Skip to content

Aviation and SATCOM roadmap

Status: WebSDR Phase 10 has accepted 1090ES/shared-foundation and UAT aircraft slices; wider protocol, broadcast and analytics requirements remain open. Phase 11 receive/product work has partial evidence and open operator, reliability and isolation gates. Server acceptance does not close these gates.

Dependencies and verification

Use the aviation guide, UAT criteria, SATCOM guide and GOES verification. Both domains reuse lane ingress, rational timing, decoder isolation, replay and the shared map; neither may acquire an independent receiver path.

These are retained acceptance records, not qualification of the current source. Checked items keep their recorded scope; implementation, qualification and operator acceptance are distinct. Later work does not close unchecked gates.

Phase 10 — Aviation / Aeronautical RF Intelligence

Phase 10 turns the decoder, learned-intelligence, investigation, map, and timeshift foundations into one coherent aviation subsystem. It is not a replacement aircraft-tracking application bolted onto RF Sentry and it must not introduce a second IQ path, map, clock, selection model, or unbounded Pi service.

Current implementation includes 1090ES and UAT, shared aircraft intelligence, FAA enrichment, map tracks and paged history. The first foundation acceptance and later UAT aircraft acceptance do not close all Phase 10 requirements. Live FIS-B weather acceptance remains open. Current source can implement an unchecked requirement without establishing its required qualification.

Protocol priorities

1090ES and 978 UAT provide complementary aviation coverage. Receiver bandwidth, antenna/filter/LNA suitability and observed traffic must be verified for each installation; no original deployment locality or antenna is a requirement. Shared identity, capture time, metadata and map contracts precede new adapters.

Architectural invariants

  • Reuse DecoderManifestV2, DecoderManager, the bounded decoder input worker/WASM ABI, capture-stamped NormalizedDecoderEvent, and the existing event store. Full-rate IQ must not pass through React or the main thread.
  • Reuse the committed-analysis and intelligent-activation boundary. A channel allocation is a prior; a valid preamble, frame, parity/CRC/FEC result, or protocol state is evidence. Frequency alone never confirms an aviation protocol.
  • Add a provider-neutral AviationIntelligenceService rather than forcing aircraft into amateur StationIntelligenceService callsign identity. The two services may share worker/store patterns, capture-time rules, map and investigation shells, and generic selection contracts without merging their identities or persistence.
  • Keep the ICAO 24-bit address, registration, transmitted flight ID/callsign, operator, aircraft type, and a temporal flight/track instance distinct. Addresses and registrations can be reassigned or entered incorrectly; flight IDs change by operation; TIS-B and privacy addresses may not identify a stable airframe.
  • Aviation observations use RF capture time and replay-stable identity. An older replayed frame must not become the latest aircraft state or inflate counts, and out-of-order ingestion must converge to the same history.
  • Store retained aviation observations/tracks in a dedicated browser IndexedDB tier, expose compact immutable aircraft/message projections from RAM, and keep ordinary live rows/markers free of SQLite, OPFS, or IndexedDB reads. Large history and searches may be worker/query driven.
  • Extend the shared investigation target with explicit aircraft, aviation ground-station, and aviation-message targets. Do not manufacture a learned amateur station or FCC identity for an aircraft callsign.
  • Extend the protocol-neutral map presentation with aircraft and aviation ground-station entities and aviation relationships. The accepted MapStore, GPU overlay, receiver geography, geodesy, layer/filter, selection, dossier, and pop-out architecture remain authoritative.
  • The active lane's MapReferenceTime is also the aviation reference time. Live state advances with captured IQ; pause, rewind, and historical playback query aircraft state, trails, messages, ground stations, and relationships strictly as of the capture cursor. Never project a future position or later metadata observation backward.
  • The Raspberry Pi may fetch, validate, version, and atomically publish bounded metadata artifacts. Aircraft decoding, track fusion, search, investigation, and map rendering remain browser/client work. Metadata refresh must not wake an SDR and failure must retain the last good generation.
  • No received or decoded aviation information is safety-certified. Weather, traffic, emergency, navigation, and collision-avoidance presentation must be labelled experimental and must not be represented as suitable for flight operations.
  • Keep data browser-local by default. Do not feed FlightAware, OpenSky, ADS-B Exchange, OGN, ACARS aggregators, or any other third party without a later, explicit operator-approved outbound integration.

