4. Anthropomorphic Test Device (ATD) considerations

Submarining is a kinematics-driven failure mode: the occupant's pelvis slides forward and downward beneath the lap belt, allowing the belt to load the abdomen instead of the pelvic bone. Because the ATD is the measurement instrument through which we judge whether a restraint system prevents this behavior, the fidelity of the ATD's pelvis and the placement of its sensors directly determine whether a test can detect submarining at all. A restraint design that looks acceptable on a poorly instrumented dummy may fail catastrophically on a properly instrumented one — and vice versa. This chapter examines the three ATD-related levers that govern submarining detection: pelvis mechanical design, sensor placement and channel selection, and the metrics used to declare that submarining has occurred.

Pelvis Design and Its Influence on Kinematics

The pelvis is the load path through which the lap belt transfers restraint force to the occupant. In the Hybrid III family, the pelvis is a rigid, vinyl-clad aluminum structure with a molded flesh form; its geometry and surface friction characteristics strongly influence whether the lap belt stays on the iliac crest or slips off. The Hybrid III 50th percentile male pelvis is comparatively stiff and its flesh form is relatively smooth, which tends to promote belt migration in aggressive frontal pulses. The THOR (Test Device for Human Occupant Restraint) family addresses this with a more anatomically representative pelvis that includes a compliant iliac wing region and improved flesh geometry, producing belt engagement behavior closer to that of a human.

Two design details matter most for submarining work. First, the pelvis flesh stiffness and contour determine how deeply the lap belt embeds before it develops enough normal force to restrain the pelvis. A softer, more contoured flesh form allows earlier belt engagement and reduces the tendency to slide under. Second, the pelvis-to-spine interface — the lumbar spine attachment and its range of motion — controls how much the pelvis can rotate relative to the torso. Excessive pelvis rotation (retroversion) is a precursor to submarining because it opens the gap between the lap belt and the anterior superior iliac spine (ASIS). When selecting an ATD for a submarining study, document the pelvis part number and flesh durometer, because results are not directly transferable between Hybrid III and THOR, or even between different Hybrid III build configurations.

Seating also interacts with pelvis design. The pelvis must be positioned so the lap belt crosses the ASIS rather than the soft abdomen. In practice, this means setting the H-point per the manufacturer's procedure, verifying the pelvis angle with an inclinometer, and confirming belt routing with a physical check before the test. A pelvis that is rotated too far forward at setup will begin the event already predisposed to submarining.

Sensor Placement and Channel Selection

Detecting submarining requires instrumentation that can resolve pelvis motion relative to the belt and relative to the vehicle. The following channels form a practical minimum set for a frontal sled or crash test focused on submarining:

  • Pelvis accelerometers (triaxial): mounted on the pelvis structure, typically at the sacrum or iliac crest, to capture pelvis acceleration in X (fore-aft), Y (lateral), and Z (vertical). The vertical and fore-aft components reveal the downward-forward slide characteristic of submarining.
  • Pelvis rotation rate sensors (angular rate about Y): to capture retroversion, the pelvis rotation that precedes belt migration.
  • Lap belt load cells: placed in-line with the lap belt webbing on both sides to measure belt tension. A sudden drop in lap belt tension during the pulse, combined with pelvis displacement, is a strong submarining indicator.
  • Belt displacement or string potentiometers: routed to measure the relative motion between the lap belt and the pelvis, or between the pelvis and the seat cushion.
  • Abdomen pressure sensors (where available): to correlate belt loading with abdominal intrusion.

Placement precision matters as much as channel selection. Pelvis accelerometers should be rigidly mounted to the pelvis structure, not to the flesh form, so that they measure skeletal motion rather than soft-tissue deformation. The mounting location should be documented in the test report, because a sensor at the sacrum and one at the ASIS will produce different signatures for the same event. Lap belt load cells must be oriented so that the sensing axis aligns with the belt path; a misaligned load cell will under-report tension and mask the tension drop that signals submarining.

