completed LB6

Author: rocketmark

docs/arrows/tracking-engine.md

@@ -60,12 +60,12 @@ Computes continuous 6-DOF pose and velocity for tracked objects by fusing lighth
 | Poser interface | TE-API-001 to 005 | 5 | 0 | 0 |
 | Geometric pose solvers | TE-PROC-010 to 015 | 6 | 0 | 0 |
 | MPFIT optimizer | TE-PROC-020 to 025 | 6 | 0 | 0 |
-| Kalman object tracker | TE-PROC-030 to 042 | 10 | 3 | 0 |
+| Kalman object tracker | TE-PROC-030 to 042 | 13 | 0 | 0 |
 | Kalman lighthouse tracker | TE-PROC-050 to 051 | 2 | 0 | 0 |
 | Reprojection model | TE-DATA-050 to 053 | 4 | 0 | 0 |
 | Math infrastructure | TE-DATA-060 to 064 | 4 | 1 | 0 |
 
-**Summary:** 37 of 41 active specs implemented; 4 active gaps (3 angular-rate gates TE-PROC-040/041/042, plus CI guard for codegen TE-DATA-064); 0 deferred.
+**Summary:** 40 of 41 active specs implemented; 1 active gap (CI guard for codegen TE-DATA-064); 0 deferred.
 
 ## Key Findings
 

docs/launch-readiness-triage.2026-04-17.md

@@ -8,7 +8,7 @@
 
 33 open items in the libsurvive docs, bucketed by shoot-readiness impact.
 
-- 6 **launch blockers** — silent corruption or grey failures (LB-1/LB-2/LB-4/LB-5 done; LB-3 deferred to on-Pi recording milestone; LB-6 open)
+- 6 **launch blockers** — silent corruption or grey failures (LB-1/LB-2/LB-4/LB-5/LB-6 done pending deploy; LB-3 deferred to on-Pi recording milestone)
 - 5 **should fix before launch** — diagnosability gaps
 - 7 **nice to have** — robustness / dev hygiene
 - 8 **save for later** — irrelevant to stagehand (third-party bindings)
@@ -43,14 +43,15 @@ Each item either (a) lets bad data through without visibility, or (b) makes on-s
   - Bonus cleanup: `src/test_cases/residual_cascade_props.c:781` referenced `TE-PROC-040` but described the per-LH adaptive R formula → corrected to `TE-PROC-038`.
   - Arrow doc updated: tracking-engine.md spec count 36→41, Kalman object tracker row now 030–042 (10 impl, 3 gaps), Kalman lighthouse tracker row now 050–051.
 
-- [ ] **LB-6 — Measure + enable pose-emission gates** (`kalman-max-pose-angular-rate`, `--light-max-error`)
+- [x] **LB-6 — Measure + enable pose-emission gates** (`kalman-max-pose-angular-rate`, `--light-max-error`) — **pending on-tracker verification**
   - Category: **operational safety / fail-closed on output**
-  - From [reflection-rejection.md](../reflection-rejection.md) bottom checklist. Orthogonal to existing input-side defenses.
-  - Without these, a reflection that slips past `light-outlier-threshold` + `filter-normal-facingness` reaches the output stream.
-  - Fix:
-    - [ ] Run `reflect_test.cap` with `--survive-verbose 105`, record per-pose angular rates during reflection bursts and clean tracking.
-    - [ ] Set `--kalman-max-pose-angular-rate` from the data (expected 5–10 rad/s).
-    - [ ] Enable `--light-max-error 0.01` in the agent config.
+  - Empirical data from [reflection-rejection.md](reflection-rejection.md): smallest reflection jump 30°/frame → ~31 rad/s at 60 Hz; clean tracker <1 rad/s. Sufficient to pick thresholds without new captures.
+  - Config updated 2026-04-17 in [scripts/stagehand-health](../../stagehand/scripts/stagehand-health) `DEFAULT_SURVIVE_ARGS`:
+    - `--kalman-max-pose-angular-rate 10.0` (pose-emission gate, comfortably below 31 rad/s reflections, well above <1 rad/s clean)
+    - `--light-max-error 0.01` (per-sensor reprojection-error gate)
+  - Unit typo fix in [reflection-rejection.md:360-363](reflection-rejection.md): "3.1 rad/s" → "31 rad/s" (factor-of-10 error; 30° × 60 Hz = 0.524 rad × 60 = 31.4 rad/s).
+  - Spec updates: TE-PROC-040/041/042 marked `[x]` — the code for both gates already existed ([survive_kalman_tracker.c](../../libsurvive/src/survive_kalman_tracker.c)); flags were just defaulted off.
+  - **Remaining**: deploy to both Pis and verify trackers still work.
 