Phase 10A — Contracts, local metadata, and Aviation workspace

  • [ ] Define versioned aircraft-address, airframe, registration, operator, flight/track, ground-station, aviation-message, weather-product, signal- observation, and source-provenance contracts. Preserve raw protocol fields and validation evidence while exposing bounded normalized projections.
  • [ ] Define address namespaces and qualifiers for ICAO-assigned, self-assigned, anonymous/privacy, TIS-B/track-file, Mode A/C-only, FLARM, and unknown identities. Never treat every 24-bit value as a permanent ICAO airframe.
  • [x] ~~Define the first validated-event adapter into capture-time aviation observations. Its 1090ES observation identity includes receiver/path, capture epoch/sample position, protocol, and raw-frame signature rather than job, tab, arrival, or replay time.~~
  • [x] ~~Implement the dedicated aviation store/service with incremental RAM projections, replay deduplication, and IndexedDB observation history.~~
  • [ ] Extend persistence diagnostics from the current duplicate/failure counts to explicit lost/pruned/incomplete-history semantics before configurable aviation retention is introduced.
  • [x] ~~Extend shared selection/investigation with an explicit aircraft target while preserving existing station and regulatory targets.~~
  • [ ] Add aviation ground-station, message, and broadcast-product targets as their protocols arrive.
  • [x] ~~Build an adaptive Aviation workspace inside the existing pop-out shell without creating another receiver session.~~
  • [x] ~~Implement the protocol-neutral Surveillance aircraft/history table and shared bounded investigation pane, including persisted column layout and stable first-seen sorting.~~
  • [ ] Add protocol-shaped Messages for ACARS/VDL2/HFDL, Broadcast products for UAT FIS-B, and Signals/receivers views without turning the workspace into one giant sparse table.
  • [ ] For position-bearing protocols, expose flight ID, ICAO/address type, registration, model/type, operator, position provenance, pressure and geometric altitude separately, groundspeed, track/heading, vertical rate, squawk/emergency, integrity/quality fields, receiver-relative range/bearing, first/last seen, RF facts, and a capture-time trail.
  • [ ] For messaging protocols, emphasize direction, aircraft and ground- station identity, frequency, link/protocol stack, message/application type, flight ID, multipart/reassembly state, validation, decoded content, and safe raw-field inspection. Do not imply that a message-derived position is a continuous track.
  • [x] ~~Window the retained aircraft table and observation investigation DOM; use RAM snapshots/subscriptions for ordinary live rendering.~~
  • [ ] Add query-driven IndexedDB paging/search before observation histories can exceed the currently hydrated RAM corpus.
  • [x] ~~Add a generic aviation map adapter that publishes RAM snapshots and keyed deltas into the existing map. Aircraft markers, heading, altitude, recency, trails, selection, receiver range/bearing, filters, and dossier focus use the existing GPU and map-selection paths.~~
  • [ ] Make aircraft trails derived from capture-time positions with explicit gaps and source changes. Never interpolate across a discontinuity, impossible jump, address change, missing interval, or replay boundary without marking the gap.

Local-first aviation metadata

Use the established updater/publication pattern:

