gps-denied-onboard/.planning/phases/01-eskf-core/01-01-SUMMARY.md at 1b069e2bb3b2867d608f05513ed20f62d64b4837

mirror of https://github.com/azaion/gps-denied-onboard.git synced 2026-04-23 09:26:38 +00:00

Files

T

Yuzviak 094895b21b feat(phases 2-7): implement full GPS-denied navigation pipeline

Phase 2 — Visual Odometry:
  - ORBVisualOdometry (dev/CI), CuVSLAMVisualOdometry (Jetson)
  - TRTInferenceEngine (TensorRT FP16, conditional import)
  - create_vo_backend() factory

Phase 3 — Satellite Matching + GPR:
  - SatelliteDataManager: local z/x/y tiles, ESKF ±3σ tile selection
  - GSD normalization (SAT-03), RANSAC inlier-ratio confidence (SAT-04)
  - GlobalPlaceRecognition: Faiss index + numpy fallback

Phase 4 — MAVLink I/O:
  - MAVLinkBridge: GPS_INPUT 15+ fields, IMU callback, 1Hz telemetry
  - 3-consecutive-failure reloc request
  - MockMAVConnection for CI

Phase 5 — Pipeline Wiring:
  - ESKF wired into process_frame: VO update → satellite update
  - CoordinateTransformer + SatelliteDataManager via DI
  - MAVLink state push per frame (PIPE-07)
  - Real pixel_to_gps via ray-ground projection (PIPE-06)
  - GTSAM ISAM2 update when available (PIPE-03)

Phase 6 — Docker + CI:
  - Multi-stage Dockerfile (python:3.11-slim)
  - docker-compose.yml (dev), docker-compose.sitl.yml (ArduPilot SITL)
  - GitHub Actions: ci.yml (lint+pytest+docker smoke), sitl.yml (nightly)
  - tests/test_sitl_integration.py (8 tests, skip without SITL)

Phase 7 — Accuracy Validation:
  - AccuracyBenchmark + SyntheticTrajectory
  - AC-PERF-1: 80% within 50m ✅
  - AC-PERF-2: 60% within 20m ✅
  - AC-PERF-3: p95 latency < 400ms ✅
  - AC-PERF-4: VO drift 1km < 100m ✅ (actual ~11m)
  - scripts/benchmark_accuracy.py CLI

Tests: 195 passed / 8 skipped

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

2026-04-02 17:00:41 +03:00

1.3 KiB

Raw Blame History

plan, phase, status, started, completed

plan	phase	status	started	completed
01-01	01-eskf-core	complete	2026-04-01T20:36:07Z	2026-04-01T20:45:00Z

Summary

Implemented the 15-state Error-State Kalman Filter (ESKF) from scratch using NumPy only. Covers ESKF-01 through ESKF-05.

Key Files Created

src/gps_denied/schemas/eskf.py (68 lines) — ConfidenceTier, IMUMeasurement, ESKFConfig, ESKFState schemas
src/gps_denied/core/eskf.py (359 lines) — ESKF class with predict(), update_vo(), update_satellite(), get_confidence(), initialize_from_gps()

Commits

57c7a6b feat(eskf): add ESKF schema contracts (ESKF-01, ESKF-04, ESKF-05)
9d5337a feat(eskf): implement 15-state ESKF core algorithm (ESKF-01..05)

Tasks Completed

Task	Status	Notes
Task 1: ESKF schema contracts	✓	All 4 classes importable
Task 2: ESKF core algorithm	✓	All acceptance criteria passed

Deviations

None — implemented as specified in plan.

Self-Check: PASSED

ESKF class importable
predict() propagates state and grows covariance
update_vo() reduces position uncertainty via Kalman gain
update_satellite() corrects absolute position
get_confidence() returns correct tier
initialize_from_gps() uses CoordinateTransformer
All math NumPy only, no external filter library

1.3 KiB Raw Blame History