3. Occupant restraint system architecture

In a frontal crash, the restraint system is the only interface between the occupant's mass and the vehicle's deceleration pulse. Its job is deceptively simple: couple the occupant to the vehicle early, then manage the energy transfer so that no single body region absorbs more than it can tolerate. HIC_Nij (head) and femur force (leg) are the two metrics most sensitive to how that energy management is tuned, and they respond to restraint design in opposite directions. Understanding the architecture — and the trade-offs baked into it — is a prerequisite for any multi-objective optimization study.

Architectural Layers and Their Functions

A modern frontal restraint system is a layered stack, not a single device. Each layer has a distinct job in the timeline of a crash:

• Seat belt webbing and geometry — the primary load path. The lap belt anchors the pelvis; the shoulder belt routes across the chest and clavicle. Geometry (D-ring position, anchor points, belt angle) determines how load is distributed and whether the occupant "submarines" or loads the belt improperly.
• Retractor pretensioner — fires within the first ~10–20 ms, removing slack (typically 40–80 mm of webbing) so the occupant engages the belt before the deceleration peak. Earlier coupling means the occupant rides the pulse longer, reducing peak acceleration but increasing the risk of high belt force.
• Buckle/anchor pretensioner — pulls the buckle down and the outboard anchor inward, improving pelvis coupling and reducing lap-belt ride-up.
• Load limiter — a force-limiting element in the retractor or anchor (torsion bar, tear stitch, or pyrotechnic switchable limiter) that caps shoulder-belt force, typically in the 3–6 kN range for a 50th-percentile male. This is the single most influential parameter for chest and head metrics.
• Airbags — frontal driver and passenger airbags provide a compliant surface that decelerates the head over a longer distance, and venting controls the pressure-time history. Side and knee airbags address lateral and lower-extremity loading respectively.

How Each Element Moves HIC_Nij and Femur Force

The coupling between restraint parameters and injury metrics is not monotonic, which is exactly why multi-objective optimization is needed.

Pretensioner timing and stroke. Firing earlier removes slack sooner, so the occupant's head and chest begin decelerating with the vehicle rather than after free travel. This generally reduces HIC because the head's relative velocity at airbag contact is lower. However, if the pretensioner fires too early relative to the airbag deployment, the occupant may be pulled forward into the deploying airbag — a well-known out-of-position risk. In optimization terms, pretensioner fire time is a constrained variable: it must sit inside a window defined by airbag deployment timing.

Load limiter force level. Lowering the limit force reduces shoulder-belt force and chest deflection, which helps chest criteria and, indirectly, head metrics by allowing more forward excursion (more ride-down distance). But excessive forward excursion increases the head's travel toward the airbag and the steering wheel, and can raise femur loading if the occupant slides forward into the knee bolster or instrument panel. Raising the limit force does the reverse: better femur protection via tighter pelvis coupling, but higher chest and head loading. This is the classic HIC-versus-femur trade-off that dominates frontal restraint tuning.

Airbag venting and pressure. A larger vent or lower inflator output produces a softer bag that absorbs head energy over a longer stroke, lowering HIC — but only if the occupant actually reaches the bag. If the load limiter allows too much excursion, the head may bottom out the bag against the steering wheel hub, producing a sharp HIC spike. Conversely, an over-stiff bag can produce high head acceleration even with good ride-down.

Knee airbags and lower-extremity loading. Femur force is driven by the knee's contact with the knee bolster and the pelvis's restraint by the lap belt. A knee airbag distributes load across the femur and tibia and reduces localized force, but it also pushes the knees rearward, which can alter pelvis position and lap-belt fit. The interaction is bidirectional and must be evaluated together, not in isolation.

Reading the Architecture as a Signal Chain

The figure below traces the crash pulse through the restraint stack to the two injury metrics, showing where each tunable parameter acts and which metric it primarily influences. Note that the load limiter sits on a shared path — it affects both head and femur outcomes, which is why it is the pivot point of the optimization.

Practical Tuning Example

Consider a 50th-percentile male driver in a 56 km/h full-frontal rigid barrier test. A baseline restraint configuration uses a retractor pretensioner firing at 12 ms, a single-stage load limiter at 4.5 kN, and a driver airbag with a 60 mm vent. Suppose the baseline results show HIC₁₅ = 620 and peak femur force = 4.8 kN — both within regulatory limits but with little margin on the femur side.

A naive fix — raising the load limiter to 5.5 kN to reduce femur force — would tighten pelvis coupling and lower femur load, but it also raises shoulder-belt force and chest deflection, and reduces forward excursion, which can push HIC upward. The correct approach is to treat the parameters jointly:

1. Hold the load limiter at 4.5 kN and instead increase pretensioner stroke by 15 mm. Earlier, deeper coupling reduces femur force by limiting pelvis forward travel into the knee bolster, with minimal HIC penalty because the airbag still provides ride-down.
2. If HIC rises above the target, enlarge the airbag vent from 60 mm to 70 mm to soften the bag and recover head margin.
3. Re-evaluate knee airbag pressure: a slightly softer knee bag reduces localized femur force without pushing the knees so far rearward that lap-belt fit degrades.

The observable verification criteria are straightforward: HIC₁₅ and Nij must remain below the target thresholds (e.g., HIC₁₅ ≤ 700, Nij ≤ 1.0), femur force must stay below the criterion limit (typically 10 kN for the 50th percentile, with a design margin), and the head must not contact the steering wheel hub or A-pillar. If the head excursion trace shows contact with a rigid structure before full airbag engagement, the load limiter is too soft or the pretensioner too late — a symptom that appears as a sharp HIC spike late in the pulse.

Why This Architecture Matters for Optimization

The restraint stack is a multi-input, multi-output system with coupled, non-monotonic responses. Pretensioner timing, load limiter force, airbag venting, and knee airbag pressure each influence both HIC_Nij and femur force, often in opposing directions. That coupling is precisely what makes the restraint design space a genuine multi-objective problem: there is no single setting that minimizes both metrics simultaneously, only a Pareto front of trade-off configurations. The next step in any optimization workflow is to map that front by sampling the parameter space — pretensioner fire time, limiter force, vent area, and knee bag pressure — and evaluating each candidate against both injury criteria under the same crash pulse.