feat: qualify complete K1 replay on worker

docs/13_LIDAR_WORKER_PRODUCT_AND_ROADMAP.md

@@ -5,7 +5,7 @@ Status: accepted architecture plan; L0/L1 implemented; L2 diagnostic A/B
 complete; full GOOSE and RELLIS qualification complete; L2.6d K1 replay
 local-surface temporal qualification, operator triage and prior-plane residual
 explainability implemented; L2.6e recorded-source-paced bounded shadow
-qualified; physical K1 shadow next
+qualified; E28 complete worker replay accepted; dynamic-observation layer next
 Scope: passively received real-time K1 point/pose evidence, immutable replay and
 future live shadow processing
 Explicitly out of scope: K1 firmware modification, a new onboard exporter, new
@@ -423,8 +423,10 @@ Dataset expansion is no longer the next gate.
 - [x] Replay the same profile through a bounded latest-wins shadow queue at the
   recorded 1× source rate; no K1 command, navigation or safety authority is
   added.
-- [ ] Bind that runtime to physical live K1 point/pose evidence through the
-  existing authenticated external-worker seam.
+- [x] Bind the complete raw replay to the normalizer, point/pose binder and
+  bounded estimator through the authenticated external-worker seam.
+- [ ] Commission the same seam against physical live K1 evidence only when
+  hardware/network behavior becomes the test subject.
 
 The implemented `missioncore.k1-local-surface/v1` derivative is reproducible
 through `experiments/perception/run_k1_local_surface.py` and is exposed
@@ -437,36 +439,40 @@ local-discontinuity candidates. The candidates are not semantic curb
 recognition or ground truth. The React contract is provider-neutral; K1 remains
 a bound backend source rather than UI implementation knowledge.
 
-The complete RAVNOVES00 run covered all `526/526` available LiDAR samples.
-There were zero stale-pose, insufficient-surface or failed-fit frames. Observed
-diagnostic distributions are: derived sensor-to-surface height `1.295 m` p50,
-roughness `0.054 m` p95, slope `3.172°` p95 and pose-binding age `19.113 ms`
-p95. These values describe this recording only; they are not calibration,
-ground truth or a navigation gate.
+The complete RAVNOVES00 pack contains `4,489` timeline frames and `9,207,270`
+points. `3,928` frames carry available LiDAR evidence and all `3,928` produced
+a valid result; `561` timeline frames are explicitly unavailable. There were
+zero stale-pose, insufficient-surface or failed-fit frames. Observed diagnostic
+distributions are: derived sensor-to-surface height `1.2816 m` p50, roughness
+`0.06736 m` p95, slope `4.1895°` p95 and pose-binding age `21.8545 ms` p95.
+These values describe this recording only; they are not calibration, ground
+truth or a navigation gate.
 
 The L2.6b qualification scores each available frame against a local plane built
 only from the preceding TTL window; the frame being scored is excluded from
-the prediction input. It produced `525` independent next-frame samples. The
-distribution of per-frame median absolute residual has `0.038543 m` p50 and
-`0.050398 m` p95. Prediction inlier fraction has `0.96354` p50 and `0.95133`
-mean. The temporal comparison detected `12/525` strict-threshold jumps; height
-delta is `0.022887 m` p95, slope delta `0.09572°` p95 and roughness delta
-`0.003317 m` p95.
+the prediction input. It produced `3,927` independent next-frame samples. The
+distribution of per-frame median absolute residual has `0.04065 m` p50 and
+`0.07304 m` p95. Prediction inlier fraction has `0.94643` p50 and `0.91937`
+mean. The temporal comparison detected `205/3,927` strict-threshold jumps;
+height delta is `0.02891 m` p95, slope delta `0.12410°` p95 and roughness
+delta `0.00520 m` p95.
 
 The central surface is therefore repeatable at roughly four-centimetre median
 error on this replay. The tail is not yet planner evidence: the distribution
-of per-frame p95 residual has `0.41496 m` p95 and reaches `1.083 m`. Step/curb
-candidates occur in every available frame (`68.5` cells and `83` points p50),
+of per-frame p95 residual has `0.66193 m` p95 and reaches `1.735 m`. Step/curb
+candidates occur in every available frame (`75` cells and `99` points p50),
 which proves that the review layer is active but also that it is broad and
 still requires independent review. Free space remains unavailable; none of
 these metrics grants navigation, command or safety authority.
 
