4. Data acquisition for occupant injury analysis

Occupant injury analysis is only as trustworthy as the measurement chain that feeds it. A physical injury criterion such as HIC, chest deflection, or femur axial force is ultimately a number derived from a sensor voltage sampled at a specific instant, filtered through a defined channel-frequency class, and synchronized with the crash event. This chapter examines the four primary data sources used in crash testing — accelerometers, load cells, high-speed video, and post-test physical measurement — and how they are acquired, conditioned, and validated so that the resulting injury metrics are defensible.

The Measurement Chain and Its Reference Conventions

Every channel in a crash test follows the same path: a sensing element converts a mechanical quantity into an electrical signal, a signal conditioner amplifies and filters it, a data acquisition system (DAS) digitizes it, and the recorded time history is processed into injury metrics. Before any of this, the test must establish a coordinate convention. SAE J211/1 defines the sign convention used throughout occupant testing: for a vehicle or sled, the +X axis points forward (direction of travel), +Y points to the left, and +Z points upward, with rotations positive according to the right-hand rule. All accelerometer polarity, load cell orientation, and displacement signs in this chapter follow this convention unless stated otherwise. A sensor mounted backwards will produce a plausible-looking but sign-inverted trace, and injury criteria that depend on direction (for example, distinguishing tension from compression in a neck load cell) will be silently wrong.

Accelerometers: Selection, Mounting, and Filtering

Accelerometers dominate occupant data acquisition because most injury criteria are acceleration-based. Three practical classes appear in a typical test:

• Vehicle/sled accelerometers — measure the crash pulse applied to the structure; typically ±500 g to ±2000 g range, mounted rigidly to the sled or vehicle body.
• Dummy accelerometers — head, chest, and pelvis units inside the ATD (anthropomorphic test device); head accelerometers are commonly triaxial units in the ±500 g class, chest units in the ±200 g class.
• Angular rate sensors — gyroscopes used to compute head angular velocity and displacement for rotational injury criteria such as BrIC.

Two parameters matter more than any other for data quality. First, range and sensitivity: an accelerometer that saturates (clips) during the pulse produces a flat-topped trace, and any HIC computed from it is an underestimate. Second, mounting stiffness: a sensor that resonates on a compliant bracket will show mounting-induced oscillation that contaminates the signal. Dummy accelerometers are mounted to the ATD structure per the dummy manufacturer's specification, and the mounting surface must be flat and torqued to spec.

Filtering is not optional. SAE J211 defines channel frequency classes (CFC) — CFC 60, CFC 180, CFC 600, CFC 1000 — each a fourth-order Butterworth low-pass with a specified cutoff. Head acceleration for HIC uses CFC 1000; chest acceleration for the 3 ms clip criterion uses CFC 180; neck and femur loads use CFC 600 or CFC 1000 depending on the criterion. Applying the wrong CFC changes the peak value and therefore the injury number. A common workflow check is to confirm the filter class in the post-processing configuration against the criterion definition before reporting any metric.

Load Cells: Force Measurement at Interfaces

Where acceleration describes motion, load cells describe the forces transmitted through interfaces. In occupant testing they appear as:

• Seat belt load cells — in-line units in the shoulder and lap belt webbing, measuring belt tension (commonly 0–20 kN range).
• Neck load cells — six-axis units at the head/neck interface of the ATD, resolving axial force, shear, and bending moment.
• Femur and tibia load cells — axial force transducers in the dummy legs.
• Steering column and knee bolster load cells — measuring the reaction force at the occupant contact surface.

Load cells are calibrated with known dead weights or a proving ring, and their zero must be captured immediately before the test with the dummy in its final seated position — belt tension preload, for example, will offset the shoulder belt channel. Six-axis neck cells require a defined local coordinate system at the occipital condyle; the sign of the axial channel determines whether the reported value is tension or compression, and neck injury criteria treat these asymmetrically. As with accelerometers, load channels are filtered to the CFC specified by the relevant criterion (CFC 600 for neck axial force and moment in most protocols).

High-Speed Video: The Independent Check

High-speed cameras typically run at 1000 frames per second (sometimes 2000 fps for close-up occupant kinematics), giving a 1 ms frame interval. Video serves three roles: it validates sensor data (a head trace that shows motion the accelerometer did not record indicates a mounting or channel fault), it captures events sensors cannot measure (belt slip, dummy contact with structure, airbag interaction), and it supports photogrammetric tracking of targets on the dummy for kinematic reconstruction. Camera synchronization to the DAS time zero is essential — a common practice is a flash or LED event marker recorded simultaneously on video and as a DAS trigger channel, so the two data streams share a common t=0. Without this, correlating a video frame to an acceleration peak is guesswork.

Post-Test Measurement

After the test, physical inspection adds information that no sensor captures: dummy damage and deformation, belt webbing abrasion and payout marks, contact witness marks on interior surfaces, and residual deformation of the restraint system. These observations are recorded on a post-test inspection sheet and cross-referenced with the time histories. For example, a shoulder belt load spike combined with a webbing abrasion mark at a specific payout length corroborates that the belt loaded at that instant. Post-test measurement is also where the dummy itself is checked for damage that could invalidate the run — a bent neck load cell or a cracked pelvis insert means the recorded loads may not reflect a valid ATD response.