Ticket Backlog¶

69 implemented · 17 planned · 86 tickets · 2200 tests

Counts are the files in this directory. Verify them: ls docs/tickets/*.md | grep -v README | wc -l and uv run pytest --collect-only -q | tail -1.

This directory tracks every capability from idea to implementation. Completed tickets are kept as historical records. Open tickets describe what to build next and why.

What's already built¶

Area	Tickets	Ships as
Core schemas & CLI	001–003, 090, 103	`estimate`, JSON envelopes, golden fixtures, public API, `migrate` schema upgrades, the per-command exit-code contract
Energy & feasibility	010–012, 061, 063, 097, 099, 110	Energy budget, mass/density/SoC energy fidelity, scalar `usable_capacity_fraction` derating, geofence (2D + altitude bounds), landing-zone reachability, RTH reserve timeline with opt-in feasibility gate
Scenario runner	020–021, 069	Lost-link policies, per-event contingency overrides, contingency events, assertions
Route physics	030–039, 044, 100	Fidelity v2, layered/spatiotemporal wind, terrain alt, geodesic Dubins divert, obstacle and terrain clearance
SITL (ArduPilot)	040–043	`sitl` command, MAVLink upload, telemetry recording, `compare`
Stochastic propagation	047–049, 086	`propagate` command, twin-state EKF, closed-loop tracking controller; split into `propagation/` subpackage
Output formats	055, 057, 072–075	GeoJSON/KML exports, `summary`, `checklist`, `profile`, `sensitivity`, `size-battery`
Batch, import & export	060, 064, 085, 091	`batch` command with estimate/scenario/propagate run types, `convert` QGC importer with `--vehicle-profile`, `--format csv`, `export` mission→QGC `.plan` writer
Real-world data	052–053, 056, 059	Fetch scripts, community vehicle profiles, infeasible Alpine demo
Regulatory pre-assessment	094–095	SORA Ground Risk Class (iGRC), Air Risk Class, and GRC×ARC→SAIL with applicable OSOs via the `sora` command
Weather & GO/NO-GO	092–093	Enforced `max_wind_mps`/`max_crosswind_mps` weather minimums with a checklist Weather-limits row; time-varying geofence activation windows
Correctness fixes	062, 065, 109	Wind-triangle divert correction, stochastic spatial infeasibility tracking, `WIND_COVERAGE_MISSING` so an out-of-coverage wind query is missing evidence instead of a silently clamped value
Documentation	096	MkDocs Material site published to GitHub Pages, auto-built from `main`
Release tooling	098	`bump` command (semver bump + changelog roll + `--check` drift guard); version-agnostic golden fixtures
Calibration & validation	080–083, 102	`flight-trace.v1` ingestion (incl. binary `.bin`/ULog), `phase-segments.v1` segmentation, `validation-report.v2` predicted-vs-observed metrics and acceptance gate via the `validate` command, `calibration-profile.v2` parameter fitting via the `calibrate` command

Full feature list: docs/tickets/README.md implemented section · Test coverage: uv run pytest

Planned work¶

Items are ordered by impact. Pick one, read its ticket file, open a PR.

Quick wins¶

#	Ticket	What it adds
088	Performance benchmarks	`pytest-benchmark` suite + CI regression gates (no production code changes)
089	Preflight report command	Single `preflight` command combining estimate + scenario + Monte Carlo into one operator briefing

Backend integration readiness¶

Gaps surfaced while reviewing whether the engine can be driven cleanly from a long-running service (the Mission Control web backend) as a one-process-per-job subprocess. None blocks single-shot CLI use; each sharpens the contract a queue worker depends on.

This track is complete: the per-command exit-code contract (103), atomic output writes and clean cancellation (104), contract-version discovery (105), machine-readable run progress (106), and the machine-readable preflight validation envelope (107) are all implemented — see the full list below.

Model fidelity gaps¶

#	Ticket	What it adds
108	Terrain-aware contingency paths	Make RTH and divert paths obey the same terrain, obstacle, and geofence evidence the planned route already obeys; today they are energy-only. Deferred because closing it changes the meaning of `rth_is_feasible` and every divert verdict

Output & visualisation gaps¶

#	Ticket	What it adds
076	Departure window finder	Scan a forecast window for the first feasible departure time
077	Mission comparison report	Side-by-side `diff` of two mission estimates
066	Stochastic GeoJSON export	`--format geojson/kml` for `propagate` command
067	Propagation progress feedback	Progress bar / stderr updates during long particle runs
068	Divert route GeoJSON layer	Add computed divert paths as a separate layer in scenario GeoJSON/KML

Hardware validation ladder¶

#	Ticket	What it adds
045	PX4 SITL — launch & upload	PX4 SITL adapter behind the existing evidence contract
046	PX4 SITL — telemetry & evidence	PX4 artifact recording and evidence bundle assembly
087	HITL adapter	Real flight controller hardware (Pixhawk) in the validation loop

Integrations & platform¶

#	Ticket	What it adds
054	Reference calibration inputs	Curated reference datasets for model calibration and import
058	NOTAM / live airspace	Dynamic no-fly zone ingestion from live NOTAM feeds
050	REST API & web UI	HTTP service wrapper and browser-based preflight interface
070	Operational integration seams	Hooks for flight-ops systems, flight plans, and operator dashboards
071	Live comms, Remote ID, traffic	UTM/U-space, Remote ID broadcast, and traffic awareness

Calibration & validation track¶

A separate track for post-flight model calibration from real flight logs. Tickets 080, 081, 082, and 083 are complete; only 084 (holdout validation reports) remains.

