2. Child occupant anatomy and injury biomechanics

A crash dummy is only as valid as the biological target it represents. Before any ATD (anthropomorphic test device) geometry, mass distribution, or instrumentation channel can be justified, the designer must answer three questions: how does a child's body differ structurally from an adult's, how do those differences translate into different injury thresholds, and how are the resulting injuries classified so that dummy response can be mapped onto a clinically meaningful outcome? This section addresses those three questions in sequence, because they form a dependency chain — scaling defines the target, tolerances define the pass/fail boundary, and AIS coding defines the language in which the boundary is expressed.

Why Children Are Not Scaled Adults

The naive approach to child dummy design is geometric scaling: take the 50th-percentile adult male, shrink it by a mass ratio, and adjust spring rates accordingly. This fails because pediatric anatomy is not a uniform reduction. Three structural differences dominate the injury mechanics of the child occupant.

First, the head is disproportionately large and heavy. At birth the head accounts for roughly 25% of total body mass; by age 6 it is still around 16–18%, compared with about 6% in adults. The center of gravity of the head sits higher relative to the neck, and the neck itself is short, with immature, incompletely ossified cervical vertebrae and relatively lax ligamentous restraint. The consequence is a head-neck system with a high inertial moment and limited bending stiffness — the classic mechanism behind the higher incidence of cervical spine and brain injury in young occupants.

Second, the pediatric thorax is highly compliant. The rib cage is cartilaginous, the ribs are more horizontally oriented, and the chest wall offers far less resistance to deflection than the adult rib cage. This compliance protects against rib fracture but permits large internal organ excursion, which is why pulmonary and cardiac contusion, and aortic injury from rapid deceleration, occur at deflection levels that would be benign in an adult. It also means that chest deflection measured on a child ATD must be interpreted against child-specific, not adult, tolerance curves.

Third, the pelvis and abdomen are immature. The iliac crests are not fully ossified, the pelvic structure provides less effective lap-belt anchorage, and the abdominal wall is thin with organs positioned relatively anteriorly. This produces the well-documented "seat-belt syndrome" pattern — submarining under the lap belt with abdominal organ compression and lumbar spine distraction — at belt loading levels that an adult pelvis would tolerate without injury.

Scaling Principles Used in Dummy Design

Because direct pediatric cadaver and volunteer data are scarce, dummy designers rely on scaling relationships derived from dimensional analysis and from the limited pediatric biomechanical literature. The most widely used framework is mass-based geometric scaling, in which a reference adult response is transformed to a child target using a length scale factor λ and a mass scale factor λ³, with derived scaling for force (λ²), acceleration (λ⁰, i.e., unchanged), velocity (λ), time (λ), and bending moment (λ³).

In practice, the designer does not apply a single global scale factor. Instead, body regions are scaled independently. A typical design workflow for a 6-year-old (6YO) ATD might proceed as follows, using published pediatric anthropometry as the input:

Region        Adult ref.   6YO target    Scale basis
Head mass     4.5 kg       2.2 kg        pediatric head mass data
Head CG height 1.62 m      1.05 m        seated anthropometry
Neck length   0.12 m       0.07 m        cervical spine imaging
Chest depth   0.23 m       0.16 m        chest CT anthropometry
Chest mass    25 kg        9 kg          thorax mass fraction
Pelvis width  0.36 m       0.22 m        bi-iliac breadth

These are illustrative target values consistent with published 6YO anthropometry, not measured outputs of a specific dummy. The point is that each region carries its own scale basis, and the designer must document which source supports each target. Where pediatric data are absent — for example, cervical ligament stiffness — the designer falls back to scaled adult data and flags the resulting tolerance as low-confidence.

Injury Tolerances and Their Regional Structure

Injury tolerances for child ATDs are expressed as region-specific criteria, each with a threshold and a duration or probability basis. The table below summarizes the criteria most relevant to child dummy design and the physical quantity each instrumented channel must capture.

Two structural features of these tolerances matter for dummy design. First, they are probability-based: a threshold such as HIC = 1000 corresponds to a defined risk of a given AIS level, not a binary safe/unsafe line. The designer must therefore decide which risk level the dummy is intended to discriminate, and instrument accordingly. Second, several criteria are interaction criteria — N_ij combines axial compression or tension with bending moment — which means the dummy must measure both channels simultaneously and with matched bandwidth, or the criterion cannot be evaluated.

AIS Coding as the Output Language

The Abbreviated Injury Scale (AIS) provides the ordinal severity classification that connects dummy response to clinical outcome. AIS ranks injuries from 1 (minor) to 6 (currently untreatable), with 9 reserved for unknown severity. For child occupant work, the relevant codes cluster in specific regions and mechanisms.

Head injuries in children are coded by lesion type: cerebral contusion and subdural hematoma carry AIS 3–4 depending on extent, while diffuse axonal injury and brainstem lesions escalate to AIS 5. Neck injuries are coded by structure: cervical spine fracture without cord involvement is typically AIS 2–3, while cord injury with incomplete or complete deficit reaches AIS 4–5. Thoracic injuries follow a similar logic — rib fracture is AIS 2, flail chest AIS 4, aortic laceration AIS 5 — but in children the more common thoracic lesions are pulmonary contusion and cardiac injury at lower force levels. Abdominal injuries are coded by organ and by whether the capsule is breached: liver or spleen laceration without major hemorrhage is AIS 2–3, with major hemorrhage AIS 4.

The design consequence is that the dummy's instrumentation must resolve the lesion type, not merely the force magnitude. A chest deflection channel that reports a single peak value cannot distinguish rib fracture from pulmonary contusion, yet these map to different AIS codes and different dummy design priorities. This is why modern child ATDs increasingly carry multiple, regionally distributed sensors rather than a single load cell per body region.

Putting the Chain Together

The three elements form a single design logic. Anatomy defines which regions are vulnerable and why; scaling converts adult reference data into child-specific targets where pediatric data are missing; tolerances convert those targets into measurable thresholds; and AIS coding converts thresholds into clinical meaning. A child dummy that measures the right quantity at the right location, against a tolerance traceable to a documented scaling basis, and reports it in a form mappable to AIS, is a defensible design. One that skips any link in that chain — for example, applying an adult chest deflection threshold to a compliant pediatric thorax — will produce a device that passes its own criteria while misrepresenting the injury risk it was built to assess.

The figure above makes the dependency explicit: anatomy sets the vulnerable regions, scaling converts adult reference data into child targets, tolerances turn those targets into measurable thresholds, and AIS coding gives the thresholds clinical meaning. The dashed feedback path is not decorative — it represents the iterative refinement that occurs when dummy response is compared against post-mortem human subject (PMHS) data or field accident reconstructions, and it is the mechanism by which low-confidence scaled tolerances are eventually replaced with pediatric-specific values.