Before every submarining test, perform a static belt-routing check: with the ATD seated and belted, confirm that the lap belt passes over the ASIS and that the load cell axis is aligned with the webbing. Photograph the routing. This single step catches the most common cause of false-negative results.

Submarining Detection Metrics

No single channel declares submarining; the determination is made from the combination of pelvis kinematics, belt tension behavior, and belt-to-pelvis relative displacement. The most widely used detection logic rests on three observable signatures:

  1. Pelvis forward-downward displacement: the pelvis translates forward and downward relative to the seat and belt beyond a threshold established from non-submarining baseline tests. This is the primary kinematic signature.
  2. Lap belt tension drop: a rapid decrease in lap belt tension during the loading phase, indicating the belt has slipped off the pelvis and is no longer restraining it.
  3. Pelvis retroversion: angular rate about the Y axis exceeding a baseline threshold, indicating the pelvis has rotated into a position that permits belt migration.

In practice, engineers define a submarining detection criterion by first running a baseline test with a known-good restraint (no submarining) and recording the peak pelvis displacement, minimum lap belt tension, and peak pelvis rotation rate. The detection threshold is then set as a margin beyond those baseline values — for example, pelvis forward displacement exceeding the baseline peak by a defined percentage, or lap belt tension falling below a defined fraction of its peak during the loading window. These thresholds are test-program-specific and must be documented; they are not universal constants.

The figure below summarizes how the ATD's pelvis design, sensor placement, and the three detection signatures combine into a single determination logic.

ATD Pelvis Hybrid III / THOR Rigid pelvis structure + flesh form ASIS Lap belt path Triaxial pelvis accel (sacrum) Pelvis angular rate (Y) Lap belt load cells (L/R) Belt-pelvis displacement Detection Signatures 1. Pelvis displacement Forward + downward beyond baseline threshold 2. Lap belt tension drop Rapid decrease during loading phase 3. Pelvis retroversion Angular rate about Y exceeds baseline Combination, not single channel Determination Baseline test Known-good restraint Threshold set Margin beyond baseline Submarining declared Signatures exceed threshold Iterate: adjust pelvis setup / sensor placement

Practical Workflow and Common Pitfalls

A reproducible submarining evaluation follows a consistent sequence. First, select and document the ATD configuration — pelvis part number, flesh durometer, and any build-specific notes. Second, set the H-point and pelvis angle per procedure, then perform the static belt-routing check and photograph it. Third, verify sensor mounting: pelvis accelerometers rigid to the structure, load cells aligned with the belt path, displacement sensors zeroed at the belted position. Fourth, run the baseline (known-good) test and record peak pelvis displacement, minimum lap belt tension during loading, and peak pelvis rotation rate. Fifth, run the candidate configuration and compare against the baseline-derived thresholds.

Common failure symptoms and their causes are worth memorizing. If the lap belt tension trace shows no drop but the pelvis displacement is large, suspect a load cell misalignment or a sensor mounted on the flesh form rather than the structure. If pelvis rotation rate is flat while displacement is large, suspect a missing or saturated angular rate channel. If the baseline test itself shows a tension drop, the "known-good" restraint may not be good — re-verify belt routing before trusting the threshold. If results differ sharply between two ATDs of the same nominal type, compare pelvis part numbers and flesh durometers; build variation is a real and frequently overlooked source of inconsistency.

Do not set submarining thresholds from a single baseline test. Run at least two baseline repetitions and use the more conservative (larger displacement, lower tension) values to define the margin. A threshold derived from one noisy run will produce false positives or false negatives that are difficult to diagnose later.

The ATD is not a passive observer in submarining analysis — its pelvis design shapes the kinematics, its sensor placement determines what can be seen, and its metrics define what counts as failure. Treating these three elements as a coordinated instrumentation strategy, rather than as independent setup details, is what makes submarining detection repeatable and trustworthy across a test program.

← Previous 📖 Contents Next →