3. Dummy anatomy and instrumentation

An anthropomorphic test device (ATD) is only as useful as the measurements it produces. The dummy's metal and polymer skeleton defines its mass distribution and kinematics, but the instrumentation determines what engineers can actually observe and quantify during a crash event. This section examines the four primary sensor families used in modern ATDs — accelerometers, load cells, potentiometers, and angular rate sensors — and explains where each is mounted, what physical quantity it transduces, and how its output maps to injury criteria.

Accelerometers: Measuring Translational Response

Accelerometers convert translational acceleration into an electrical signal, typically via a piezoelectric or piezoresistive sensing element. In a crash test, they are the workhorse sensor because most injury criteria are expressed in terms of acceleration or its double integral (displacement).

Placement follows the body region being evaluated. The head accelerometer triad sits at the head's center of gravity (CG), with three orthogonal axes aligned to the SAE J211 coordinate system: +X forward, +Y to the dummy's left, +Z upward. This triad feeds the Head Injury Criterion (HIC), computed from the resultant acceleration:

HIC = max over t1,t2 of [ (t2 - t1) * ( (1/(t2-t1)) * ∫ a_res dt )^2.5 ]

where a_res is the resultant acceleration in g and the time window is limited to 36 ms. A typical frontal test might yield a peak resultant of 150–250 g at the head CG, with HIC values in the 400–700 range for a well-performing restraint system.

The chest accelerometer is mounted on the sternum or on the spine box, depending on the ATD family. The Hybrid III places a uniaxial accelerometer on the sternum for chest deflection correlation, while the THOR dummy uses a multi-axis array on the upper spine. Pelvis accelerometers are bolted to the iliac crest region to capture the occupant's lower-body deceleration, which drives submarining assessment.

Load Cells: Force and Moment Measurement

Load cells are strain-gauge bridges configured to measure force and moment along one or more axes. Unlike accelerometers, which measure the response of a mass, load cells measure the interface forces transmitted through a structural joint.

Key load cell locations include:

• Upper neck load cell — six-axis (Fx, Fy, Fz, Mx, My, Mz), mounted between the head and neck. It feeds the Neck Injury Criterion (Nij), which combines axial force and bending moment.
• Lower neck / upper thoracic load cell — captures the transition from cervical to thoracic loading.
• Lumbar spine load cell — measures the bending moment at L5/S1, relevant to seatback loading and whiplash.
• Femur load cells — axial force along the femur shaft, used for knee-thigh-hip (KTH) injury assessment.
• Tibia load cells — combined axial force and bending moment at the tibial plateau.

The six-axis neck load cell is the most critical for frontal and rear impact. Its moment output is filtered at CFC 600 (channel frequency class) per SAE J211, while forces are typically filtered at CFC 1000. The Nij calculation uses the peak axial force and moment, normalized by critical intercept values:

Nij = (Fz / Fzc) + (My / Myc)

For a 50th percentile male Hybrid III, Fzc = 6800 N in tension and Myc = 310 N·m in extension. A Nij above 1.0 indicates a probable AIS 3+ neck injury.

Potentiometers: Displacement and Deflection

Potentiometers are variable resistors that output a voltage proportional to linear or rotary displacement. In ATDs, they are used primarily for chest deflection and knee slider displacement.

The chest deflection potentiometer is mounted between the sternum and the spine box. As the sternum moves rearward under belt loading, the potentiometer's wiper translates, producing a voltage change calibrated to millimeters of deflection. The Hybrid III 50th percentile male has a maximum chest deflection of approximately 70 mm before bottoming out; the injury assessment reference value (IARV) is 63 mm for a 50th percentile male.

Knee slider potentiometers measure the relative displacement between the femur and tibia under axial impact, providing a direct indicator of knee joint compression. A typical threshold for knee injury is 6 mm of slider displacement.

Angular Rate Sensors: Rotational Velocity

Angular rate sensors (ARS), typically MEMS gyroscopes, measure angular velocity about one or more axes. They are increasingly common in ATDs because rotational head kinematics correlate strongly with diffuse axonal injury and subdural hematoma.

In the THOR and newer ATD designs, a three-axis ARS is mounted at the head CG alongside the accelerometer triad. The angular velocity signal is integrated to obtain head rotation angle, and combined with linear acceleration to compute the Brain Injury Criterion (BrIC):

BrIC = (ωx / ωxc) + (ωy / ωyc) + (ωz / ωzc)

where ωx, ωy, ωz are peak angular velocities and ωxc, ωyc, ωzc are critical values (typically 66.3, 53.1, and 41.5 rad/s for a 50th percentile male). A BrIC above 1.0 suggests a high risk of severe brain injury.

ARS placement must minimize cross-axis sensitivity. Mounting the sensor on a rigid block at the head CG, with axes aligned to the SAE J211 frame, reduces the risk of rotational crosstalk. Post-test calibration checks should confirm zero-rate offset drift of less than 0.5 °/s.

Integrated Sensor Layout and Signal Chain

The figure below shows a representative sensor layout for a 50th percentile male Hybrid III in a frontal impact configuration. Each sensor's output passes through a signal conditioning unit, an anti-aliasing filter, and an analog-to-digital converter before being recorded by the data acquisition system at a minimum sample rate of 10 kHz for accelerometers and 20 kHz for load cells.

Practical Calibration and Verification

Before every test series, each sensor channel must be calibrated end-to-end. The procedure differs by sensor type:

• Accelerometers: Perform a static tilt test. With the dummy upright, the Z-axis should read +1 g; rotate 90° forward and the X-axis should read +1 g. A deviation greater than 2% indicates a mounting or calibration issue.
• Load cells: Apply a known deadweight or use a shunt calibration resistor. For a 6-axis neck load cell, verify that a pure axial force produces negligible cross-axis output (less than 2% of full scale).
• Potentiometers: Drive the wiper through its full stroke with a micrometer and record the voltage at 10% increments. The linearity error should be less than 1% of full scale.
• Angular rate sensors: Use a rate table to apply a known angular velocity (e.g., 100 °/s) about each axis. Verify scale factor and zero offset. Cross-axis sensitivity should be less than 3%.

After the test, a post-test calibration check is equally important. A shift in zero offset greater than 5% of full scale suggests the sensor was overloaded or damaged. In that case, the affected channel should be flagged and the data reviewed for clipping or non-physical artifacts before being used in injury calculations.

Mapping Sensors to Injury Criteria

The table below summarizes the primary injury criteria and the sensors that feed them. This mapping is essential for understanding why sensor placement matters: a misplaced accelerometer or a saturated load cell directly compromises the validity of the injury assessment.

In practice, the data acquisition system records all channels simultaneously, and the post-processing software applies the appropriate filters and computes each criterion. The engineer's job is to verify that each channel is within its calibrated range, that no sensor has saturated, and that the computed criteria are consistent with the observed dummy kinematics. A HIC of 1200 with a peak head acceleration of only 80 g, for example, should trigger a review of the acceleration trace for noise or a filtering error.