3. Anthropomorphic test devices (ATDs)

Anthropomorphic test devices (ATDs) are the calibrated human surrogates that convert a crash pulse into measurable occupant injury metrics. In FMVSS occupant protection work, the choice of ATD is not a matter of preference — it is dictated by the regulation, the seating position, the restraint configuration, and the injury criterion being evaluated. This chapter covers the five ATD families most frequently encountered in FMVSS compliance programs: the Hybrid III 50th percentile male, the Hybrid III 5th percentile female, the THOR 50th percentile male, the WorldSID 50th percentile male, and the CRABI child dummy used for air bag interaction testing.

3.1 The ATD Family and Its Regulatory Anchors

Each ATD exists because a specific regulation or test procedure requires it. FMVSS 208 frontal occupant protection is built around the Hybrid III family: the 50th percentile male (part 572 Subpart E) for driver and outboard front passenger positions, and the 5th percentile female (part 572 Subpart O) for small-stature occupants and for the belted/unbelted matrix. FMVSS 208 also references the CRABI 12-month-old (part 572 Subpart R) for air bag suppression and low-risk deployment evaluation. FMVSS 214 side impact uses the WorldSID 50th percentile male (part 572 Subpart W) in the current rule, while the ES-2re remains in some legacy procedures. THOR is not yet mandated in FMVSS 208, but it is the reference ATD in NHTSA's advanced frontal research and in the THOR-based test procedures that inform future rulemaking.

The practical consequence is that a compliance engineer must know, for each test, which ATD is called out, which part 572 subpart defines its calibration, and which channels are required versus optional. A test run with the wrong ATD, or with an ATD outside its calibration window, is not a valid compliance test regardless of how clean the data looks.

3.2 Hybrid III 50th Percentile Male

The Hybrid III 50th percentile male is the workhorse of frontal compliance testing. It is a 78 kg, 175 cm seated surrogate with a segmented steel and aluminum skeleton, a vinyl skin, and a deformable chest. Its instrumentation set for FMVSS 208 typically includes:

• Head: triaxial accelerometer at the head CG, used for HIC15 and HIC36.
• Neck: upper-neck load cell measuring Fx, Fy, Fz, Mx, My, Mz for Nij.
• Chest: chest deflection potentiometer (or IR-TRACC on newer builds) plus triaxial accelerometers for chest acceleration and the 3 ms clip.
• Femur: axial load cells for femur force and the femur force criterion.
• Tibia: axial load and bending moment for the tibia index.
• Pelvis: triaxial accelerometer for pelvis acceleration.

Setup begins with the ATD sitting in the specified posture: the H-point is set using the SAE J826 HPM, the head is leveled, and the limbs are positioned per the test procedure. The seat belt is routed over the pelvis and chest, and the belt is tensioned to the specified pre-load. A common failure mode is a belt routed over the chest deflection sensor rather than beside it, which biases the deflection reading. Another is a head that is not level, which rotates the head CG accelerometer and corrupts the HIC calculation.

3.3 Hybrid III 5th Percentile Female

The 5th percentile female is a 49 kg, 150 cm surrogate used to represent small-stature occupants. Its structure mirrors the 50th male but with scaled geometry and lower mass. The critical setup difference is that the smaller occupant sits closer to the air bag module, so the out-of-position (OOP) and air bag interaction tests are more sensitive to initial posture. The chest deflection sensor has a shorter stroke, and the femur load cells have lower full-scale ratings, so the data acquisition gain must be adjusted accordingly. A 5th percentile female run that uses the 50th male channel gains will clip the femur and chest channels and produce unusable data.

3.4 THOR 50th Percentile Male

THOR (Test Device for Human Occupant Restraint) is the advanced frontal ATD. It differs from Hybrid III in three ways that matter for compliance engineering: a more human-like spine with a flexible thoracic and lumbar section, a four-point chest deflection measurement (IR-TRACC) that captures asymmetric loading, and a more biofidelic pelvis and femur. THOR also carries a greater channel count — typically 40 to 60 channels — including thoracic spine loads, abdominal pressure, and iliac wing loads.

Because THOR is not yet a mandated FMVSS 208 ATD, its use in compliance work is currently limited to research and to procedures that reference it explicitly. When setting up THOR, the most common issues are sensor zeroing drift after the ATD has been seated for a long period, and incorrect routing of the IR-TRACC cables, which can pre-load the chest and bias deflection. Always zero the chest channels after the ATD is in its final seated posture, not before.

3.5 WorldSID 50th Percentile Male

WorldSID is the side impact ATD used in FMVSS 214. It is a 72 kg, 175 cm surrogate with a fully instrumented thorax, abdomen, and pelvis. Its key channels include:

• Head: triaxial accelerometer for HIC and head excursion.
• Thorax: three rib modules (upper, middle, lower) each with a triaxial accelerometer and a deflection sensor, used for the side impact thoracic trauma index and rib deflection criteria.
• Abdomen: two rib modules with deflection sensors.
• Pelvis: triaxial accelerometer and pubic symphysis load cell.

WorldSID setup is dominated by the side impact environment: the ATD is seated on the struck-side seat, the door trim and armrest are in contact, and the ATD must be positioned so that the struck-side arm is in the specified position (typically 90 degrees forward). The rib deflection sensors are calibrated with a known displacement, and any contact between the rib modules and the door trim before impact will pre-load the sensors and invalidate the deflection data. A useful check is to verify that the rib modules read zero after the ATD is fully seated and the door is closed.

3.6 CRABI Child Dummy

The CRABI 12-month-old is a 10 kg child surrogate used for air bag interaction testing, particularly for the suppression and low-risk deployment requirements of FMVSS 208. Its instrumentation is simpler than the adult ATDs: head triaxial accelerometer, chest triaxial accelerometer, and in some builds a chest deflection sensor. The critical setup issue is the child restraint: the CRABI must be installed in the specified child seat, and the seat must be installed in the vehicle per the child seat manufacturer's instructions. The air bag module is then deployed, and the CRABI's head and chest accelerations are recorded to assess injury risk.

Because the CRABI is small and light, the air bag deployment can produce high head accelerations even at low deployment energy. The test engineer must verify that the CRABI is positioned so that the head is not in direct contact with the air bag module cover before deployment, and that the child seat is not modified in any way that would alter the interaction.

3.7 Instrumentation and Data Quality

All ATD channels are routed through a data acquisition system with anti-alias filtering and a sample rate of at least 10 kHz for acceleration channels and 10 kHz for load and deflection channels. The channels are zeroed before the test, and the zero offset is recorded. After the test, the data is filtered per SAE J211 (CFC 1000 for acceleration, CFC 600 for loads, CFC 180 for deflection) before any injury criterion is calculated.

A practical checklist for ATD setup and instrumentation:

The map above summarizes the ATD-to-test assignment and the channel set that each device contributes. In practice, the engineer should treat this as a starting point: the specific test procedure may add or remove channels, and the calibration requirements in the relevant part 572 subpart always take precedence. The next chapter moves from ATD selection into the calibration and verification procedures that keep these devices within their required performance windows.