bounded background updater on the appliance
  -> versioned source archives + checksums + manifest
  -> atomic last-good appliance publication
  -> same-origin browser fetch
  -> staging/validation in a browser worker
  -> durable browser generation + RAM hot indexes
  -> decoder/workspace/map enrichment
  • [x] ~~Add a low-priority timer/oneshot and status manifest for aviation metadata. Downloads and preprocessing must be bounded, resumable where useful, atomic at generation boundaries, and independent of receiver demand.~~
  • [x] ~~Start with the FAA Releasable Aircraft Database. It is refreshed daily and currently downloads as an approximately 60 MB archive containing the master registration data, make/model and engine references, deregistered aircraft, dealer/document indexes, and reserved N-numbers. Preserve source generation, field provenance, effective/status dates, ICAO Mode S hex, registration, type/model, and available operator/registrant facts used by the current hot projection.~~
  • [x] ~~Respect the FAA's withheld-personal-information mechanism and missing fields. A registry name or address is regulatory metadata, not an aircraft position, operating base, or proof of the current operator. Do not geocode a registrant address into a live aircraft track.~~
  • [ ] Add the FAA 28-day NASR CSV generation for U.S. airports, communication facilities/frequencies, ASOS/AWOS, navaids, ILS, and other selected ground- station context. Keep its 28-day effective date and do not use stale airport data as operational flight guidance.
  • [ ] Evaluate the current FAA contractions/operator material and any ICAO Doc 8643/8585 inputs separately. Do not redistribute scraped or copyrighted designator/operator data without an explicit compatible license.
  • [ ] Keep international airframe providers pluggable. OpenSky's downloadable metadata is technically useful but its current terms of use restrict operational use and redistribution; it must not become a default bundled/runtime corpus without written permission. Apply the same source, privacy, provenance, and redistribution audit to tar1090/Mictronics-derived databases and each national registry.
  • [ ] Permit a documented operator-supplied international metadata generation when the operator has lawful access. Missing international personal data must degrade to address, registration, transmitted category/type code, and observed state without making the aircraft look broken.
  • [ ] Benchmark browser durable representations against actual queries before locking the format. A generation-scoped SQLite/WASM + OPFS substrate with RAM address/registration summaries is a strong fit, but the FCC decision is precedent rather than automatic proof. Compare it with simpler OPFS/RAM or IndexedDB for import, activation, lookup, history joins, storage, and update cost.
  • [x] ~~Keep the current valid FAA metadata generation readable during refresh; validate archive hashes, schema, counts, address uniqueness/ambiguity, and index integrity before atomically activating a replacement. Never destroy a valid generation merely to make space for staging.~~
  • [ ] Expose provider unavailable, stale, ready/no-match, matched, ambiguous, withheld, and source-conflict states separately. Decoder/track learning must never depend on metadata readiness.

Phase 10B — 1090ES ADS-B, Mode S, and Mode A/C

This is the first RF decoder slice because it has the highest expected Central Texas yield and exercises the complete aircraft/map/replay path. readsb is a strong GPL-3-licensed research candidate; dump1090-fa remains a comparison fixture. Select exact source files only after the usual license and transitive-source audit.

  • [x] ~~Build an allowlisted browser-WASM 1090 MHz decoder that consumes bounded complex IQ or an explicitly manifested envelope/pulse domain at a qualified rate such as 2.4 MS/s. Remove direct-SDR, network-server, filesystem, feeder, and web-map code from the imported execution surface.~~
  • [x] ~~Decode CRC-valid DF17/DF18 extended squitters for identity, airborne and surface CPR, altitude, velocity, track/heading, vertical rate, squawk, and emergency/status fields currently exposed by the browser decoder.~~
  • [ ] Decode short and extended Mode S replies, 1090ES airborne/surface position, identity, velocity, status, target state, operational status, and useful Comm-B/EHS registers beyond the current DF17/DF18 tranche. Retain downlink format, subtype, parity/correction, quality fields, and raw frame.
  • [x] ~~Implement conservative airborne even/odd CPR and surface CPR resolved only against a recent valid same-aircraft position or capture-time receiver coordinate with age, range, and motion plausibility bounds.~~
  • [ ] Continue Compact Position Reporting refinement. Distinguish globally decoded even/odd pairs from receiver-relative local CPR, require a valid capture-time receiver coordinate for the latter, enforce age and plausibility limits, and never choose a visually convenient position.
  • [ ] Preserve barometric altitude, geometric altitude, track, heading, indicated/true airspeed, groundspeed, vertical-rate source, squawk, emergency, on-ground, emitter category, NIC/NAC/SIL/SDA, and version as distinct facts instead of flattening them into one aircraft state.
  • [ ] Add Mode A/C awareness from the same RF channel, including code/altitude facts and rate/count information, but do not invent a stable aircraft identity or position. A future synchronized multi-receiver MLAT phase is the appropriate path for non-position Mode S/A/C localization.
  • [x] ~~Fuse validated messages incrementally into capture-time aircraft state, RF summaries, and authoritative track points with out-of-order/replay-safe observation identity.~~
  • [ ] Add richer per-field freshness/quality state, source-frame navigation, and discontinuity handling. Apply source freshness per field, reject physically implausible jumps, expose stale/unknown values, and preserve source-frame references for investigation.
  • [x] ~~Publish aircraft and receiver-reception relationships through the generic map adapter. Click/select/focus must open the same Aviation investigation target in both the map and workspace; heading glyphs and trails belong in the retained GPU layer rather than React-per-aircraft animation.~~
  • [ ] Add fixed-channel 1090 detection/classification evidence: pulse preamble, symbol timing, short/long frame length, parity/CRC, message diversity, and burst rate. The band plan may propose the decoder, but valid frames confirm it.
  • [ ] Extend intelligent activation for the 1090 channel reference, capture bandwidth, tuner gain/headroom, burst-oriented persistence, loss/hold, decoder pin, cancellation, manual takeover, and undo. Double-click must reuse the known-good single decoder-start transaction.
  • [ ] Validate with pinned upstream IQ/frame fixtures, synthetic overlapping bursts, bit errors, malformed frames, address collisions, CPR edge cases, all compatible capture rates, live reception at the operator-selected site, IQ replay, dual lanes, and continuous audio/waterfall/DSP operation.

