Autonomous Vehicle Occupant Safety

Experts have been working to develop crash test dummies designed to assess vehicle occupant safety in reclined seating positions.

Autonomous vehicle (AV) technology has advanced in quantum leaps in recent years. However, despite all the miles logged by OEM and AV software companies as they put the technology through its paces, accidents occur and will continue to occur. This is especially true because future AVs will share the road with conventional cars for decades to come. There is much to do before the autonomous car is capable of driving more safely than the ‘average’ human driver and is able to protect the occupants in a crash. A fully autonomous vehicle will remove the duties of driving from human hands. This offers new opportunities for OEMs and Tier 1 suppliers to develop novel seats with the ultimate aim of improving ride comfort and social interaction among occupants. AV concept cars have explored entirely different seating configurations, such as lounge-style setups with rear-facing front rows, diagonal cross-seating, and increased seatback recline angles for a more relaxing ride. These innovative seating positions pose unique challenges when it comes to occupant protection.

_{A reclined seating position, which increases the risk of submarining, is expected to be more common in autonomous vehicles}

Anthropomorphic test devices (ATDs) have been developed and continually evolved over the past 65 years, designed to assess the efficiency of vehicle restraint systems. These ATDs are designed to replicate standard occupant seating postures with a seatback angle of 25° and seat pan angle of approximately 14.5°. Reclined seating postures are not covered in current safety regulations, nor have they been considered in dummy design thus far. No existing ATDs can be configured to accurately replicate an occupant in a reclined seat. 

Today’s occupant restraint system technology was developed for forward-facing seats and standard upright postures only. Other scenarios, such as a reclined seating position, oblique-oriented seats and rearward-facing seats have not been considered or evaluated in high-speed impacts. The restraint systems for these seating configurations must be thoroughly re-examined to assess their safety protection in such scenarios.

Occupant submarining can cause severe pelvic and abdominal injuries such as fractures and internal organ laceration. An occupant is far more likely to submarine when in a reclined position, greatly increasing their risk of these injuries. Improving a dummy’s submarining-related biofidelity is therefore an immediate concern in the development of autonomous vehicle restraint systems.

Humanetics has launched an initiative to modify an existing ATD design for AV crash testing. The latest 50th percentile THOR ATD (THOR-50M) was redesigned so that it could be easily positioned in reclined postures. Refinements were also made to improve neck torsion biofidelity, upper torso kinematics and submarining behavior in these complex seating scenarios. The resulting model is the THOR-AV 50M ATD.

The THOR-AV 50M has an all-new neck with updated curvature, which is more representative of a human than the straight neck design used in current ATDs. It also features an added torsion element to improve its biofidelity. The restructured upper thoracic spine provides the necessary adjustment for positioning in a reclined posture, and a more representative lumbar spine with increased bending ability was created, improving the upper torso kinematics as well as the submarining response. 

Anthropometric data from a recent study was leveraged in the design of the shape of the pelvis, and the geometry of the uncompressed pelvic flesh was generated with a technique used in a previous Humanetics project. In addition, a more representative coupling between the flesh and the bone was achieved through updated mechanical features. The newly developed abdomen, which is more durable, complies with the UMTRI (University of Michigan Transportation Research Institute) Anthropometry of Motor Vehicle Occupants and includes integrated pressure sensors to provide additional pressure measurements for predicting submarining and abdominal injuries. 

The finite element model of the THOR-AV 50M was developed concurrently with the physical ATD and was used in synergy to aid in the design of the hardware and model. In particular, the FE model was used to optimize the neck design according to the biofidelity requirements. Frontal Impact, lateral impact and torsion were all simulated in computer modeling to optimize combined biofidelity. FE models of the THOR-AV 50M, which can be configured for 25°, 45° and 60° seatback angles, are now available for users to analyze the restraint system in normal and reclined seating postures. The model will be further validated and released with the biofidelity data in conjunction with the ATD’s hardware development in the near future.

After an 18-month development program, carried out by a dedicated engineering team in collaboration with key industry stakeholders, the first THOR-AV 50M prototype was assembled in mid-May 2019 and has since been going through a series of biofidelity evaluation tests. The preliminary neck biofidelity test has a combined score of less than 2.0, which corresponds to ‘Good’ according to NHTSA’s Biofidelity Ranking (BioRank) system in which the lower the score, the better the biofidelity. The abdomen test for steering wheel and rigid bar impacts also achieve a BioRank score that corresponds to ‘Good’. The biofidelity tests with the available corridors are expected to be completed in July this year. In the meantime, NHTSA launched a biomechanics research program for AV occupant safety in the autumn of 2018. The focus of the program is to study occupant response in forward-facing, rearward-facing and reclined seatback configurations. 

The research also extends to vulnerable occupants, such as small females, the obese and the elderly. The PMHS (Post Mortem Human Subjects) tests will be conducted in these configurations to better understand the kinematics and injury mechanisms. Biomechanical response corridors will be derived from this research program and can be used as guidelines to develop suitable ATDs for AV restraint system testing.   

In the coming months, THOR-AV 50M will be subjected to the same test conditions determined in the aforementioned NHTSA PMHS research projects. The BioRank score of the dummy will be calculated based on the corridors developed from the PMHS research projects. Where necessary, design improvements to the dummy will be made to achieve satisfactory biofidelity.