2. Excursion metrics and scoring framework

Far-side occupant protection is judged by how far critical body regions travel and how that travel is converted into a defensible score. The scoring framework exists so that a physical trajectory measured in a test can be compared against a regulatory or consumer-rating threshold, and so that different test configurations, dummy types, and impact severities can be ranked on a common scale. This section defines the three core elements — head excursion, the excursion index, and the pass/fail criteria — and shows how they connect in practice.

Head excursion: the primary measured quantity

Head excursion is the maximum displacement of a defined head reference point relative to a fixed vehicle coordinate system during the impact event. The reference point is normally the head center of gravity (CG) of the anthropomorphic test device (ATD), tracked from the pre-impact position. Two conventions must be fixed before any number is meaningful:

• Coordinate frame. Excursion is reported in the vehicle's global X (longitudinal), Y (lateral), and Z (vertical) axes, with the origin at a documented datum. Far-side excursion is dominated by lateral (+Y or −Y depending on impact side) and vertical components, so the sign convention must be stated explicitly and held constant across all tests and figures.
• Time window. Excursion is evaluated over the full event, typically 0–300 ms, and the reported value is the peak magnitude, not the value at a single instant. Peak timing matters because a head that reaches maximum lateral travel late in rebound may behave differently from one that peaks during the primary impact.

In practice, excursion is derived from filtered ATD instrumentation — head CG accelerometers integrated twice, or optical/motion-capture tracking of targets on the head — and cross-checked against high-speed video. A typical far-side sled or full-vehicle test might report a peak lateral head excursion of, for example, 320 mm at 118 ms; that figure is only comparable to another test if the datum, filter class (e.g., CFC 60), and time window match.

The excursion index: normalizing travel against a limit

Raw excursion in millimeters is intuitive but not directly scorable across configurations. The excursion index (EI) normalizes the measured peak against the applicable limit for that configuration:

EI = X_peak / X_limit

where:
  X_peak  = peak head excursion magnitude (mm), filtered, over 0–300 ms
  X_limit = applicable excursion limit for the dummy, seating position,
            and impact severity (mm)

Interpretation:
  EI < 1.0  → within limit
  EI = 1.0  → exactly at limit (marginal)
  EI > 1.0  → limit exceeded

The index is dimensionless, so a 300 mm excursion against a 350 mm limit (EI = 0.86) and a 260 mm excursion against a 300 mm limit (EI = 0.87) are treated as roughly equivalent performance, even though the absolute travel differs. This is what allows a test matrix spanning different dummy sizes and impact severities to be summarized on one scale. The limit itself is configuration-specific: it depends on the ATD (e.g., THOR vs. Hybrid III), the restraint state (belted, unbelted, with or without a far-side airbag), and the pulse severity. Using the wrong limit silently invalidates the index, so the limit source must be recorded alongside every EI value.

Pass/fail criteria and the scoring decision

Pass/fail is not a single number comparison; it is a decision rule applied to the index together with supporting evidence. A defensible rule set has three layers:

1. Primary criterion. EI ≤ 1.0 for the head excursion metric constitutes a pass on that metric. Values above 1.0 fail.
2. Marginal band. An EI within a defined tolerance of 1.0 (commonly 0.95–1.00) is classified as marginal and requires review of video, sensor health, and repeatability before a final call.
3. Evidence validity gate. The result is only scoreable if the test itself is valid: sensors within calibration, no channel clipping, dummy positioning within tolerance, and the pulse within the specified corridor. An invalid test cannot produce a pass regardless of EI.

This layered structure prevents two common errors: declaring a pass from a single favorable number while ignoring a clipped accelerometer, and failing a design because one marginal run fell just over the line without checking run-to-run variation.

Worked example: from trajectory to decision

Consider a far-side oblique sled test with a belted THOR ATD, impact side +Y, and an applicable lateral head excursion limit of 350 mm. The measured peak lateral head CG excursion is 332 mm at 121 ms, with a vertical component of 48 mm. The resultant peak is √(332² + 48²) ≈ 335 mm. The excursion index is 335 / 350 = 0.96. That falls inside the marginal band (0.95–1.00), so the run is not an automatic pass. Review shows clean sensor channels, dummy positioning within ±5 mm, and a pulse inside the corridor; a repeat run gives 341 mm (EI = 0.97). Both runs are consistent and marginal, so the configuration is reported as marginal-pass with a recommendation to increase margin — for example, by adjusting the far-side airbag vent or belt geometry — rather than as a clean pass.

Reporting the score with its evidence

A score without its evidence chain is not auditable. Each reported result should carry the peak excursion with time of peak, the applicable limit and its source, the resulting EI, the decision band, and the validity status. The table below shows a compact reporting format for the worked example and a contrasting clean-pass case.

The decision uses the worst valid run (EI = 0.97, marginal), while the tuned configuration demonstrates that the margin can be recovered. This is the practical value of the framework: it converts a trajectory into a normalized, comparable, and auditable score, and it forces the engineer to state the limit, the convention, and the validity conditions that make the number trustworthy.