Phase 10C — 978 MHz UAT surveillance and broadcast

UAT is first-class U.S. aviation functionality. FAA guidance confirms that it is available below FL180 and is the only ADS-B link carrying FIS-B. The BSD-2-Clause dump978-fa demodulator is the leading upstream candidate and its roughly 2.08 MS/s input fits the current 2.4/3.2 MS/s capture choices after an exact-rate/channelizer qualification.

  • [ ] Add a separate allowlisted UAT browser-WASM manifest and decoder for aircraft uplinks/downlinks, basic and long messages, FEC/CRC, signal quality, and raw-frame provenance. Direct device, network-feeding, and bundled web UI code stay outside the decoder artifact.
  • [ ] Normalize direct UAT ADS-B, TIS-B, and ADS-R identity/source distinctly. Deduplicate or associate cross-link reports only with sufficient address, time, and state evidence; retain both source observations for audit.
  • [ ] Reuse the Phase 10 aircraft state, workspace, selection, history, and map model. Do not create a UAT-only aircraft store or a second map layer model.
  • [ ] Decode and retain supported FIS-B products with product identity, issue/valid/expiry time, geographic applicability, reception source, and completeness. Initial high-value products include METAR, TAF, PIREP, NOTAM, AIRMET/SIGMET, winds/temperatures, service status, and regional/CONUS weather imagery where the upstream decoder provides validated content.
  • [ ] Give FIS-B a bounded product/message surface in the Aviation workspace. Geographic weather products may later project through the existing generic map overlay architecture, but must preserve product time, coverage, resolution, missing tiles, and experimental/non-safety-certified labeling.
  • [ ] Add UAT-specific detector evidence and activation geometry rather than treating the approximately 1 MHz channel as generic broadband noise.
  • [ ] Make simultaneous 1090 + 978 operation an explicit multi-receiver plan. One 2.4/3.2 MS/s RTL-SDR cannot capture carriers separated by 112 MHz; the scheduler must acquire two eligible routes or tell the operator which link is not being monitored. Never retune an owned/manual lane silently.
  • [ ] Validate aircraft and ground frames, FEC failure, TIS-B/ADS-R source semantics, FIS-B text and imagery, stale products, map/replay behavior, and live local reception without turning weather output into flight guidance.

Phase 10D — VHF ACARS and VDL Mode 2 messaging

VHF messaging is the next local-value group. Conventional ACARS offers a smaller demodulation step; VDL2 supplies richer and likely higher-volume modern traffic. Both feed the same application/message intelligence through the MIT-licensed libacars. acarsdec and GPL-3-or-later dumpvdl2 are upstream research candidates subject to exact revision and dependency review.

