-
Notifications
You must be signed in to change notification settings - Fork 28
Experiment Results
JokerJohn edited this page Apr 24, 2026
·
5 revisions
The default shifted_cylinder experiment was re-validated after the full DCReg release sync:
| Parameterization | Iter | RMSE | Fitness | Trans. Error (m) | Rot. Error (deg) | LinIt | QRfb |
|---|---|---|---|---|---|---|---|
| Euler | 9 | 0.0314988 | 0.116239 | 0.0258485 | 0.0516469 | 6 | 0 |
| SE3 | 10 | 0.0315708 | 0.116636 | 0.0271197 | 0.0507196 | 6 | 0 |
| SO3 | 10 | 0.0315708 | 0.116636 | 0.0271196 | 0.0507195 | 6 | 0 |
| Quaternion | 10 | 0.0315708 | 0.116636 | 0.0271196 | 0.0507195 | 6 | 0 |
dcreg_minimal_example exposes the three core modules:
- Module 1: spectral degeneracy detection
- Module 2: physical-axis characterization
- Module 3: preconditioned linear solve
This minimal example uses a synthetic linear system, so it does not print RMSE,
fitness, or runtime. Those registration metrics are reported by dcreg_runner
and dcreg_parking_lot_example.
The parking-lot example registers one LiDAR frame against a prior map from a real localization scene. The repository bundles the source frame and initial pose metadata; the large prior map remains external and is linked from the README.

Current exported diagnostics for the bundled pk01-1976 frame:
- Iterations: 5
- Final RMSE: 0.053225
- Final fitness: 0.060365
- Runtime: machine-dependent; about 1.8 ms on the local TBB validation run
- Degenerate mask:
000100, corresponding to a weak physicalxtranslation axis - Visualization: black background, intensity-colored target/source clouds,
100 mlocal target-map crop, compact degeneracy card, and highlighted weak axis
| Demo | Characterization |
|---|---|
![]() |
![]() |
| Controlled Simulation | Real-World Evaluation |
|---|---|
![]() |
![]() |
| Runtime | Parameter Study |
|---|---|
![]() |
![]() |





