2. Occupant load assessment in MPDB
In the Moving Progressive Deformable Barrier (MPDB) test, the barrier face is not the only thing that deforms. The occupant compartment is subjected to a complex, multi-axis deceleration pulse that differs substantially from a rigid-wall or full-width barrier event. Because the MPDB is designed to expose structural incompatibility between vehicles of different mass and stiffness, the occupant load assessment must capture both the severity of the pulse and the way restraint systems manage it. This section explains how that assessment is structured: which injury criteria are evaluated, how the Anthropomorphic Test Devices (ATDs) are instrumented, and how the resulting data channels are organized, filtered, and interpreted.
Why MPDB Occupant Load Differs from Conventional Frontal Tests
In a full-width rigid barrier test, the vehicle decelerates against an essentially non-yielding surface, producing a short, high-magnitude pulse. In the MPDB, the deformable barrier element and the moving trolley together create a longer, lower-amplitude pulse with a pronounced ride-down phase. The occupant therefore experiences a different loading history: restraint coupling occurs earlier relative to peak vehicle deceleration, and the risk of "bottoming out" the restraint (airbag fully compressed, belt force limiter exhausted) shifts. This is precisely why occupant load assessment in MPDB cannot simply reuse rigid-barrier thresholds without justification; the criteria and the data processing chain must be matched to the pulse characteristics of the compatibility test.
Injury Criteria Evaluated
The occupant load assessment typically reports a defined set of injury criteria for the driver and, where specified, the front passenger. Each criterion is computed from a specific channel or combination of channels, and each has an associated acceptance limit. The table below summarizes the criteria commonly used in MPDB occupant load evaluation; the specific limits are protocol-dependent and should always be taken from the governing regulation or rating procedure rather than assumed.
| Criterion | Body region | Primary data source | What it captures |
|---|---|---|---|
| HIC15 / HIC36 | Head | Head CG triaxial accelerometer | Head injury risk from resultant translational acceleration over the integration window |
| Head rotational criteria (BrIC, or protocol-specific) | Head / brain | Head CG angular rate sensors | Rotational loading associated with diffuse axonal injury |
| Nij (neck injury criterion) | Neck | Upper neck load cell (Fx, Fz, My) | Combined axial force and bending moment at the occipital condyle level |
| Chest deflection / compression | Thorax | Chest potentiometer or IR-TRACC | Rib cage intrusion relative to the spine box |
| Chest acceleration (3 ms clip) | Thorax | Chest triaxial accelerometer | High-rate thoracic loading |
| Femur axial force | Lower extremity | Femur load cells (left/right) | Axial compression transmitted through the knee-thigh-hip complex |
| Knee / tibia loading | Lower extremity | Knee slider and tibia load cells | Localized impact and bending loads |
Two practical points deserve emphasis. First, HIC is computed from the resultant acceleration at the head center of gravity, not from a single axis; feeding only the X-axis channel into the calculation is a common and serious error. Second, Nij requires the upper-neck load cell channels to be combined with the correct moment arm and the protocol-defined critical values; the sign convention for Fz (tension positive) and My (flexion positive) must be verified against the dummy manufacturer's documentation before any calculation is attempted.
Dummy Instrumentation and Channel Mapping
The ATD is the measurement platform. For a Hybrid III 50th percentile male, the standard MPDB instrumentation set includes head CG accelerometers (Ax, Ay, Az) and angular rate sensors, an upper-neck six-axis load cell, chest accelerometers and a chest deflection sensor, and femur/knee/tibia load cells. Each physical transducer maps to a named data channel in the acquisition system. The mapping must be documented and verified, because a swapped left/right femur channel or an inverted chest deflection polarity will silently corrupt the assessment.
Data Acquisition, Filtering, and Processing Chain
Raw transducer signals pass through signal conditioning, anti-alias filtering, and digitization before they become usable channels. The processing chain matters as much as the sensors themselves. The figure below shows the typical signal path from transducer to reported criterion, including the filter classes applied at each stage.
The filter class is not a cosmetic choice. Acceleration channels are typically filtered to CFC 1000, load channels to CFC 600, and deflection channels to CFC 180, following the channel frequency class definitions in the relevant standards. Applying the wrong filter class changes the peak value and therefore the criterion. For Nij, the force and moment channels must share the same filter class and time base; mixing a CFC 1000 force with a CFC 600 moment produces a physically meaningless combination.
Worked Example: Computing HIC15 from Head CG Data
Suppose the post-processed head CG resultant acceleration a(t) is available at 10 kHz over a 150 ms window. HIC15 is defined as the maximum over all 15 ms windows of:
HIC15 = max over t1,t2 with (t2 - t1) = 15 ms of:
[ (1/(t2 - t1)) * ∫ a(t) dt from t1 to t2 ]^2.5 * (t2 - t1)In practice this is computed numerically: for each start index i, integrate the resultant acceleration over the 15 ms window, divide by the window duration to obtain the mean acceleration in g, raise to the power 2.5, and multiply by the window duration in seconds. The maximum across all windows is the reported HIC15. A common implementation error is to use the vector sum of already-filtered per-axis channels without re-computing the resultant after filtering, or to integrate in the wrong units (g versus m/s²). The result is sensitive to both.
Interpreting the Occupant Load Result in Context
A criterion value below its limit does not by itself confirm a good occupant outcome, and a value near the limit does not automatically mean failure of the structure. In MPDB, the occupant load must be read together with the barrier deformation pattern and the vehicle pulse. A stiff, heavy vehicle may produce a benign occupant pulse while imposing severe deformation on the barrier and on the opposing vehicle — the compatibility problem the test exists to reveal. Conversely, a vehicle that crushes efficiently may show a higher occupant load because the ride-down is longer and the restraint is engaged over a larger portion of the pulse. The assessor should therefore report occupant criteria alongside the vehicle deceleration pulse characteristics and the barrier intrusion measurements, and should flag cases where the occupant load and the structural compatibility metrics point in different directions.
Finally, the data channels themselves are part of the deliverable. A complete MPDB occupant load dataset includes the raw and filtered time histories for every instrumented channel, the channel-to-transducer mapping, the filter settings, the computed criteria with their windows, and the pass/fail judgment against the governing limits. Retaining this provenance is what allows a result to be reviewed, reproduced, and defended — and it is the difference between a measurement and an assertion.