Implement a complete self-improving localization system that uses BLE
RSSI triangulation as ground truth to learn per-link, per-zone weights
for the Fresnel zone fusion engine.
Key components:
- Ground truth sample collection with confidence > 0.7 and distance < 0.5m gates
- Spatial weight learner using online SGD with L2 regularization
- Validation gate that rejects updates without 5% improvement on holdout set
- Bilinear interpolation for smooth weight transitions between zones
- SQLite persistence for weights with 10,000 sample cap per person
- Position accuracy trend visualization in dashboard Accuracy panel
Backend (mothership/internal/localization/):
- groundtruth.go: BLE trilateration provider with Gauss-Newton optimization
- groundtruth_store.go: SQLite storage with weekly accuracy rollups
- spatial_weights.go: SpatialWeightLearner with SGD, validation, interpolation
- weightlearner.go: WeightLearner with error history tracking
- weightstore.go: Weight persistence to SQLite
Frontend (dashboard/js/accuracy.js):
- fetchPositionAccuracy/fetchPositionHistory functions
- drawPositionSparkline for weekly median error trend
- Position accuracy section with median error, trend indicator, sample count
Tests cover:
- Sample collection gates (confidence/distance thresholds)
- SGD weight updates after 100 samples
- Validation gate rejection of adversarial samples
- Bilinear interpolation smoothness
- SQLite sample cap enforcement
- Weight persistence across restarts
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add GetFeedbackInTimeRange() to query feedback by timestamp for accuracy computation
- Add GetUniqueScopeIDs() to extract unique link/zone/person IDs from feedback
- Update accuracy.go to use the new store methods for proper scope filtering
- Add zone breakdown visualization in accuracy panel with clickable zone items
- Add blob hover tooltips with thumbs up/down feedback buttons in 3D view
- Wire up feedback processor and accuracy computer in main startup
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add feedback_store.go with SQLite storage for detection feedback and accuracy metrics
- Add feedback_processor.go for background processing of user feedback
- Add accuracy.go for weekly precision/recall/F1 metric computation
- Add handler.go with REST API routes for feedback submission and accuracy retrieval
- Wire learning package into main_phase6.go with background processing
- Add dashboard/js/feedback.js with thumbs-up/down UI components
- Add dashboard/js/accuracy.js with accuracy panel rendering and sparkline trends
- Add comprehensive tests for feedback storage and accuracy computation
Feedback UI provides:
- Thumbs-up/down buttons for detection events
- Feedback form with false positive/negative options
- Missed detection reporting with position/zone selection
- Motivational counter showing improvement from user corrections
Accuracy panel shows:
- Circular gauge with F1 score
- Week-over-week trend sparkline
- Per-zone breakdown of precision/recall
- Total corrections count and improvement percentage
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
