This thesis ultimately aims to enhance occupant protection by incorporating aspects of realworld crash heterogeneity, often overlooked within current safety assessments. By investigating the effects of crash heterogeneity and broadening the comprehensiveness of occupant safety assessments, it seeks to support the development of more effective future vehicle safety systems. Specifically, the thesis focused on developing and applying methods to incorporate a range of heterogeneity aspects—from crash characteristics to occupant posture, anthropometry, and seat adjustments—into vehicle safety assessments.

To predict how crash avoidance systems might change the configurations of the remaining crashes, a method using counterfactual simulations was developed. The use of a novel crash configuration definition, along with a purpose-designed clustering method, reduced the number of predicted crash configurations—while being able to maintain coverage of diverse real-world situations. Three crash configurations were selected to be used in the following studies.

Non-nominal sitting postures, body sizes, and seat adjustments can influence the occupant’s response during a crash. These aspects were investigated in simulation studies employing numerical Human Body Models (HBMs) and tailor-made analysis methods. The methods focused on quantifying the influence of these aspects (including interaction effects) on the occupant’s response during a crash. Additionally, techniques were developed to streamline the setup and analysis of numerical experiments using HBMs.

The application of the methods indicated that autonomous emergency braking systems tend to move the crash locations towards the vehicle’s corners. Additionally, further studies showed that the occupants’ posture, anthropometry, and seat adjustments influenced their kinematic and kinetic crash response. Variations in lower extremity postures had the greatest effect on wholebody response across all tested crash scenarios. For example, in frontal collisions, sitting crosslegged increased pelvic movement, while seat adjustments altered load distributions between the pelvis and the lower extremities. Moreover, occupant characteristics could also induce differences:​ greater Body Mass Index (BMI) or stature correlated with larger lower extremity loading in frontal impacts. In side impacts, occupants were more sensitive to lateral movement when leaning forward.

Furthermore, the influence of individualising the shoulder belt placement on the occupant-tobelt interaction, without changing any other belt parameter, was investigated. The findings revealed that while improved initial belt placement over the shoulder is important, it alone does not guarantee improved seat belt interaction. This approach, by investigating seat belt interaction challenges for occupants with varying characteristics, paves the way for analysing further modifications in belt characteristics towards tailored occupant restraint systems.

By incorporating aspects not typically included in current safety assessments, this thesis demonstrates the potential to further enhance assessment for future vehicle safety systems, accommodating a broader range of real-world situations.