-The `missioncore.k1-local-surface-review/v1` triage contract selects a frame
+The earlier content-bound `526`-frame review generation is retained for
+operator evidence and regression history. Its
+`missioncore.k1-local-surface-review/v1` triage contract selects a frame
 when its prediction p95 residual is at least `0.45 m`, prediction inlier
 fraction falls below `85%`, or one of the content-bound temporal thresholds is
 crossed. It groups observations separated by no more than two frames into one
 episode and ranks threshold exceedance without changing the source artifact.
-RAVNOVES00 produced `37` review frames in `21` episodes: `18` prediction-tail
+That reviewed slice produced `37` review frames in `21` episodes: `18` prediction-tail
 frames, `14` inlier drops and `12` height transitions. Four frames have a
 normalized attention score of at least `2.0`.
 
@@ -517,7 +523,7 @@ immutable replay derivative: zero state, point-class or step-candidate
 mismatches and zero scalar delta. The accepted result is
 `k1-local-surface-shadow-04f14d8c580f74cbd5b0a452867563ebc6b3ef93d872129e4918680932253ab7`.
 
-The authenticated external worker now contains the next bounded seam.
+The authenticated external worker now contains the accepted bounded seam.
 `missioncore.k1-local-surface-pose-binder/v1` accepts decoded map-frame
 `lio_pcl` and `lio_pose` in either arrival order, binds by bounded worker
 host-arrival time, and publishes only matched pairs to the capacity-two
@@ -527,18 +533,28 @@ counts every point frame; the final report records pose-binding misses,
 binder/runtime overflow, exact queue accounting, processing failures and p95
 result age.
 
-The pinned physical profile is
-`experiments/perception/worker/e28_physical_k1_local_surface_profile.json`.
-It requires at least `100` bound frames, at most `5%` pose misses, `1%` binder
-point replacement, `1%` runtime replacement and `80 ms` p95 worker-local
-result age. The existing D-only PowerShell worker launcher accepts this
-physical profile and verifies that the minimal hash-addressed package contains
-the local-surface geometry, binder/runtime and ground primitives.
+The pinned profile is
+`experiments/perception/worker/e28_worker_replay_local_surface_profile.json`.
+It separates a lossless depth-eight authenticated LiDAR ingress from the
+freshness-first capacity-two estimator. The acceptance gate requires at least
+`9 Hz`, no more than `100 ms` p95 result age, at most `8%` explicitly
+superseded estimator inputs, at most `5%` pose misses, at most `1%` binder
+replacement, exact queue accounting and shadow-only authority.
 
-This is implementation readiness, not a physical result. The next gate is to
-build and stage the new content-addressed worker package on D, execute the
-profile against an attached K1 and retain the generated report. Commands,
-free-space, navigation and safety authority remain disabled.
+LAB E28 executed the complete recording on `mission-gpu` through the normal
+authenticated wire. All `4,570` LiDAR and `4,598` pose events arrived with
+zero sequence gaps; all `4,570` point frames bound with no miss or binder
+replacement. The worker persisted `4,206` results, explicitly superseded
+`364/4,570` inputs (`7.965%`), delivered `9.1945 Hz`, and measured
+`89.530 ms` full p95 result age. Processing p95 was `45.470 ms`; failures were
+zero. The accepted result is
+`e28-local-surface-wire-e1b0902d502267aff5f0bf78365b24967f98c83d7c1e6bb0f9dc2934c0684ceb`.
+
+This is a worker replay result, not a physical acquisition result. The next
+gate is a separate dynamic-observation layer plus explicit staleness/deadline
+health. Physical K1 execution is deferred until hardware-clock, field-network
+or acquisition behavior is the test subject. Commands, free-space, navigation
+and safety authority remain disabled.
 
 Exit: one immutable K1 session yields both a persistent reconstruction and a
 bounded local world state without hard-coded terrain height or scanner-side
@@ -563,10 +579,12 @@ independent gate without increasing unsafe false-free or false-dynamic output.
 
