Earth imagery · NASA Blue Marble / BMNG reference
Drag to rotate Earth · wheel to zoom · click a satellite to inspect
VEHICLE-ODI · Orbital Debris Intelligence
ODI 2.0
Orbital structure becomes visible.
A relational decision-research architecture for examining orbital population structure, debris concentration, density, uncertainty and modeled structural load using public catalog-derived evidence.
ODIEXP-001 R4
Complete audited full-catalog run
Complete audited full-catalog run
28,605
LEO objects in audited model window
LEO objects in audited model window
001B → 001F
Divergence-forensics chain
Divergence-forensics chain
Live orbital map · real Earth reference · public GP · SGP4
See where the public satellite population is around Earth now.
A correctly oriented real-Earth reference replaces the abstract sphere. ODI propagates the latest cached public GP elements to the current UTC time in the browser, so latitude and longitude can be read against recognizable continents and oceans. The map starts with the public ACTIVE population; debris and rocket-body layers remain a separate next layer.
LOADING ORBITAL DATA
Preparing current GP mirror…
Earth referenceNASA Blue Marble / BMNG imagery
PropagationSGP4 from OMM/GP elements · UTC now
Current phasePublic ACTIVE population · selected ground subpoint
Next ODI layerDebris · rocket bodies · structural-load overlays
Current audited evidence
One environment. Two distinct structures.
ODI 2.0 now makes the population effect explicit instead of compressing everything into a generic “high-risk” label.
70,661SATCAT source rows
28,605LEO objects analyzed
27/27Preregistered sensitivity runs
PASSInternal result audit
AUDITED SNAPSHOTchecking age…
Full-catalog structural load
LEO 200–600 km
8,246.06modeled load- Objects
- 14,152
- Active payloads
- 13,305
- Fine maximum
- 475 km / 52.5°
With the complete public catalog population used by Core 2.0, this broad region ranks first. The present result is strongly shaped by the large active-payload population in lower LEO.
Structural counterfactual · no active payloads
LEO 600–900 km
4,030.26recomputed load- Objects
- 6,574
- Sensitivity
- 27 / 27
- Fine maximum
- 825 km / 97.5°
When active payloads are excluded before H/X/V are recomputed, the historical 600–900 km broad region returns to rank #1 in every preregistered sensitivity run.
Interpretation: Core 2.0 is sensitive to contemporary occupancy and composition. The audited divergence is primarily a population-structure effect, not evidence of a simple implementation failure. The counterfactual remains an in-model experiment, not a prediction of the physical environment after real removals.
Evidence path
The divergence was investigated, not hidden.
Each experiment narrowed one question while preserving the frozen Core 2.0 and the evidence record.
Legacy current control
Exact v0.2b methodology on the current legacy population preserved LEO 600–900 as the top broad region.
Same-population A/B
Core 2.0 on the same curated population preserved LEO 600–900, including all 27 sensitivity configurations.
Scale-aligned full catalog
With 28,605 LEO objects, broad ordering changed and LEO 200–600 became #1.
Population attribution
Count effect explained 92.05% of the regional load gap; active payloads were the largest positive additive contributor.
Structural counterfactual
Removing active payloads and recomputing restored LEO 600–900 in 27/27; removing debris preserved LEO 200–600 in 27/27.
Public dossier · research and institutional materials
Documentation visible, readable and version-aware.
Current evidence, reproducibility materials, legacy demonstrators, policy research and institutional documents are presented as separate evidence classes.
Research & reproducibility
Zenodo research recordFrozen public reproducibility materials · DOI 10.5281/zenodo.20077230
GitHub repositoryPublic code and reproducibility materials
VEHICLE-ODI Technical BriefArchitecture and methodology
ODI 2.0 evidence snapshotMachine-readable web evidence layer
ODI 2.0 public claim ledgerAllowed language, prohibited overclaims and evidence boundaries
ODI AI ReferencePublic machine-readable project context
Public research proposals & demonstrators
Orbital Debris Responsibility & Removal PaperPolicy and legal-technical research proposal; not current ODI 2.0 operational evidence
Legacy Decision Surface DemoEarlier demonstrator retained for historical reproducibility
v0.2b frozen snapshotHistorical demonstrator evidence associated with the submitted manuscript
Claim boundary
Ambitious, but not exaggerated.
What the laboratory has demonstrated internally
Audited public-catalog computation, full-catalog population sensitivity, same-population controls, scale alignment, additive attribution and a preregistered structural population counterfactual through ODIEXP-001F.
What this public release does not claim
It does not claim live telemetry, validated collision probability, operational conjunction warnings, maneuver authority, removal authority, ESA validation, or external certification. OTI and structural load remain research signals.
Not telemetryCatalog-derived GP/SATCAT evidence and browser visualizations are not spacecraft telemetry.
Not collision probabilityStructural load and OTI are not validated Pc.
External benchmark remains openThe tested ESA-window comparison did not show robust external convergence at the preregistered resolutions.
Counterfactual means in-modelCohort exclusion is a structural experiment, not a forecast of the physical orbital environment after real removal.
Continue exploring
Return to MADRE and the VEHICLE Systems Lab home.
Company membership
NVIDIA Inception Program
VEHICLE Systems Lab is a member of NVIDIA Inception. The NVIDIA Inception program is designed to help startups accelerate innovation and growth through developer resources, training, ecosystem access and program benefits.
Claim boundary: this company membership does not imply NVIDIA endorsement, funding, certification, sponsorship, technical validation, adoption, or approval of VEHICLE-ODI.