Skip to content

Ticket 094: SORA Ground Risk Class (iGRC)

Status

Implemented.

Goal

Compute the intrinsic Ground Risk Class (iGRC) for a mission by evaluating the declared operational footprint against conservative population evidence and the aircraft's maximum characteristic dimension and speed, following JARUS SORA 2.5. Output per-leg and whole-mission iGRC pre-assessment evidence.

Why This Is High Impact

SORA (Specific Operations Risk Assessment) is the regulatory framework that governs BVLOS approvals in the EU, UK, and a growing number of other jurisdictions. The first number every SORA submission needs is the intrinsic Ground Risk Class — a table lookup derived from the aircraft's characteristic dimension and the population density beneath the flight path.

bvlos-sim already models the route, terrain, wind, energy, and geofences. The one thing it cannot answer is the question a regulator asks first: "what is the ground risk if this aircraft comes down here?" Adding iGRC computation moves bvlos-sim from an engineering feasibility tool into a regulatory pre-assessment tool — by far the most compelling reason a professional BVLOS operator would adopt it.

Background: the SORA iGRC table

The intrinsic GRC is a lookup on two axes:

  • Aircraft column: the first Table 2 column whose maximum characteristic dimension and maximum speed both cover the aircraft. The 250 g / 25 m/s exception is handled separately.
  • Population density across the assessed operational footprint: controlled ground area, <5 ppl/km², <50, <500, <5 000, <50 000, >50 000 (gatherings).

The cell value is the iGRC. Higher dimension and higher population both raise the class. Values above 7 are reported but annotated as outside the standard specific-category envelope.

Scope

New population-density grid asset

A new evidence asset, parallel to the terrain grid, supplying conservative population density per grid cell in people per km²:

# population_grid.v2
schema_version: population-grid.v2
origin_lat: 47.04
origin_lon: 8.29
step_lat_deg: 0.001
step_lon_deg: 0.001
density_ppl_km2:
  - [12.0, 12.0, 340.0, ...]   # one row per latitude step, south to north
  - ...
metadata:
  source: Authority-approved conservative population map
  population_year: 2026
  native_resolution_m: 100.0
  effective_resolution_m: 100.0
  value_semantics: conservative_cell_maximum
  authority_assessment_reference: POP-001 rev 1
  valid_from: 2026-01-01T00:00:00Z
  valid_until: 2026-12-31T23:59:59Z
  transient_population_assessment_reference: EVT-001 rev 1
  operational_footprint_assemblies_present: false

Referenced from the mission:

assets:
  population_grid_file: assets/population.yaml

New vehicle field

# schemas/vehicle.py — VehicleProfile
characteristic_dimension_m: float | None   # maximum aircraft dimension

When omitted, iGRC is not computed and a POPULATION_DENSITY_DIMENSION_MISSING advisory warning is emitted (consistent with other provider-dependent features).

Computation

  • For each route leg, conservatively cover the declared footprint using route samples and the population grid's effective resolution.
  • Take the maximum conservative cell value that can overlap that footprint.
  • Select the aircraft column using maximum characteristic dimension and maximum speed, then map the maximum density to iGRC via the SORA 2.5 table.
  • The mission iGRC is the maximum iGRC across all legs.

Output integration

  • New --format ground-risk for estimate: a Markdown table of per-leg iGRC, population density, and the governing waypoint, plus the mission-level iGRC.
  • --format checklist gains a Ground risk class row showing the mission iGRC.
  • --format json result envelope gains a ground_risk block: 
    "ground_risk": {
  "characteristic_dimension_m": 1.5,
  "mission_igrc": 4,
  "legs": [{"leg_index": 1, "max_density_ppl_km2": 340.0, "igrc": 4}]
}
  • --format geojson route legs gain an igrc property so the route can be colour-coded by ground risk in QGIS / Google Earth.

Fetch script

scripts/fetch_population.py pulls WorldPop point samples for diagnostics. Its output is deliberately marked ineligible for operational SORA use; it does not provide conservative cell maxima, authority approval, transient-population assessment, or bounded evidence validity.

New schema and enum additions

File Change
schemas/vehicle.py Add characteristic_dimension_m
schemas/mission.py Add assets.population_grid_file
estimator/core/enums.py Add POPULATION_DENSITY_DIMENSION_MISSING warning
estimator/core/results.py Add GroundRiskEstimate + GroundRiskLegEstimate
estimator/environment/population.py New population grid provider
estimator/execution/ground_risk.py New iGRC computation
adapters/assets/population_grid.py New loader
adapters/ground_risk_markdown.py New --format ground-risk renderer
scripts/fetch_population.py New fetch script
tests/test_ground_risk.py New unit + integration tests
docs/USAGE.md New ## Ground Risk (SORA iGRC) section

Non-goals

  • This ticket computes the intrinsic GRC only. Mitigations were out of its original scope and belong to Ticket 101. The current SORA 2.5 contract uses M1(A/B/C) and M2; the former M3 ERP treatment is not a ground-credit input.
  • Air Risk Class and SAIL determination are Ticket 095.
  • The population grid is offline; live population data fetch is best-effort in the fetch script only, never in core estimation.

Composition

  • The population grid loads exactly like the terrain grid (Ticket 032 pattern), through assets.population_grid_file, resolved relative to the mission file.
  • iGRC computation reuses the sub-segment sampling already implemented for wind.
  • Output flows through the existing envelope/markdown/geojson surfaces.

Acceptance criteria

  1. A mission over a 12 ppl/km² area with a 1 m aircraft returns the correct low iGRC; the same route over a 5 000 ppl/km² cell returns a higher iGRC, matching the implemented SORA-style table.
  2. A vehicle with no characteristic_dimension_m emits POPULATION_DENSITY_DIMENSION_MISSING and omits the ground_risk block.
  3. estimate --format ground-risk produces a per-leg iGRC table and a mission-level iGRC.
  4. estimate --format geojson includes an igrc property on each route leg.
  5. The SORA lookup table is unit-tested at every dimension, speed, and density boundary.
  6. A mission with no population_grid_file is unaffected (backward compatible).