  • [ ] Add a multi-channel VHF ACARS decoder for configured regional channels that fit within one capture. Preserve channel/frequency, direction, mode, block/message number, label, acknowledgement, registration/flight ID, parity/error correction, raw text, and multipart state.
  • [ ] Add VDL2 physical/link decoding for one or more 25 kHz channels within the captured span, beginning with the widely used 136.975 MHz common signalling channel and operator-configured regional channels. Preserve burst power/noise/SNR, frequency offset, header/FEC corrections, AVLC direction, ground-station identity, and raw frame.
  • [ ] Route validated ACARS-over-AVLC and supported X.25/CLNP/COTP/ATN payloads into the same application layer as VHF ACARS. Support only the CPDLC, ADS-C, MIAM, media-advisory, and related message families proven by the pinned libacars revision.
  • [ ] Reassemble multipart messages in capture order with bounded lifetime and memory. Preserve missing fragments and raw components; never fabricate a complete message or execute content carried in a message.
  • [ ] Render all free-form text and structured values as untrusted data. Bound lengths/depth/rate, escape content, and do not auto-open URLs, upload messages, or infer a flight route that was not actually decoded or supplied by an approved local metadata source.
  • [ ] Treat aircraft/ground-station identities as message participants, not as proof that either endpoint was at a particular location. An ADS-C/CPDLC position report may create explicitly sourced geographic evidence at its capture time; a registration, ground-station name, or airline designator may not.
  • [ ] Add ACARS MSK/burst and VDL2 D8PSK/symbol/framing evidence to the existing detector/classifier and intelligent-activation policy. Multi-channel start must remain one bounded decoder job fed from the existing capture fan-out.
  • [ ] Validate with upstream raw-IQ/frame fixtures, multipart and malformed payloads, simultaneous channels, local VHF reception, message search, selection, replay, and active audio/waterfall/DSP workloads.

Phase 10E — HFDL and long-range messaging

HFDL is valuable after the VHF application layer exists because it carries ACARS/AOC, CPDLC, ADS-C, and MIAM over long-range HF. The installed hf_receiver plus verified 125 MHz Ham It Up path is relevant, although propagation, weak signals, strong adjacent HF emitters, and current channel schedules make reception more variable than local 1090/VHF traffic.

  • [ ] Qualify a pinned GPL-3-or-later dumphfdl signal/application subset for browser WASM before considering a native broker. Its documented 250 kS/s multichannel example fits the current receiver, but only channels inside one actual capture and analog passband may run together.
  • [ ] Keep configured/updated HFDL ground-station and active-frequency data in a versioned local artifact with source and age. Frequency schedules are routing hints, never protocol confirmation.
  • [ ] Expose HFDL slot/burst, ground station, aircraft ID, direction, frequency, signal quality, LPDU/SPDU/MPDU validation, message reassembly, and shared libacars application content in the Aviation workspace.
  • [ ] Add HFDL waveform/framing evidence, carrier-offset handling, channel selection, and multi-channel admission to detector/activation policy. Keep HF tuner gain and strong-signal headroom measurement within the accepted staged-gain path.
  • [ ] Project a message-carried ADS-C position only with its message time and provenance. Sparse reports form discrete historical evidence/trail points, not a continuously measured trajectory.
  • [ ] Validate day/night propagation cases, weak and adjacent strong signals, multiple channels, gaps, clock epochs, multipart application messages, IQ replay, and failure isolation on the existing HF route.

Phase 10F — Specialized and hardware-gated aviation signals

  • [ ] Add SELCAL paired-tone detection as an audio-domain adjunct for selected VHF/HF aviation voice channels. A code may be correlated with independently known aircraft/message context but must not create a definitive aircraft identity by itself.
  • [ ] Reassess FLARM/FAMP after the surveillance and workspace foundations are stable. FLARM now publishes a receive-only FAMP Public Protocol for non-commercial research/private use, but GPL redistribution and exact terms still require a legal/license gate. North American operation spans approximately 902.2–927.8 MHz and uses time-dependent channel hopping, much wider than one current 2.4/3.2 MS/s capture; decoder acceptance therefore also requires a workable hopping/timing strategy and a suitable 915 MHz antenna path. See the FLARM public-protocol announcement and current official specification/license before implementation.
  • [ ] If FLARM is qualified, model its identity and position as an aviation provider and feed the same aircraft selection, history, map, trail, and replay contracts. Do not depend on or upload to OGN during normal operation.
  • [ ] Evaluate receive-only 1030 MHz Mode S/TCAS/SSR interrogation awareness for radar/interrogator activity, coverage, and RF analysis. Do not represent an interrogation as an aircraft position or promise multilateration from one unsynchronized receiver.
  • [ ] Evaluate VOR/ILS identifiers and DME/TACAN pulse-pair awareness as infrastructure/signal intelligence. Clearly separate decoded station facts from navigation solutions and label all output unsuitable for navigation.
  • [ ] Keep Inmarsat AERO Classic and STD-C deferred until a suitable L-band antenna, filter/LNA, receiver bandwidth, open decoder, and redistribution path are proven. If implemented, application messages should reuse the same aircraft/message model and libacars layer.
  • [ ] Keep Wi-Fi/Bluetooth Remote ID and other 2.4/5 GHz systems outside the current RTL-SDR promise. They require different hardware and capture paths; a future adapter may join the aviation model only after that architecture is explicitly approved.