#	Ticket	Status
080	Flight log ingestion & trace normalisation	implemented
081	Flight phase segmentation	implemented
082	Predicted vs. observed validation metrics	implemented
083	Calibration profile data & fitting	implemented
084	Holdout validation reports	planned
102	Binary flight-log ingestion (.bin, ULog)	implemented

Contributing¶

Before opening a PR, read the ticket file for the work you're doing. Each ticket states its acceptance criteria, composition with existing code, and what tests are expected.

Rules¶

Keep core execution deterministic. Randomness lives in Monte Carlo and stochastic layers only.
Add adapter layers only after core contracts are stable.
All schemas and public outputs are versioned. Bump the version string when the shape changes.
Reject unsupported inputs explicitly — never approximate silently.
No live external network calls in core CI.
Update docs, tests, and golden fixtures in the same commit when public behaviour changes.

Integration standard¶

Every new ticket must compose through the existing surfaces:

YAML schemas: mission.v7, vehicle.v6, scenario.v1, uncertainty.v2, stochastic.v2, batch.v1 — bvlos-sim schema-versions prints the authoritative set
Examples: examples/missions/, examples/vehicles/, examples/scenarios/
CLI commands: estimate, scenario, sample, propagate, batch, convert, export, migrate, sitl, compare, verify, size-battery, sora, ingest-log, validate, calibrate, bump, schema-versions
Output contracts: canonical JSON envelopes, Markdown reports, golden fixtures
Public Python API: estimator.try_estimate_mission_distance_time, estimator.run_scenario, estimator.execution.monte_carlo.run_monte_carlo

New capabilities should work with existing pieces, not alongside them in isolation.

Implemented tickets¶

Full list (69 tickets)¶

001 Estimator CLI and envelope
002 Versioning and golden fixtures
003 Technical debt hardening
010 Deterministic energy feasibility
011 Static geofence feasibility
012 Static landing-zone reachability
020 Scenario runner core
021 Comms link and contingency policies
030 Fidelity controls plus independent layered wind and sub-segments
031 Fidelity v2 — turns and fixed-wing loiter
032 Terrain-referenced altitude execution
033 Continuous spatiotemporal wind grid
034 Resource and link feasibility abstractions
035 Dynamic landing-zone availability
036 Computed divert routing
037 Monte Carlo uncertainty modelling
038 Bank-angle and Dubins path optimisation
039 Path planning model gaps
040 SITL adapter contract and evidence schema
041 ArduPilot SITL launch and mission upload
042 SITL telemetry recorder and evidence bundle
043 SITL scenario comparison report
047 Stochastic state propagation
048 Observation model and twin-state EKF
049 Stochastic closed-loop control
052 Real-world data fetch scripts
053 Airspace geofence fetch script
055 GeoJSON and KML route export
056 Community vehicle profiles
057 Summary output format
059 Infeasible demo mission
060 Import, export, and batch workflows
062 Wind-corrected divert energy
063 Return-to-home reserve check
065 Geofence and LZ in stochastic propagation
069 Per-event lost-link policy override
072 Route altitude profile report
073 Pre-flight checklist output
074 Energy reserve sensitivity report
075 Minimum battery sizing command
085 QGC convert vehicle profile selection
086 Stochastic propagator module split
094 SORA Ground Risk Class (iGRC)
095 SORA Air Risk Class and SAIL determination
091 QGC mission export
092 Weather minimums and automatic GO/NO-GO
093 Time-varying geofence activation
096 Documentation site on GitHub Pages
099 Energy-model fidelity
100 Obstacle database and vertical clearance checks
101 SORA 2.5 mitigation depth (original M1–M3/tactical-credit design superseded)
097 Opt-in RTH reserve feasibility gate
061 3D geofence altitude bounds
080 Flight log ingestion and trace normalization
081 Flight phase segmentation
082 Predicted vs. observed validation metrics
098 Version bump and release tooling
083 Calibration profile data and fitting
104 Atomic output writes and clean cancellation
105 Contract-version discovery command
106 Machine-readable run progress
107 Machine-readable preflight validation report
044 Geodesic Dubins divert path
064 Batch scenario and batch propagate modes
090 Schema-version migration tooling (migrate, mission.v6 → mission.v7)
102 Binary flight-log ingestion (.bin, ULog)
103 Backend-facing CLI exit-code contract
109 Wind coverage fails closed (WIND_COVERAGE_MISSING)
110 usable_capacity_curve replaced by scalar usable_capacity_fraction