MetriPlane v0.2.0 - Physical Observability, Evidence Bundles, and Command Center

MetriPlane v0.2.0 — Physical Observability, Evidence Bundles, and Command Center

MetriPlane v0.2.0 is an open-source physical-observability release for bounded planar workcells. It converts calibrated camera/replay-derived state into named assets, traces, spatial contracts, incidents, evidence bundles, operator-review reports, and generated physical regression tests.

This release is the primary software artifact for the planned SoftwareX paper. The historical DOI-archived baseline remains MetriPlane v0.1.4.

Main additions

• Physical-observability release identity
• Sentinel observe-only auditing
• Atlas Evidence Review / Cell Black Box workflow
• Domain packs and deterministic assembly-cell replay
• Physical event ledger
• Cell Truth Report
• Evidence bundle verification
• Generated physical regression tests
• Command Center review UI/API
• Camera trust, local evidence Q&A, counterfactuals, and forecasting evidence
• ROS 2 / Isaac / Omniverse adapter surfaces with bounded claims
• Docker local replay/demo smoke evidence, bounded to build/start, health endpoint JSON, and cleanup logs
• SoftwareX paper evidence package under evidence/paper_v2_0/
• Reviewer kit under docs/review_kit/

Reproduce the paper demo

python3 -m venv .venv
source .venv/bin/activate
pip install -e .

python -m metriplane.cli doctor
PYTEST_DISABLE_PLUGIN_AUTOLOAD=1 python -m pytest -q

./tools/mp.sh deterministic-replay

metriplane atlas validate-pack configs/domain_packs/assembly_cell

metriplane atlas run \
  --session-jsonl datasets/demo/atlas/assembly_cell_missing_tool.jsonl \
  --pack configs/domain_packs/assembly_cell \
  --out runs/atlas/assembly_cell_missing_tool

metriplane atlas bundle verify \
  runs/atlas/assembly_cell_missing_tool/evidence_bundles/INC-0001.zip

metriplane atlas test \
  runs/atlas/assembly_cell_missing_tool/regression_tests/INC-0001.yaml

metriplane atlas dashboard build \
  --run-dir runs/atlas/assembly_cell_missing_tool

Evidence status

The captured v0.2.0 paper evidence package records:

• 580 passing tests
• Deterministic replay pass
• Atlas assembly-cell run: 6 events, 1 incident
• Evidence bundle verification pass
• Generated regression test pass
• Dashboard build pass
• Python sdist/wheel build pass
• twine check pass
• Docker local replay/demo smoke captured as bounded smoke evidence only

Key paper artifacts

The release includes a dedicated paper evidence package:

evidence/paper_v2_0/

Important generated artifacts include:

evidence/paper_v2_0/atlas_run/cell_truth_report.md
evidence/paper_v2_0/atlas_run/cell_truth_report.html
evidence/paper_v2_0/atlas_run/atlas_dashboard.html
evidence/paper_v2_0/atlas_run/physical_event_log.jsonl
evidence/paper_v2_0/atlas_run/reality_graph.json
evidence/paper_v2_0/atlas_run/process_trace.json
evidence/paper_v2_0/atlas_run/evidence_bundles/INC-0001.zip
evidence/paper_v2_0/atlas_run/regression_tests/INC-0001.yaml
evidence/paper_v2_0/CHECKSUMS.sha256

DOI and citation

The v0.2.0 Zenodo archive is available at:

10.5281/zenodo.20736619

Zenodo record:

https://zenodo.org/records/20736619

Recommended citation:

Parkkinen, M. (2026). MetriPlane v0.2.0: Open-Source Physical Observability for Workcell Evidence, Replay, and Regression Testing (Version 0.2.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.20736619

Historical baseline

The historical DOI-archived baseline remains:

MetriPlane v0.1.4 — 10.5281/zenodo.20631037

v0.1.4 provides the earlier repository-stabilized baseline for camera-first planar metric state streaming, deterministic replay, zone analytics, and benchmark evidence. v0.2.0 extends that baseline into physical observability, evidence bundles, operator review, and regression testing.

Scope and limitations

MetriPlane is observe-only.