Phase 10G — Aviation investigation and analytics

These are deliberate follow-ons after reliable decoding and capture-time state exist. They must use retained local observations, not an Internet flight- tracking service masquerading as RF Sentry evidence.

  • [ ] Add complete aircraft/history investigation across addresses, registrations, flight IDs, metadata generations, messages, tracks, receivers, frequencies, signal quality, and source provenance.
  • [ ] Add operator, registration, aircraft type/category, altitude, range, speed, receiver/path, protocol, validation, and recency filters. Support a direct aircraft within X miles view using the existing receiver geodesy.
  • [ ] Add factual emergency/squawk alerts, rapid-descent or unusual-state indicators, and new/rare/interesting-aircraft rules with explicit operator configuration and bounded notification rate. Never call an anomaly an emergency or violation without the corresponding decoded evidence.
  • [ ] Add track and message-history comparison, dwell/route summaries where supported by retained positions, takeoff/landing-area association with clear inference labels, and export of source-linked observations.
  • [ ] Add per-receiver RF coverage analysis: maximum range by azimuth/altitude, message/frame rate, CRC/FEC performance, signal strength, gaps, antenna/path comparison, and time-of-day trends. Do not mistake aircraft availability for calibrated receiver sensitivity.
  • [ ] Consider synchronized multi-node MLAT only as a separate later phase with measured clock/position uncertainty, at least the required independent receivers, calibration, residual/error presentation, and no use of ordinary browser arrival timestamps.
  • [ ] Add cross-protocol correlation carefully: 1090/UAT addresses, ACARS/VDL2/ HFDL flight IDs, registration metadata, and position/time agreement may produce an explainable association, never an irreversible identity merge.

Detection, activation, replay, and map acceptance contract

Every Phase 10 decoder slice must extend, rather than bypass, the same shared contracts:

  • [ ] Add band/channel priors and waveform-specific evidence to the existing candidate analyzer. Burst-oriented aviation signals need appropriate persistence and recurrence rules; they must not be discarded merely because a continuous parent envelope is absent.
  • [ ] Publish detector confidence separately from decoder validation. Valid frames can confirm protocol; failed or absent decodes preserve unknown rather than proving the signal is not aviation.
  • [ ] Add decoder recommendations and supported activation geometry to the existing committed-analysis handoff. Manual decoder start always remains available; automatic start remains an explicit intelligent-activation or future scan-policy action.
  • [ ] Keep receiver selection and concurrency honest. 1090, 978, VHF data, and HF data occupy separate captures; two receivers can cover at most two incompatible spans simultaneously, and antenna/path compatibility matters.
  • [ ] Replaying retained IQ must run through the same decoder and aviation ingestion path with replay-stable deduplication. Historical map/workspace state must be reproducible as of the cursor without rewriting newer live evidence or retuning the physical receiver.
  • [ ] Map position, heading, trail, recency, relationship activity, and receiver range/bearing must use capture-time evidence and MapReferenceTime. A future position, metadata update, or later flight ID may not leak backward during replay.

