3. Child anthropomorphic test devices (ATDs)

Every child restraint simulation ultimately answers a question about a body: what loads does a real child experience, and would those loads cause injury? Since we cannot instrument living children in crash tests, the field substitutes anthropomorphic test devices (ATDs) — physical dummies whose geometry, mass distribution, and joint stiffness are engineered to approximate a child of a given age. In simulation, the ATD becomes a finite element or multibody model, and its fidelity determines how much confidence we can place in the predicted injury metrics. This chapter examines the two dominant families used in child occupant protection — the Q-series and the Hybrid III child dummies — and, critically, where their biofidelity breaks down.

Two Design Philosophies

The Q-series was developed in Europe (originally by TNO, now maintained under the Humanetics Q-dummy line) specifically for child restraint evaluation, and it is the reference dummy for UN Regulation No. 129 (the "i-Size" regulation). The Hybrid III family descends from the adult 50th-percentile male dummy developed in the United States and was scaled down to child sizes (the 3-year-old, 6-year-old, and 10-year-old variants are the most common). The two families differ in more than dimensions: they embody different assumptions about how a child's body should respond.

The Q-series uses a segmented, articulated spine with rubber elements tuned to child stiffness, a deformable thorax with integrated deflection measurement, and — in the Q0 through Q6 sizes — a design emphasis on frontal and side impact in child seats. The Hybrid III child dummies inherit the adult dummy's rigid metallic spine, a single-piece deformable chest with a chest potentiometer, and a head with a compliant vinyl skin over an aluminum skull. The Hybrid III 3-year-old is the workhorse of FMVSS 213 compliance testing in the United States.

Anatomy of the Models and What They Measure

In a crash simulation, the ATD model is a kinematic chain of rigid bodies connected by joints, with contact surfaces and instrumentation channels. The channels that matter most for child occupants are:

• Head resultant acceleration (three-axis accelerometer at the head CG) — used for HIC15/HIC36 and, in Q-dummies, for the head injury criterion applied in i-Size.
• Chest deflection — measured by a potentiometer (Hybrid III) or integrated deflection sensor (Q-series); the primary thoracic injury predictor.
• Neck forces and moments — upper and lower neck load cells; relevant to both families but with different injury reference values.
• Pelvis and abdomen — Q-series includes abdominal pressure sensors (Q3 and above) that Hybrid III child dummies lack entirely.

These channels are not interchangeable between families. A chest deflection of 40 mm in a Q3 model and 40 mm in a Hybrid III 3-year-old do not represent the same physical event, because the measurement location, the thorax stiffness, and the deflection definition differ. Any simulation study that mixes ATD families in a comparison must state which dummy and which channel definition it used.

Biofidelity Limitations

Biofidelity is the degree to which an ATD reproduces the human response to impact. Both families have documented limitations that a simulation engineer must carry into the interpretation of results.

Geometry and mass scaling. Child ATDs are scaled approximations, not anthropometric replicas. The Hybrid III 3-year-old, for example, has a head mass and stiffness that were tuned for frontal compliance testing rather than derived from pediatric biomechanics data. The Q-series improved on this with child-specific anthropometry, but the smallest sizes (Q0, Q1) still rely on limited neonatal and infant data.

Spine and thorax stiffness. The Hybrid III child's rigid spine and relatively stiff chest tend to underpredict the flexibility of a real child's torso, which can bias chest deflection and head excursion predictions. The Q-series' segmented spine is more compliant but introduces its own modeling challenge: the joint stiffness values are calibrated to a narrow set of cadaver and volunteer data, and their validity outside the frontal impact regime is uncertain.

Soft tissue and abdominal injury. Neither family captures abdominal organ injury well. The Q-series' pressure sensors are a step forward, but the abdomen is still a simplified cavity. Hybrid III child dummies have no abdominal instrumentation at all, so any simulation claiming to assess abdominal injury risk with a Hybrid III model is extrapolating beyond the dummy's capability.

Side impact and oblique loading. The Q-series was designed with side impact in mind (Q6 and Q10 are used in side-impact regulation), but the biofidelity of the shoulder and pelvis under lateral loading remains an active research area. Hybrid III child dummies were primarily validated for frontal loading; using them for lateral or oblique simulations without acknowledging this is a common methodological error.

Choosing and Using an ATD Model in Simulation

The choice of ATD model should follow the regulatory or research question, not convenience. For i-Size / UN R129 evaluations, use the Q-series (Q0, Q1, Q3, Q6, Q10 as appropriate to the seat category). For FMVSS 213, use the Hybrid III child dummies specified in the standard. For research into abdominal or side-impact injury, prefer the Q-series and treat Hybrid III results as a lower-bound estimate of instrumentation coverage.

A practical workflow for a frontal sled simulation with a Q3 in a forward-facing child seat:

1. Import the Q3 FE model and verify its mass (approximately 15 kg) and joint stiffness against the model documentation.
2. Position the dummy per the installation procedure for the seat and restraint, and record the H-point and joint angles as model inputs.
3. Apply the sled pulse (e.g., a 50 km/h frontal pulse) as a boundary condition to the seat and floor.
4. Request the head acceleration, chest deflection, and neck load channels at the standard sampling rate (typically 10 kHz for physical tests; FE models may output at the solver's time step).
5. Post-process HIC15 and chest deflection, and compare against the Q3 reference values for the loading direction.

If the predicted chest deflection is implausibly low (near zero) while head excursion is large, the likely cause is a contact definition error between the harness and the dummy thorax, or an incorrect restraint routing — not a biofidelity issue. If chest deflection is high but the pulse is mild, check the thorax material model and the deflection sensor orientation before concluding the restraint is inadequate.

Interpreting Results with Biofidelity in Mind

When a simulation predicts a chest deflection of, say, 32 mm for a Q3 in a forward-facing seat under a 50 km/h pulse, the number is only meaningful relative to the Q3's reference value for that loading condition. The same physical restraint evaluated with a Hybrid III 3-year-old might yield a different deflection because the thorax stiffness and measurement definition differ. The engineer's job is not to declare one dummy "correct" but to state which ATD was used, which channels were read, and which reference values were applied — and to flag where the ATD's known limitations (abdominal injury, side impact, infant sizes) leave the conclusion uncertain.

In practice, this means treating the ATD model as a measurement instrument with a documented calibration range. Use it inside that range, report the instrument alongside the result, and where the question falls outside the instrument's validated domain, say so explicitly rather than presenting a number as if it were a direct measure of child injury risk.