Hybrid III has been the default frontal crash dummy since its introduction in the 1970s. Half a century of use has made it the common language of occupant protection. Regulators wrote procedures around it. Labs built decades of comparable results on it. Engineers can read its traces almost by instinct.

Its instrumentation is focused and well understood: head acceleration, neck forces and moments, chest acceleration and deflection at the sternum, plus femur and tibia loads. Those channels feed the injury criteria that most passive safety programs still report against.

THOR stands for Test device for Human Occupant Restraint. It matches Hybrid III in size. Its developers had a specific research goal: a dummy that could reveal injury mechanisms older designs struggle to capture. THOR carries expanded instrumentation across the head, neck, thorax, abdomen, pelvis, femurs, lower legs and feet.

The practical effect is a much richer dataset from every test. A restraint engineer can see where the belt concentrates load on the ribcage and how the pelvis moves under the lap belt, along with the headline injury values.

Regulators have taken notice. NHTSA and Euro NCAP are specifying the THOR-50M for frontal crash tests in place of the Hybrid III 50M, although its path into consumer safety ratings has been gradual. For labs, that shift lands alongside the other Euro NCAP 2026 protocol changes already stretching sled calendars and dummy fleets.

Exact channels vary by build level and protocol. Always work from the applicable specification.

Head and Neck

A real occupant's head translates and rotates at the same time while the neck takes tension, shear and bending together. Hybrid III measures head acceleration and neck loads well enough to have anchored decades of analysis. THOR's neck records more of that combined loading, which matters when engineers are tuning how the head meets the airbag.

Thorax

Hybrid III reads chest deflection at a single point on the sternum. THOR reads it at several points across the ribcage, so an off-centre belt load shows up clearly in the data. Studies of the smaller female versions make the point well. The THOR-05F recorded much larger deflection on the buckle side of the chest than the central Hybrid III measurement. A single central reading would have missed that asymmetry.

Abdomen and Pelvis

The lap belt is supposed to hold the pelvis. When it rides up onto the abdomen, the occupant can slide beneath it, a failure known as submarining. Hybrid III represents this region loosely. THOR adds an instrumented abdomen and a more human-like pelvis, giving engineers a clearer view of submarining risk. The difference surfaced in NHTSA's rear seat sled work, where the Hybrid III did not submarine in any test.

Biofidelity describes how closely a dummy responds like a human under the same loading. It varies by body region and by type of response, so no single score crowns a winner.

The research reflects that. One whole-body frontal sled study found Hybrid III had marginally better time history biofidelity and THOR had better injury prediction, with generally small differences between the 2 when all body regions were pooled. The Stapp study adds a caution for anyone reading THOR numbers at face value. Compared with CISS field data, THOR-50M overpredicted chest deflection injury risk by at least 4 times and BrIC-based risk by at least 10 times, and the authors call for further work on dummy construction, injury risk curves and seating procedures.

Put simply, THOR gives engineers more to work with. The injury risk curves used to interpret that data are still being refined.

Plenty of active protocols still call for Hybrid III. Its long record gives engineers a baseline for comparing a new vehicle with the ones before it. Technicians can set it up quickly and analysts read its traces with confidence. Many labs with a THOR in the fleet keep their Hybrid IIIs busy too.

The choice belongs in the test plan, settled well before anyone books a sled. 5 considerations usually decide it.

• Protocol. When a regulation or rating protocol names an ATD, the decision is already made.

• Purpose. Restraint development and injury research benefit most from THOR's wider set of channels.

• Comparability. A program built on Hybrid III baselines may need to stay on Hybrid III. Results from the 2 dummies cannot be compared one to one.

• Data load. Extra channels only help when they land in a central test data record with the dummy's configuration attached.

• Traceability. Auditors working to ISO/IEC 17025 expect every result to trace back to the exact dummy, build level, calibration state and test.

The crash itself is over in a fraction of a second. The data it produces can be referenced for years in design reviews, audits and recall investigations.

That data only holds up if its context does. Every result depends on a chain that runs from the dummy's serial number and build level through channel configuration, calibration status, test article, restraint setup and test conditions. A THOR carries many more sensors than a Hybrid III, each with its own calibration schedule, which is why experienced labs treat crash test dummies as precision instruments. When one link in that chain lives in a spreadsheet and another sits in an email thread, reconstructing a test months later becomes slow and uncertain work.

Most ATD problems start small. One THOR gets booked onto 2 sleds for the same afternoon. A rib module drifts past its calibration date the week before a rating test. Hit counts sit in one technician's notebook. Any of these can turn a clean test into a retest.

TITAN's ATD management module keeps each dummy's master data, configuration, calibration, bookings and usage in one place. It handles Hybrid III, THOR, CRABI, SID and child dummies, plus any custom ATD types a lab defines.

In day-to-day use that looks like this:

• Profiles hold master data, status flags, CMM locations and supporting documents for every dummy.
• Channels and scenarios define what gets measured and where, then map straight into test requests.
• Calibration runs as one-time or recurring work orders against each ATD, with certificates stored on the record.
• Bookings go through the lab calendar, which flags overlapping reservations before they reach the test floor.
• Barcode check-in and check-out lets technicians update status and attach setup or damage photos from the floor.
• Hit and task counts build a usage history that informs recalibration, servicing and retirement.
• Test request links tie each dummy to the tests it ran, so its configuration and calibration state become part of the test record.

Test organizations with strict audit demands, including a national type approval authority, already run their test operations on TITAN. For any lab running THOR alongside Hybrid III, a connected history is what makes the extra data worth trusting. The dummy is part of the test, so its record belongs in the test record.

This keeps a credibility signal and an internal link. The homologation case study is already anonymized, so it carries no disclosure risk.

1. Is THOR-50M replacing Hybrid III?

For some protocols it already has. NHTSA and Euro NCAP have specified THOR-50M for frontal tests. Hybrid III remains required in many other programs, so most labs will run both for years to come.

2. Is THOR more biofidelic than Hybrid III?

It depends on the body region and the response being studied. Research has found small differences overall, with THOR stronger on injury prediction in some conditions.

3. Can THOR and Hybrid III results be compared directly?

Rarely. The 2 dummies differ in construction, instrumentation and response. Any comparison has to account for the protocol, the measurement definition and the injury criterion.

4. Why does THOR produce more data?

Its additional instrumentation captures loading across more body regions, including the abdomen and several points on the chest.

5. Why does ATD management matter for crash testing?

Calibration status, configuration and usage history all affect whether a result is valid and traceable. Keeping them in one system ties every measurement to the dummy and the test that produced it.