Phase 10 validation and exit gates

  • [ ] Pin exact upstream revisions, licenses, notices, patches, source hashes, browser assets, and golden fixtures for every decoder and metadata provider. Audit transitive code/data licenses independently.
  • [ ] Compare each browser decoder with its authoritative upstream tool over identical raw IQ/frames. Record recall, false frames, corrected bits, state accuracy, decode latency, CPU, RAM, queue pressure, and discontinuity recovery rather than accepting a few plausible rows.
  • [ ] Fuzz RF frames, application payloads, multipart assembly, metadata rows, free-form text, and binary weather products. Bound nesting, text, objects, rate, track history, and corrupt-input work.
  • [ ] Validate capture-time ordering, out-of-order ingestion, repeated IQ replay, address reassignment/ambiguity, stale fields, missing metadata, provider refresh/failure, incomplete history, and atomic generation replacement.
  • [ ] Validate workspace/map shared selection, map-to-aircraft investigation, aircraft-to-map focus, receiver distance/bearing, trails, filters, 2D/globe, reference-time rewind/play, and no future-state leakage.
  • [ ] Validate simultaneous high-rate aircraft traffic with waterfall, dB meter, audio, detection, candidate analysis, decoder work, IQ DVR, FCC enrichment, and the map active. Optional aviation consumers must drop or degrade explicitly before they disrupt realtime receiver/audio paths.
  • [ ] Perform live acceptance at the operator-selected site for 1090, 978, ACARS, and VDL2 when propagation/traffic permits, plus controlled fixtures for every phase. HFDL and specialized modes require their own path-appropriate live evidence.
  • [ ] Reconcile AGENTS.md, README/operator guidance, third-party notices, updater/runbook behavior, and the completion ledger only after each accepted slice is actually deployed and validated.

Phase 11 — SATCOM reception and products

SATCOM is a protocol-neutral Decoder Workspace domain alongside Aviation and station intelligence. It reuses the existing serial-routed receiver lanes, shared IQ ingress, rational WASM channelizer, isolated decoder workers, DVR capture-time semantics, and browser-local evidence retention. A SATCOM module must never open a competing receiver connection or silently retune a lane.

Phase 11A — GOES HRIT receive and alignment foundation

The accepted initial receive checkpoint targets the GOES-R HRIT/EMWIN channel at 1694.1 MHz. NOAA's 927 ksymbol/s BPSK, rate-1/2 convolutional, four-way RS(255,223) geometry is manifested as a 1.854 Msps complex-IQ worker input. Frequency remains a target/configuration fact; only validated protocol output establishes HRIT reception and this tranche does not claim spacecraft identity.

  • [x] ~~Evaluate pinned goestools, libcorrect/libfec, and SatDump receive-chain implementations for correctness, licensing, and browser/WASM suitability.~~
  • [x] ~~Add a bounded worker-WASM acquisition path with complex DC removal, carrier/timing recovery, encoded-ASM correlation, soft Viterbi, HRIT NRZ-M, CCSDS derandomization, dual-basis conversion, and four-way Reed-Solomon frame validation.~~
  • [x] ~~Exercise the production channelizer and decoder with compact generated raw IQ covering valid CADU, carrier/DC offset, noise, corrupt integrity, and source discontinuity. Qualify separately against an independently recorded 5 Msps OTA HRIT capture without redistributing that unlicensed recording.~~
  • [x] ~~Add compact lane/workspace navigation with an extensible SATCOM → GOES module while preserving Aviation, FT8/FT4, NRSC-5, station intelligence, decoder state, and pinned-signal access.~~
  • [x] ~~Add continuous dish-alignment trends and explicit 8-second, 30-second, 2-minute, and 10-minute cancellable tests. Report dBFS/FFT contrast, clipping/continuity, synchronization, validated/corrected/uncorrectable frames, lock loss, reacquisition, configuration changes, limitations, and justified next steps.~~
  • [x] ~~Make RTL-SDR tuner gain capability-specific. The gateway learns the tuner type/count from the RTL-TCP header, publishes E4000 versus R82xx steps, rejects known unsupported settings, and distinguishes accepted commands from unavailable gain readback and from tuner/digital/browser AGC.~~
  • [x] ~~Demonstrate local OTA HRIT-format lock on satcom1, including validated CADUs, live dBFS/contrast/clipping diagnostics, a complete two-minute report, counter-reset-safe evidence, and a cancelled partial reliability report. Frequency remains insufficient to assert GOES-19 identity.~~
  • [x] ~~Accept the lane-local capture-center refinement: residual software mean correction precedes VFO translation without rewriting raw IQ; supported E4000 driver offset tuning prevents the measured broad zero-IF pedestal while preserving logical tuning and HRIT bandwidth; capture metadata and reports distinguish both settings, unavailable classic-RTL-TCP readback, frame/FEC evidence, and transport drops.~~
  • [ ] Complete operator acceptance, extended reliability, reconnect/restoration, and concurrent receiver-isolation validation while spectrum, waterfall, DVR, map, audio, and another lane remain responsive.