It does not:

• control robots or machines
• certify safety
• approve quality release
• recognize people
• claim marker-free tracking
• claim full 3D reconstruction
• claim production collision avoidance
• claim factory-wide deployment validation

ROS 2, Isaac, Omniverse, Docker, Jetson, and exporter paths are bounded adapter/deployment surfaces unless separately evidenced.

Docker evidence in this release is local replay/demo smoke only. It is not benchmark, production-runtime, live-camera, reliability, or safety evidence.

SoftwareX status

This release is the primary software artifact for a planned SoftwareX submission. No accepted or peer-reviewed SoftwareX publication is claimed by this release.

v0.1.4 - Repository stabilization

Repository stabilization release. No benchmark numbers or evidence values changed from v0.1.3.

Changed

• Cleaned the repository root so the canonical Python package lives under metriplane/
• Removed duplicate root package directories and root Python module shadows (~7,700 lines of duplication)
• Consolidated command-line helper scripts into the canonical root tools/ directory
• Moved experiment configuration YAMLs into configs/examples/, leaving config.example.yaml at the root
• Added a tracked demo replay fixture at datasets/demo/session_001.jsonl — ./tools/mp.sh deterministic-replay now works on a clean clone with no local data
• Demo runs now write outputs to runs/demo-evidence/ by default; set METRIPLANE_EVIDENCE_OUT=1 to regenerate canonical evidence
• Added CI and script guards for ROS 2 pytest plugin autoload collisions

Validation

• 193/193 tests passing, verified on a fresh clone in a clean environment
• Deterministic replay bit-exact against the tracked fixture
• Evidence data, checksums, and benchmark values untouched

Full reports: docs/reports/phase00_repo_cleanup_report.md and docs/reports/phase00b_reviewer_path_report.md

MetriPlane v0.1.3 - Paper B Provenance-Synchronized Evidence Release

MetriPlane v0.1.3 - Paper B Provenance-Synchronized Evidence Release

This release is a public proof-hygiene and provenance synchronization release for the Paper B benchmark manuscript:

Benchmarking Camera-First Planar Digital Twins: A Reproducible Protocol and MetriPlane Evaluation

What changed

• Synchronizes release metadata, citation metadata, reference audit, manifest, checksums, and proof-check documentation with the v0.1.3 public evidence surface.
• Keeps benchmark measurements unchanged.
• Keeps benchmark measurement CSVs unchanged.
• Maintains the same canonical Paper B metric values.
• Records that no archival DOI, peer-reviewed publication status, accepted-paper status, or ACM artifact badge is claimed for this release.

Verification

From a clean checkout:

git clone --branch v0.1.3 --depth 1 https://github.com/Miko997/metriplane.git metriplane-v013-audit
cd metriplane-v013-audit
sha256sum -c evidence/CHECKSUMS.sha256
python scripts/audit_evidence.py

MetriPlane v0.1.2 - Paper B Provenance-Synchronized Evidence Release

MetriPlane v0.1.2 - Paper B Provenance-Synchronized Evidence Release

This release does not change benchmark measurements. It synchronizes README, evidence index, manifest metrics, checksums, title metadata, citation metadata, reference proof records, and proof-check documentation with the canonical Paper B evidence values.

No archival DOI is claimed for this release. The associated Paper B manuscript remains a preprint-style benchmark manuscript and is not yet peer-reviewed or published.

Canonical evidence:

• docs/eval/CANONICAL_EVIDENCE.md
• evidence/manifest.csv
• evidence/CHECKSUMS.sha256
• ARTIFACTS.md

Validation:

• sha256sum -c evidence/CHECKSUMS.sha256
• python scripts/audit_evidence.py
• PYTEST_DISABLE_PLUGIN_AUTOLOAD=1 python -m pytest -q

Scope:

• Planar XY tracking only.
• ArUco/fiducial markers required.
• No marker-free tracking claim.
• No full 3D reconstruction claim.
• No peer-reviewed publication, accepted paper, archival DOI, or ACM artifact badge is claimed for this release.

