4. Anthropomorphic test devices (ATDs) for children

Child restraint system (CRS) development and regulatory compliance both depend on a physical surrogate that behaves like a child during a crash: the child anthropomorphic test device (ATD), commonly called a crash dummy. Unlike adult ATDs, which converged on a relatively small number of standardized designs, the child ATD landscape is fragmented across several families, each with its own geometry, mass distribution, joint stiffness, instrumentation, and regulatory pedigree. Understanding why these families exist and how they differ is essential before interpreting any injury criterion or comparing test results across programs.

From scaled adults to dedicated child surrogates

Early child crash testing simply scaled down adult dummies. The rationale was geometric similarity: if a 50th-percentile adult male Hybrid II could be scaled by mass and length ratios, the resulting surrogate would approximate a child. This assumption failed for two reasons. First, a child is not a uniformly scaled adult — the head is proportionally much larger and heavier relative to the body, the neck is short and compliant, the rib cage is more cartilaginous and deflects differently, and the pelvis is immature. Second, scaling laws cannot reproduce the nonlinear stiffness of biological tissue, so a scaled adult dummy produced head accelerations and chest deflections that did not match pediatric post-mortem human subject (PMHS) or field data.

The first purpose-built child surrogates appeared in the 1970s and 1980s, driven by the introduction of child restraint legislation. These early devices were crude, often with a rigid head and a simple hinged neck. The turning point came with the development of the Hybrid III family, which was extended downward from the adult 50th-percentile design to a 3-year-old (and later a 6-year-old and a 10-year-old) through the Society of Automotive Engineers (SAE) and NHTSA-sponsored work. In parallel, European researchers pursued a different philosophy that eventually produced the Q-series, and the CRABI (Child Restraint Air Bag Interaction) family was created specifically to study air-bag–child interaction in the late 1990s.

The three principal families

Hybrid III 3-year-old. The Hybrid III 3-year-old (HIII-3YO) is a direct descendant of the adult Hybrid III. It shares the adult design language: a segmented, deformable chest with a chest deflection potentiometer, a six-axis upper-neck load cell, a head with a triaxial accelerometer block, and a lumbar spine that transmits load to a rigid pelvis. Its mass is approximately 15.9 kg and its standing height about 0.95 m. Because it inherits the Hybrid III biofidelity philosophy, it is the reference dummy for FMVSS 213 (US child restraint standard) and is widely used in sled and vehicle crash testing. Its principal limitation is that it was calibrated primarily against a limited set of PMHS data and does not represent an infant or a newborn; it is a single-point design in the child size range.

Q-series. The Q-series was developed in Europe (originally by TNO and now maintained under UN ECE Regulation 44/129 and Euro NCAP) as a family spanning Q0 (newborn), Q1, Q1.5, Q3, Q6, and Q10. Each size is a dedicated design rather than a scaled version of another. The Q-series uses a different instrumentation architecture: load cells are integrated into the dummy's spine and pelvis, and the head and chest use accelerometers and deflection sensors calibrated to European PMHS and field data. The Q3, roughly equivalent in size to the HIII-3YO, has a mass of about 15 kg but a noticeably different head mass distribution and neck stiffness. The Q-series is the reference for UN R129 (i-Size) and is increasingly used in Euro NCAP frontal and side impact assessments.

CRABI. The CRABI family (CRABI 12-month and CRABI 18-month, with a 6-month variant) was developed by NHTSA specifically to evaluate child–air bag interaction. Its defining feature is a very compliant, low-stiffness neck designed to represent the weak cervical spine of an infant, combined with a head and chest instrumented to capture the loading produced when a deploying air bag strikes a rear-facing child. CRABI is not a general-purpose crash dummy for restraint certification; it is a specialized tool for out-of-position (OOP) and air-bag interaction studies. Its mass is approximately 10 kg (12-month) and 12 kg (18-month).

