Evidence status
- Implementation & independent numerical reference
- Verified — recorded regression evidence
- Published literature reproduction
- Not currently available
- Physical test correlation
- Not performed
A dedicated external published Full-Car literature reproduction has not yet been added to the current evidence set and is not currently claimed.
Recorded verification describes checks backed by the canonical regression suites and reviewed release evidence. It is distinct from the computed PASS/FAIL plots. Acceptance criteria below come from the existing registry or unchanged canonical test assertions.
Canonical model verification
FC-001exact zero equilibrium on four flat roadsVerified
- Category
- Analytical invariant
- Reference type
- Analytical zero equilibrium
- Acceptance criteria
- Every dynamic channel is exactly zero on flat roads with zero initial state.
- Test provenance
tests/validation/full-car.validation.test.ts
FC-002exact zero roll and identical corner evidence under left/right symmetryVerified
- Category
- Symmetry
- Reference type
- Symmetric excitation
- Acceptance criteria
- Roll angle, rate and acceleration are exactly zero; left/right axle channels are exactly equal.
- Test provenance
tests/validation/full-car.validation.test.ts
FC-003collapses to unchanged trusted Half-Car across all common physical channelsVerified
- Category
- Model reduction
- Reference type
- Unchanged trusted Half-Car
- Acceptance criteria
- Every mapped common channel differs by < 1e-13; symmetric track/inertia changes leave the complete response exactly equal.
- Test provenance
tests/validation/full-car.validation.test.ts
FC-004agrees with independent state-space free/forced references, including all states and a final partial stepVerified
- Category
- Independent reference
- Reference type
- Independent energy-assembled M/K/C state space
- Acceptance criteria
- Each free/forced state error is strictly below its ordered bound: [1e-9, 1e-8, 1e-9, 1e-8, 1e-9, 1e-8, 5e-9, 2e-7, 5e-9, 2e-7, 5e-9, 2e-7, 5e-9, 2e-7,]. Position bounds use m or rad; velocity bounds use m/s or rad/s. State order interleaves position and velocity. Independent stiffness/damping matrices are symmetric.
- Test provenance
tests/validation/full-car.validation.test.ts
FC-005converges at fourth order on the documented smooth free-response caseVerified
- Category
- Numerical convergence
- Reference type
- Independent matrix-exponential reference
- Acceptance criteria
- Successive time-step refinement error ratios are > 10 and < 22. The measured channel is front-left wheel velocity (m/s).
- Test provenance
tests/validation/full-car.validation.test.ts
FC-006all four upward wheel offsets have hand-calculated heave/pitch/roll signsVerified
- Category
- Equation/sign check
- Reference type
- Hand-calculated force and moment
- Acceptance criteria
- Each of four upward wheel offsets gives exactly the hand-calculated heave, pitch, roll and wheel accelerations.
- Test provenance
tests/validation/full-car.validation.test.ts
FC-007preserves request ids, exact direct/adapter parity and all 34 transferred buffers with detachmentVerified
- Category
- Worker integration
- Reference type
- Direct execution and transferred evidence
- Acceptance criteria
- Direct and adapter records are exactly equal; all 34 distinct buffers transfer and detach; request identity is preserved.
- Test provenance
tests/validation/full-car.validation.test.ts
FC-008rejects unsupported identities, malformed inputs and cross-model/schema pairings without fallbackVerified
- Category
- Contract rejection
- Reference type
- Strict model/schema/solver contract
- Acceptance criteria
- Unsupported identities and cross-model definitions throw; the Worker returns the original request ID with SIMULATION_ERROR. No fallback model is accepted.
- Test provenance
tests/validation/full-car.validation.test.ts
The independent implementation assembles mass, stiffness and damping matrices separately from production physics, compares all fourteen states on free and forced cases with four independent roads, and includes a final partial step. The smooth free-response refinement tests fourth-order convergence; reduction checks every common physical channel against Half-Car.
Existing independent-reference review record
Recorded numerical evidence
These values come from the reviewed Full-Car physics measurement record. They describe the canonical free/forced and reduction cases, with acceptance criteria shown above. They are not a new literature benchmark or a physical measurement.
FC-004 · All fourteen state errors
| State | Unit | Free maximum error | Forced maximum error |
|---|---|---|---|
| body heave displacement | m | 7.5056e-12 | 7.9465e-12 |
| body heave velocity | m/s | 4.9980e-10 | 5.7572e-10 |
| body pitch angle | rad | 2.1760e-11 | 4.2364e-12 |
| body pitch rate | rad/s | 1.3596e-9 | 3.4375e-10 |
| body roll angle | rad | 2.3944e-12 | 9.8895e-12 |
| body roll rate | rad/s | 1.4644e-10 | 6.6862e-10 |
| front left unsprung displacement | m | 9.7805e-10 | 2.6160e-10 |
| front left unsprung velocity | m/s | 5.6337e-8 | 1.3862e-8 |
| front right unsprung displacement | m | 1.4528e-10 | 2.7120e-10 |
| front right unsprung velocity | m/s | 8.4351e-9 | 1.4721e-8 |
| rear left unsprung displacement | m | 6.5880e-10 | 2.5087e-10 |
| rear left unsprung velocity | m/s | 3.7150e-8 | 1.3119e-8 |
| rear right unsprung displacement | m | 1.2160e-10 | 1.3548e-10 |
| rear right unsprung velocity | m/s | 7.9873e-9 | 6.8320e-9 |
FC-003 · Full-Car → Half-Car reduction errors
The reviewed canonical case has exact zero difference in every mapped channel, within the unchanged strict acceptance bound.
| Mapped channel | Maximum absolute difference |
|---|---|
| time | 0 |
| front left road displacement | 0 |
| front right road displacement | 0 |
| rear left road displacement | 0 |
| rear right road displacement | 0 |
| body heave displacement | 0 |
| body heave velocity | 0 |
| body heave acceleration | 0 |
| body pitch angle | 0 |
| body pitch rate | 0 |
| body pitch acceleration | 0 |
| body roll angle | 0 |
| body roll rate | 0 |
| body roll acceleration | 0 |
| front left body attachment displacement | 0 |
| front right body attachment displacement | 0 |
| rear left body attachment displacement | 0 |
| rear right body attachment displacement | 0 |
| front left unsprung displacement | 0 |
| front right unsprung displacement | 0 |
| rear left unsprung displacement | 0 |
| rear right unsprung displacement | 0 |
| front left unsprung velocity | 0 |
| front right unsprung velocity | 0 |
| rear left unsprung velocity | 0 |
| rear right unsprung velocity | 0 |
| front left suspension travel | 0 |
| front right suspension travel | 0 |
| rear left suspension travel | 0 |
| rear right suspension travel | 0 |
| front left dynamic tire deflection | 0 |
| front right dynamic tire deflection | 0 |
| rear left dynamic tire deflection | 0 |
| rear right dynamic tire deflection | 0 |
FC-005 · Smooth free-response convergence
| Time step (s) | Front-left wheel velocity error (m/s) |
|---|---|
| 0.004 | 1.4951e-5 |
| 0.002 | 9.1140e-7 |
| 0.001 | 5.6337e-8 |
Successive refinement ratios: 16.4046, 16.1776. This is the specified smooth free-response check for fourth-order RK4 convergence.
Review provenance: docs/review/full-car-physics-measurements.json; implementation milestone 1.3.0-alpha.1. Display rounding does not alter recorded arrays or acceptance.
Evidence boundary
These checks verify the documented passive linear mathematical model and software path. Generic presets have no measured production-vehicle correlation. Review model and evidence boundaries.
Read the model equations and assumptions →