MetriPlane v0.1.1 - Paper Canonical Evidence Release

MetriPlane v0.1.1 - Paper B Canonical Evidence Release

This release freezes the canonical public evidence package for the Paper B benchmark manuscript:

Benchmarking Camera-First Planar Digital Twins: A Reproducible Evaluation of MetriPlane

Canonical evidence

• Latency: 4,387 timing samples; detect.cam0 p95 1.242 ms; detect.cam1 p95 1.684 ms; fuse p95 0.184 ms.
• Mapping: 0.63 cm mean error; 1.07 cm max error; N=9 grid points.
• Static continuity: IDs 4, 7, and 12 at 100.0% coverage over 4,387 frames.
• Motion continuity: 88,475 frames; primary-marker coverage 98.39–99.25%.
• Replay determinism: 302 frames; 906 object pairs; 0.0 cm positional difference; 0 event mismatches.
• Backpressure: 120 Hz synthetic overload; queue_max=5; 995 published; 2,605 dropped; pass=true.
• Fusion jitter: 0.067–0.080 mm; absolute fused accuracy not measured in this artifact.
• CPU/GPU equivalence: 13,161 samples; 0.0 cm RMSE and max difference.
• CPU/GPU performance: GPU backend is correct but slower than CPU for tested N=1–1000 fusion-compute workloads.
• Zone analytics: four zones; 877.85 object-seconds dwell; 112 transitions.

Scope

This release supports a scoped benchmark claim for camera-first planar digital twins. It does not claim marker-free recognition, full 3D reconstruction, safety-certified industrial control, or measured live Omniverse/ROS 2 latency.

Key files

• ARTIFACTS.md
• docs/eval/CANONICAL_EVIDENCE.md
• evidence/manifest.csv
• evidence/CHECKSUMS.sha256
• evidence/experiments/*.csv

MetriPlane v0.1.0 - Initial Public Release

MetriPlane v0.1.0 - Initial Public Release

MetriPlane is a camera-first planar digital-twin toolkit for reproducible indoor object-state tracking, metric world-coordinate mapping, multi-camera fusion, zone analytics, WebSocket streaming, JSONL recording, deterministic replay, and benchmark-backed evaluation.

This is the first public release of the MetriPlane source and evidence package.

Highlights

• Camera-first planar tracking using ArUco fiducial markers.
• Metric XY world-coordinate mapping with Z=0.
• One- and two-camera tracking pipeline.
• Multi-camera fusion backend.
• Zone dwell and transition analytics.
• WebSocket state streaming.
• JSONL recording and deterministic replay.
• Operator dashboard and release demo workflow.
• Docker/demo proof and operator UI smoke evidence.
• Benchmark evidence for latency, mapping error, ID continuity, replay determinism, overload behavior, fusion jitter, and CPU/GPU fusion tradeoffs.
• Public demo videos linked from the README.

Benchmark evidence

Primary evidence is available under:

• evidence/
• evidence/manifest.csv
• docs/eval/evidence_index.md
• docs/eval/evidence_matrix.md

Key reported results include:

• Mapping error: 0.63 cm mean, 1.07 cm max over nine grid points.
• Timing run: 4,384 frames over 15.056 s, mean 291.2 FPS.
• Detection p95 latency: 1.50 ms for cam0 and 1.72 ms for cam1.
• Fusion p95 latency: 0.19 ms.
• Static fiducial continuity: 100% for primary markers.
• Motion fiducial continuity: 97.42–99.11% for primary markers.
• Replay determinism: 301-frame fixed-clock replay, 0.0 cm maximum positional difference, 0 event mismatches.
• Backpressure: 120 Hz synthetic overload, bounded queue depth of 5, 2,600 stale frames dropped.
• CPU/GPU equivalence: 0.0 cm RMSE difference over 13,161 samples.
• CPU/GPU fusion benchmark: GPU backend correct but slower than CPU for tested N=1–1000 workloads.

Demo videos

Public demos are linked from the README and the project YouTube channel:

• Public release demo.
• Operator UI and live fusion.
• Deterministic replay and provenance.
• Backpressure and health behavior.
• Two-camera fusion.
• ArUco-driven digital-twin tracking.
• ROS 2 / Omniverse-facing concept demos.

Known limitations

• Planar XY tracking only; Z=0.
• Fiducial markers required.
• Not a marker-free object-recognition benchmark.
• Not a safety-certified industrial-control system.
• GPU benchmark covers only the fusion-compute block, not camera capture or marker detection.
• Large raw JSONL sessions may be archived separately from the GitHub repository.

Provenance

MetriPlane was previously developed under the internal name VisionTwin. Historical evidence artifacts may preserve the old name where editing would break checksums or provenance.