**2.1 Occupant model validation**

The Global Human Body Models Consortium (GHBMC) M50-O average male seated occupant is a widely used and validated HBM. The detailed M50-O v4.5

occupant model has been validated at the tissue, component, and full body crash test levels during the development. GHBMC models were developed from external anthropometry and posture specific medical image data [10, 11]. A multi-modality image dataset from volunteers representing various target anthropometries were used. Over 14 thousand images across three imaging modalities (CT, MRI, and upright MRI) were collected for the M50 model including scans in the supine, standing, and seated postures [12]. Geometries for the M50 were developed from these image datasets using a variety of segmentation techniques. Segmented data were verified against or augmented with data from literature sources [13].

Component level validation has been conducted in each major body region. In the head, the model was validated by comparing the response in matched simulation of impacts to bony structures (e.g. maxilla [14], zygoma [10], nasal bone [15], and skull [16]). Various soft tissue injuries were validated including intracranial pressures [17, 18], and relative brain to skull motion [19, 20]. In the neck, segment level tests were validated using individual functional units [21–24], functional units and full spine in tension [25], ligamentous strain [26–29], and axial rotation [30, 31]. The full cervical spine was tested in various configurations including rear impact [32, 33], lateral impact, and frontal impact [34, 35]. The thorax was validated at the rib cage level using a denuded rib cage study [36], pendulum impacts [37–39], and table top impacts [40]. The individual response of a single rib was the subject of an optimization study [41]. The abdomen was validated in various tests using bar impacts with a free back [42, 43], belt loading with a free back [43], airbag loading with a fixed back [43], belt loading at the mid abdomen [44, 45], pendulum impact [46], organ level validation at impact [47], lumbar flexion [48], lateral impact [49] and side airbag loading [50]. The pelvis was validated in lateral compression at the acetabulum [51–53] and pubic symphysis [54]. The lower extremity has been validated in various loading conditions including axial loading. The ankle has been specifically studied in impacts for axial loading, ankle inversion, eversion, dorsiflexion, and rotation [55–58]. Furthermore, the tibia has been validated in a three-point bending setup as well as axial loading for the entire leg [59, 60].

The model has also been extensively validated at the full body level in classical macro-level injury biomechanics studies [61]. Along with validation in dynamic simulations, the mass distribution of the GHBMC M50 model was validated [62] by virtually sectioning the model into body regions and comparing masses to anthropometric PMHS data from McConville et al. [63] and Robbins [64]. Rigid impacts to body regions (e.g. thorax, pelvis, etc.) in frontal, oblique, and lateral directions were applied to the model based on experimental designs in the literature [43, 65–69]. Care was taken to closely approximate the experiment, including considerations of motion constraints on impactors and the inclusion of gravity, seat backs, etc. Three sled tests in frontal [70, 71] and lateral [72] directions have been validated. An example of these simulations is shown below (**Figure 1**).

The M50-OS v1.8.3 model provides relevant biomechanical output data from the same body habitus as the detailed model, but at a substantially reduced computational cost. The simplification process included reducing the total number of elements through re-meshing, consolidating contact definitions, utilizing simplified material properties, and implementing kinematic joints throughout the body. The M50-OS model exhibits roughly a 40-fold decrease in run time (**Table 1**). Since joint definitions and meshes were designed to maximize the ability to position the model, a semiautomated positioning "tree" was programmed into LS-PrePost allowing the user to dynamically adjust joint angles prior to running a simulation.

*Accidental Injury Analysis and Protection for Automated Vehicles DOI: http://dx.doi.org/10.5772/intechopen.105155*

#### **Figure 1.**

*Sample validation using the GHBMC M50 model. Left to right: Shaw et al. sled buck model setup, head Z displacement and lower left rib displacement.*


#### **Table 1.**

*Run time results for the simplified M50 occupant model.*

The published work on the M50-OS v1.8.3 model [73] reported thirteen validations and robustness simulations, which included denuded rib compression at 7 discrete sites, 5 rigid body impacts, and one sled simulation. Perez-Rapela et al. [74] compared simulated kinematics in their far-side impact sled tests with the M50-OS v1.8.3 model to the PMHS responses. Results showed that, in general, the model captured lateral excursion in oblique impact conditions but overpredicted in purely lateral impact conditions. The human body model obtained a "good" CORA score for the correlation of their evaluation.

## **2.2 Injury risks assessment**

The injury measures are Head Injury Criterion (HIC36 for side and HIC15 for frontal) and Brain Injury Criterion (BrIC) for the head region, Chest Lateral Deflections (side and frontal) for the chest, Abdominal VC for the abdomen, Pubic symphysis peak force for the pelvis, and Femur Force for the KTH region, and Upper Tibia Force and RTI for the lower extremities, respectively. The Full Body Injury Index FBII was defined as a summation of all the body region injury probabilities. **Table 2** summarizes these injury measures and the body region injury risk functions.

In this study, the human occupant injury risks were estimated with these probability functions. The estimations served as comparative measures for the body region injury severities among different analysis cases.