 - [x] Add a provider-neutral bounded LiDAR local-surface queue independent of
   camera cadence and qualify it at recorded 1× source pace.
-- [x] Bind authenticated decoded physical `lio_pcl + lio_pose` to that queue
-  with bounded buffers, exact accounting and a pinned physical acceptance
+- [x] Bind authenticated decoded replay `lio_pcl + lio_pose` to that queue
+  with bounded buffers, exact accounting and a pinned worker acceptance
   profile.
-- [ ] Run the accepted K1 local-surface/local-map profile on the NVIDIA worker.
+- [x] Run the complete accepted K1 local-surface profile on the NVIDIA worker.
+- [ ] Commission physical K1 acquisition separately; do not block replay
+  development on device presence.
 - [ ] Fuse K1 geometric evidence with E26 camera evidence as independent
   sources; a LiDAR-native detector remains optional.
 - [ ] Publish `agree`, `single-source`, `conflict` and `unknown`; unknown remains
@@ -625,15 +643,18 @@ natural-ground recall from `51.76%` to `76.22%` and stays below `23.41 ms` p95
 across 961 frames. Full RELLIS qualification covers `2,413` validation frames
 and rejects Patchwork++ because the small Ground-IoU gain came with unacceptable
 obstacle loss. Public-dataset ground qualification is therefore complete
-enough for the current decision. The first K1 local-surface replay slice now
-covers all available `RAVNOVES00` samples and is visible in the operator
-interface. Leave-current-frame-out temporal qualification now covers `525`
-samples: the median surface error is stable, while the p95 tail remains too
-large for a free-space claim. Deterministic triage has reduced the first manual
-inspection set to four high-priority frames in two episodes. Prior-plane
-residual explainability now shows that episode `09` is a localized positive
-structure rather than symmetric plane drift. The highest-value immediate work
-is therefore the physical K1 run through the now-implemented bounded
-live-shadow seam, followed by a separate dynamic-observation layer. Nvblox,
+enough for the current decision. The complete K1 local-surface derivative now
+covers `4,489` timeline frames, including `3,928` available LiDAR frames, and
+is visible in the operator interface. Leave-current-frame-out temporal
+qualification covers `3,927` samples: the median surface error is stable,
+while the p95 tail remains too large for a free-space claim. The retained
+review generation reduces its first manual inspection set to four
+high-priority frames in two episodes. Prior-plane residual explainability
+shows that episode `09` is a localized positive structure rather than
+symmetric plane drift. The complete worker replay proves the bounded
+live-shadow seam at `9.1945 Hz` and `89.530 ms` p95 result age without a
+connected scanner. The highest-value immediate work is therefore the separate
+dynamic-observation and staleness-health layer. Physical K1 commissioning
+follows only when hardware behavior is in scope. Nvblox,
 raw-scan detectors and alternative SLAM remain optional later gates because the
 current report contract does not carry their required ray/timing semantics.

docs/14_LIDAR_DATASET_GATEWAY.md

@@ -49,11 +49,13 @@ Implemented now:
 - a content-addressed `missioncore.k1-local-surface/v1` replay derivative over
   immutable `RAVNOVES00`, with dynamic height/slope/roughness/confidence,
   explicit pose-binding age and conservative observed occupied/unknown policy;
-- leave-current-frame-out next-frame qualification over `525` samples,
-  temporal jump evidence and unverified local-discontinuity candidates;
+- a complete `4,489`-frame timeline derivative with `3,928` valid LiDAR
+  results, `9,207,270` points and leave-current-frame-out qualification over
+  `3,927` samples;
 - deterministic `missioncore.k1-local-surface-review/v1` replay triage:
-  `37` attention frames grouped into `21` episodes, with four
-  high-priority frames and direct source-frame navigation;
+  the retained `526`-frame review generation has `37` attention frames grouped
+  into `21` episodes, with four high-priority frames and direct source-frame
+  navigation;
 - point/cell-aligned prediction evidence in
   `missioncore.k1-local-surface-frame/v2`: the prior-only plane, current
   lower-cell coordinates, signed residual and derived inlier mask are
@@ -61,6 +63,9 @@ Implemented now:
 - a provider-neutral read-only local-surface view in
   **Парк → Диагностика LiDAR**, synchronized to the five existing
   RAVNOVES00 scene selectors and a clickable complete-recording timeline;
+- a complete recorded-host-paced 1× worker replay through authenticated
+  ingress, K1 normalization, point/pose binding and a bounded latest-wins
+  surface estimator: `4,206` results at `9.1945 Hz` and `89.530 ms` p95 age;
 - worker storage admission for `D:\NDC_MISSIONCORE\datasets` and
   `/mnt/d/NDC_MISSIONCORE/datasets`;
 - a dedicated, honest dataset catalog in **Полигон → Датасеты**;
@@ -93,11 +98,11 @@ Not implemented:
 - no RELLIS ROS bag admission, continuous synchronized playback or production
   promotion;
 - no ray-cleared free-space or planner-authoritative rolling occupancy map;
-- the accepted K1 local-surface profile now passes a 15-second
-  recorded-source-paced bounded shadow gate; authenticated physical
-  `lio_pcl + lio_pose` binding and its pinned acceptance profile are
-  implemented, but no attached-K1 run has qualified them yet. The residual
-  overlay remains non-authoritative.
+- the accepted K1 local-surface profile now passes both the historical
+  15-second equivalence gate and the complete external-worker replay gate.
+  Physical K1 acquisition, hardware-clock and field-network commissioning
+  remain separate and intentionally unqualified. The residual overlay remains
+  non-authoritative.
 
 ## Product surface boundary