in PSS_On27R, respectively. To confirm whether there were differences in PSS effectiveness because of postural adjustment between PSS_On27R and PSS_On27F, we compared the occupant’s postures at 1200 ms, which PSS operation was completed. As shown in Fig. 5, the changes in torso rotation angle by PSS operation in PSS_On27R and PSS_On27F were from 63.1° ± 0.4° to 68.0° ± 0.3° and from 63.7° ± 0.7° to 68.5° ± 0.6°, respectively; these were not markedly different. In contrast, the changes in head rotation angle in PSS_On27R and PSS_On27F were from 44.3° ± 1.8° to 40.0° ± 1.7° and from 41.0° ± 1.8° to 36.0° ± 1.5°, respectively; these represented a difference of approximately 5° between the two conditions. These results mean that PSS_On27R and PSS_On27F had different neck rotation angles in each initial posture depending on altered head rotation angle and head center of gravity. Occupants in PSS_On27R tended to lean more heavily on the headrest, which may have been because of the difference in center of gravity caused by the lower body posture and seat pan location. Furthermore, these differences led to pelvic-on-femoral osteokinematics. When the seat pan location moves rearward with the supralumbar trunk stationary on the seat back, the anterior pelvic tilt (flexion) will occur, and coincidentally the lumbar spine would likely have a larger curvature toward the forward. The human body motion is the summation that consists of the external force and internal force. Previous studies had mainly used the mass-spring-damper systems in representing the mechanical impedance of human body motion29,30.

$$ me(t)ddot{x}(t) + be(t)dot{x}(t) + ke(t)x(t) = fe(t) $$


Equation (1) is the mass-spring-damper model representing the joint torque. This is the second-order dynamic equation where me(t), be(t) and ke(t) the impedance parameters, which denote the mass, damping factor, and stiffness of joints, respectively; and fe(t) presenting the force exerted to joints. It means the human body motion depends on the joint stiffness and damping characteristics. Furthermore, the muscular structure and properties would likely significantly affect the degree of joint stiffness. However, the joint and muscle characteristics highly rely on their particular directions and states (e.g., muscle state of contraction/relaxation, joint state of flexion/extension, and so on). Hence, different occupant’s initial postures by seat pan moving directions of PSS would likely be the main factor that generates disparate motion characteristics during AEB. Additionally, to identify the regularity of muscular interaction, we compared the amounts of muscle activation for each PSS strategy. Although the motion and muscle characteristics can be different due to using what PSS strategy to reach the NSP before AEB operation, the test results indicate that most head and neck excursions decreased when applied PSS. Therefore, it may consider that PSS strategies serve meaningful effects to restrict more safely the occupant motion.

Figure 5
figure 5

Frontal motion trajectories of an occupant in the sagittal (x–z) plane. Photographs were taken by Musculoskeletal Bioengineering Laboratory, Sejong University.

The maximum belt tensioning force on the shoulder and lap belt was not significantly different under each test condition (P > 0.05). On the other hand, the maximum muscular activities indicated significant differences diversely under each test condition (P < 0.05). These differences show that the possibility of muscle restraining activity in SCM was ameliorated, compared with the muscle activation in PSS_Off27. Furthermore, it means that characteristics of muscle activation-particularly longissimus muscles-would likely have a difference regarding each posture of an occupant. The muscle activities measured in all seat positions peaked at approximately 2400 ms after activation of partial braking, at approximately 100–200 ms after the excursions had reached their maximum values. This point would support the phenomena of damping force occurrences with respect to changing joint stiffness and muscle contraction. However, all muscle activities of the occupants were below 50% of the MVC values. These results may have been limited by the low-speed conditions-the partial braking phase of AEB-used in the tests. Hence, further research is needed to quantify the actual risk of muscle injury. In this study, although the muscle activities during PSS operation were not expected to reflect a risk of muscle injury, these results are valuable for improving the overall understanding of occupant motions during AEB with PSS.

Several factors that can directly or indirectly influence the responses of occupants during the tests (e.g., habituation, lack of awareness, and startle response) have been studied14,20,31,32. Blouin et al.32 noted the effects of habituation and acoustic startle stimulus of the test platform on resulting occupant responses. They suggested that it is important to perform tests without habituation and acoustic startle stimulus to clearly observe occupant responses. Beeman et al.31 studied the effects of contracted muscles on occupant kinematics and kinetics; they showed that muscle condition before the test had a significant effect on the occupant motion response. These studies suggested that effective exploration of experimental parameters would facilitate better characterization of occupant motion responses14,20. Therefore, although the objective of the present study was to examine the test parameters of the PSS that influence the results, comprehensive observation was not possible because of limitations regarding the test environment and conditions. The selection of PSS strategy was restrictive because of considerations regarding test volunteer safety. However, because the occupant motion characteristics for the OOSP measured here were similar to the characteristics reported in previous studies, the experimental parameters may have been sufficient and our results can be considered both reliable and valid14. The sample size and specific type of tested subjects selected in this study might be limitations to precisely attest the PSS effectiveness. However, the restrictive sample size was the situational limitation of this study, and the particular type of volunteer may be more beneficial to define the criterion of human models. If we verified with volunteers of more people and various types in the future study, we would get better results to identify the effects of PSS.

Furthermore, the results of this study are highly meaningful because we identified the effectiveness of PSS implementation during AEB operation. Importantly, the acquisition of human characteristics for PSS effectiveness in this study was novel and the test parameters were covered sufficiently; thus, our results can be considered reliable and valid. These results represent essential information regarding the improvement of safety device effectiveness to ameliorate occupant restraint and counteract any deterioration of occupant safety. Further studies are required to devise a PSS mechanism suitable for a more diverse range of emergency maneuvers, such as evasive swerving, to further improve occupant safety. This study provides a basis for obtaining a clear understanding of occupant motion responses for active PSS and for the identification of potential discomfort or injuries to the occupant during AEB operation.

Overall, the results of the present study indicated that the PSS can improve occupant safety in OOSP during AEB operation. The occupant’s frontal motion was generally decreased with PSS operation; the characteristics of frontal excursion showed a small range of standard error, compared with the PSS_Off condition. In the fully reclined seat position, PSS operation led to a reduction of the characteristic motion associated with submarining of the occupant.

In conclusion, PSS operation can improve the safety of occupants in OOSP. However, the occupant safety effects according to various PSS operation strategies must be further verified. To secure occupant safety with various PSS operation strategies, additional confirmation is required for more diverse emergency maneuvers, such as evasive swerving and/or the use of recent occupant restraint configurations (e.g., belt-in-seat components). Moreover, if the PSS control system can perform with better classification of the occupant’s condition (e.g., body type, awareness/lack of awareness, and upper body center of gravity) in autonomous vehicle driving, the optimized algorithm for PSS strategies may provide greater occupant safety improvements.