Where the design differences actually matter

The most consequential difference between families is the neck. The Hybrid III 3-year-old neck is a molded butyl-rubber cylinder with a defined moment–angle response calibrated to adult-derived scaling; it is relatively stiff in flexion and extension. The Q3 neck uses a different elastomer and geometry tuned to European pediatric data, producing a softer, more progressive response. The CRABI neck is softer still, deliberately so, because an infant's cervical spine cannot carry the moment that an older child's can. When the same CRS is tested on an HIII-3YO and a Q3, the head excursion and neck injury metrics (Nij, Nkm) can differ by 15–30% purely because of neck stiffness — a difference that has real consequences when a restraint passes one program and fails another.

The chest is the second differentiator. The Hybrid III chest is a steel-and-rubber assembly with a potentiometer measuring sternal deflection; its force–deflection corridor was derived from adult data scaled to child size. The Q-series chest is built around a deformable rib structure with integrated load cells in the spine, allowing direct measurement of spinal compression rather than inferring it from sternal displacement. For CRS work, this matters because a rear-facing restraint loads the child's back, not the sternum, so a chest-deflection-based criterion can be misleading; the Q-series spine load cells capture the actual load path.

The head differs in mass and mass moment of inertia. The HIII-3YO head is heavier relative to its neck than the Q3 head, which raises the inertial moment during a frontal pulse and tends to produce higher head accelerations. CRABI's head is smaller and lighter, consistent with an infant, and its instrumentation is optimized for the short-duration, high-acceleration event of an air-bag deployment rather than a 50 ms sled pulse.

Practical implications for test selection

Choosing the right ATD is a test-design decision, not a formality. A few rules of thumb:

• If the deliverable is a US FMVSS 213 certification or a US NCAP rating, the HIII-3YO (and its 6YO/10YO siblings) is mandatory; results from a Q3 are not accepted as a substitute.
• If the deliverable is UN R129 / i-Size approval or a Euro NCAP assessment, the Q-series is the reference; the Q3 is the closest analogue to the HIII-3YO but is not interchangeable.
• If the question is whether a deploying air bag injures an infant in a rear-facing CRS, CRABI is the appropriate surrogate; using an HIII-3YO for this purpose overestimates neck stiffness and can mask a real hazard.
• When comparing data across programs, never pool HIII and Q-series results without documenting the surrogate; the injury criteria (HIC, Nij, chest deflection) are computed the same way but the underlying dummy response differs.

Calibration and verification in practice

Before any test series, each ATD must be verified against its calibration manual. For the HIII-3YO, the standard checks include a head drop test (drop height and peak acceleration window), a neck pendulum test (moment–angle corridor at specified velocity), and a chest calibration (force–deflection corridor). For the Q3, the manual specifies equivalent head drop, neck, and chest checks with different corridors. A practical verification sequence is:

1. Record ATD serial number, family, and calibration due date.
2. Head drop: release from specified height; confirm peak resultant
   acceleration falls inside the manual's corridor (e.g., HIII-3YO
   head drop peak typically 200–250 g; confirm against manual).
3. Neck pendulum: mount in fixture, apply specified pulse; plot
   moment vs. angle; confirm the curve lies within the corridor.
4. Chest: apply calibration force; confirm deflection within corridor.
5. Log ambient temperature and humidity; elastomer stiffness is
   temperature-sensitive and out-of-band conditions invalidate results.
6. If any check fails, do not test; replace the affected component
   and repeat the full verification.

Temperature is the most frequently overlooked variable. The butyl-rubber neck and chest elements stiffen at low temperature and soften at high temperature; a dummy calibrated at 22 °C and tested at 15 °C will produce systematically different neck moments. Most manuals specify a test window of roughly 20–22 °C and 30–70% relative humidity, and the ATD should be soaked at test conditions for a defined period before the run.

Finally, instrumentation channels must be zeroed and ranged correctly. Head accelerometers are typically ranged to ±2000 g for child dummies; neck load cells to the manual's specified ranges; chest potentiometers to the full deflection range. A saturated channel produces a clipped peak that under-reports injury risk, so always inspect the raw time histories for flat-topped peaks before accepting a computed HIC or Nij value.