Document Istanbul drift-tuning validation in README and public log.

Record the post-tuning demo and public regression gate results so release
checks and smoke numbers stay aligned with the new Istanbul preset.

Co-authored-by: Cursor <cursoragent@cursor.com>

Author: rsasaki0109

CHANGELOG.md

@@ -18,6 +18,7 @@
 - add `docs/pose_covariance.md` for `/pcl_pose` covariance semantics (issue #72)
 - document Istanbul public demo RMSE run variance; identical seed/map can yield ~`1.2 m` or late-run drift outliers
 - tune public Istanbul 60 s preset with `local_map_crop` and `recovery_retry_from_last_pose` (`r3`, `gap<=1 s`, `seed<=15 m`)
+- reconfirm public regression suite after drift tuning; Istanbul gate `1.00 m` and HDL IMU checks pass (2026-06-10)
 - guard `use_odom` integration until initial pose is valid and keep pose finite for NDT init guess
 - ignore non-finite `/initialpose` payloads and stop eager scan replay during pose reset
 - v1.1 relocalization is documented as a validated dry-run `/initialpose` command artifact only

README.md

@@ -61,12 +61,13 @@ What it does:
 Typical first-run time: about **15–30 minutes** (network download dominates).
 Re-runs with `--resume` are much faster.
 
-Latest local smoke on Autoware Istanbul `60 s` (dataset already present, 2026-06-10):
+Latest local smoke on Autoware Istanbul `60 s` (dataset already present, 2026-06-10,
+new preset with `local_map_crop` + `recovery_retry_from_last_pose`):
 
-- translation RMSE: `1.21 m` (fresh rerun; outlier run `4.74 m` on same seed/map)
-- rotation RMSE: `0.40 deg`
-- matched samples: `61`
-- wall time: about `70 s`
+- public demo best run: translation RMSE `0.95 m`, rotation `2.13 deg`, matched `220`
+- public regression gate: translation RMSE `1.00 m`, rotation `2.33 deg`, matched `104` — **pass**
+- outlier runs still happen on identical seed/map (e.g. `3.31 m`, `19` matched); see variance notes
+- wall time: about `70 s` per demo run
 
 Istanbul replay is run-to-run variable; see [docs/benchmarking.md](docs/benchmarking.md)
 for the variance notes and prefer `scripts/run_public_regression_suite.sh` for release gates.

docs/public_validation_log.md

@@ -3,6 +3,59 @@
 This file records reproducible public-data validation snapshots. It is not a
 hardware field-test log.
 
+## 2026-06-10 Main After Istanbul Drift Tuning
+
+- commit: `f6503d3`
+- source branch: `main`
+- validation commands:
+
+```bash
+source scripts/setup_local_env.sh
+scripts/run_public_demo.sh \
+  --output-dir /tmp/lidarloc_public_demo_post_tune_r2 \
+  --skip-fetch --skip-build --ros-domain-id 212
+scripts/run_public_regression_suite.sh \
+  --output-dir /tmp/lidarloc_public_regression_post_tune \
+  --ros-domain-base 220
+```
+
+Result:
+
+- public demo (best of two fresh runs): `pass` trajectory eval
+  - translation RMSE: `0.955 m`
+  - rotation RMSE: `2.130 deg`
+  - matched samples: `220`
+  - pose rows: `222`
+- public demo outlier on same preset (same day): translation RMSE `3.308 m`, matched `19`
+- public regression suite: `pass`
+  - Istanbul `60 s`: `pass`
+    - pose rows: `106`
+    - matched samples: `104`
+    - translation RMSE: `1.002 m`
+    - rotation RMSE: `2.328 deg`
+    - recovery retry recovered: `1`
+    - IMU prediction active rows: `0`
+  - HDL `60 s`: `pass`
+    - runs: `2 baseline / 2 candidate`
+    - pose rows median: `559.0 -> 560.0`
+    - alignment-time median: `0.045634 s -> 0.046194 s`
+    - IMU prediction active rows median for the IMU-enabled candidate: `6.0`
+    - max IMU disable fallback events: `1`
+
+Artifacts:
+
+- `artifacts/public/demo/latest/demo_report.md`
+- `artifacts/public/public_regression_suite/summary.json`
+- `artifacts/public/public_regression_suite/summary.md`
+- `/tmp/lidarloc_public_demo_post_tune_r2/run/`
+- `/tmp/lidarloc_public_regression_post_tune/`
+
+Interpretation:
+
+- Drift-tuned Istanbul preset clears the public regression gate while still showing
+  run-to-run variance; prefer the regression suite over single demo runs for release checks.
+- HDL IMU safety / throughput behavior remains stable after the Istanbul preset change.
+
 ## 2026-05-22 Main Release Regression
 
 - commit: `2a5f11f`