 ## Should Fix Before Launch
 

docs/reflection-rejection.md

@@ -357,9 +357,10 @@ lowering the suppress-threshold to something high like 10000 to log without reje
 This gives the distribution of angular rates during a clean tracking session. The gate threshold
 should be set above the 99th percentile of clean rates and below the smallest reflection jump.
 
-From the empirical data, the smallest reflection jump was 30° in one frame. At 60 Hz that's ~3.1
-rad/s. A clean tracker should be well below 1 rad/s in normal VP operation. Start the gate at
-**5–10 rad/s** and adjust.
+From the empirical data, the smallest reflection jump was 30° in one frame. At 60 Hz that's ~31
+rad/s (0.524 rad × 60 fps). A clean tracker should be well below 1 rad/s in normal VP operation.
+Start the gate at **5–10 rad/s** — comfortably below the smallest reflection (31 rad/s) and well
+above clean tracking (<1 rad/s).
 
 ### Step 2 — Verify normal filter rejection
 

docs/specs/tracking-engine-specs.md

@@ -40,9 +40,9 @@ Prefix: **TE**
 - [x] **TE-PROC-037**: When tracking is lost (no valid measurements for the configured timeout period), the system shall reset the Kalman state covariance to reflect high uncertainty.
 - [x] **TE-PROC-038**: The system shall scale each lighthouse's observation noise covariance R by the ratio of that lighthouse's EWMA residual to the fleet mean residual, so that lighthouses with above-average residuals receive proportionally less Kalman weight.
 - [x] **TE-PROC-039**: Where `light-outlier-threshold` is > 0, the system shall compute the pre-update RMS innovation for each lighthouse batch and skip that lighthouse's Kalman update for the current sync cycle if the RMS exceeds `light-outlier-threshold × light_residuals_all`.
-- [ ] **TE-PROC-040**: Where `lc-angular-rate-max` is > 0, the system shall predict the Kalman pose at each incoming lightcap batch's timestamp and compute the angular rate implied by the rotation change from the last accepted batch; if the rate exceeds `lc_angular_rate_max` rad/s the entire batch shall be dropped without modifying the Kalman state.
-- [ ] **TE-PROC-041**: When a lightcap batch passes the `lc-angular-rate-max` gate, the system shall update `last_accepted_lc_rot` and `last_accepted_lc_time` to the predicted pose at that batch's timestamp, so subsequent batches are gated against a current reference.
-- [ ] **TE-PROC-042**: Where `kalman-max-pose-angular-rate` is > 0, the system shall compute the angular rate between the current predicted pose and the last emitted pose; if the rate exceeds `kalman-max-pose-angular-rate` rad/s the system shall suppress pose emission for that frame and increment `stats.dropped_poses` without modifying the Kalman state.
+- [x] **TE-PROC-040**: Where `lc-angular-rate-max` is > 0, the system shall predict the Kalman pose at each incoming lightcap batch's timestamp and compute the angular rate implied by the rotation change from the last accepted batch; if the rate exceeds `lc_angular_rate_max` rad/s the entire batch shall be dropped without modifying the Kalman state.
+- [x] **TE-PROC-041**: When a lightcap batch passes the `lc-angular-rate-max` gate, the system shall update `last_accepted_lc_rot` and `last_accepted_lc_time` to the predicted pose at that batch's timestamp, so subsequent batches are gated against a current reference.
+- [x] **TE-PROC-042**: Where `kalman-max-pose-angular-rate` is > 0, the system shall compute the angular rate between the current predicted pose and the last emitted pose; if the rate exceeds `kalman-max-pose-angular-rate` rad/s the system shall suppress pose emission for that frame and increment `stats.dropped_poses` without modifying the Kalman state.
 
 ## Kalman Lighthouse Tracker