The detailed evidence, open-source revisions, measurement semantics, and test commands live in satcom-goes-hrit.md. This is a receive/readiness foundation, not completion of GOES product support.

Phase 11B — HRIT/LRIT products and investigation

The 2026-11-07 publication approval covers workspace, scenery and radar corrections, not blanket Phase 11 acceptance. Faster native telemetry remains deferred (1 Hz native; Direct IQ diagnostics up to 4 Hz). The 2026-11-05 headless scenery smoke could not create an EGL/WebGL context; mobile/workstation rendering, GPU resources and sustained performance remain unqualified by it. The 2026-11-06 responsive fixture covered 1440/768/390 px, source disclosures, keyboard inspection and retained content, not installed/live reception. The replacement field recorder presents decoded observations in a shared sensor timeline and links samples to their DCS messages. Its interactive preview uses synthetic data. Operator acceptance and sustained performance during reception remain open for this replacement.

This acceptance candidate turns RS-validated CADUs into a reusable receive-to-product path. GOES-19 is the current target and live qualification source, while receiver, lane, configured target, service profile, decoded relay identity, originating product source, and acquisition session remain separate. Only the independently qualified NOAA GOES-R HRIT/EMWIN RF profile is enabled; other HRIT/LRIT profiles remain explicit future work.

Candidate validation on SATCOM1 includes a clean ten-minute live interval with 10,768 validated CADUs and no transport, decoder-input, or product-handoff drops; 208 complete live EMWIN products persisted across a production reload. Independent goestools/libaec/NOAA evidence qualifies exact file bytes, Rice imagery, graphics, DCS, and text handlers. Live imagery channels were observed mid-file, so a live file-start-to-completion image and operator acceptance remain open rather than being inferred from independent fixtures.

  • [x] ~~Route exact validated CADUs directly from decoder WASM into a bounded product worker without generic-media or React-per-frame traffic.~~
  • [x] ~~Demultiplex VCDUs/M_PDUs and reconstruct CRC-checked CCSDS packets and HRIT/LRIT files across frame boundaries, interleaved channels, sequence wrap, duplication, loss, reconnects, decoder sessions, and replay.~~
  • [x] ~~Keep simultaneous lanes and live/replay acquisition sessions isolated; never combine file or image segments solely because their protocol identifiers match.~~
  • [x] ~~Process uncompressed and CCSDS-Rice imagery, segmented mosaics, partial-image missing lines, EMWIN ZIP/text/graphics, administrative text, and documented DCS transport records while retaining unknown originals.~~
  • [x] ~~Provide a product-first SATCOM workspace with overview, searchable archive, imagery viewer/animation, weather/message reader, DCS records, details/provenance, original and rendered exports, deletion, and honest empty/unsupported/incomplete states.~~
  • [x] ~~Provide bounded Acquisition Activity with stable keyed rows, per-channel/file stages, exact-or-unknown progress, unique versus wire bytes, packet/frame/segment loss, reasons, source-time rates, replay-processing rates, and presentation suspension while hidden.~~
  • [x] ~~Persist original reconstructed files and derived previews in browser-local OPFS with IndexedDB metadata/fallback, content deduplication plus reception history, configurable retention/budget, and bounded memory, decompression, queues, images, and assemblies.~~
  • [x] ~~Qualify deterministic raw IQ through the production channelizer, demodulator, FEC, CADU, LRIT, and product stack; compare published VCDU recordings byte-for-byte with pinned goestools output and exercise an upstream libaec Rice vector.~~
  • [ ] Complete operator acceptance of browsing/export/deletion/activity UX, observe a live imagery file from start through completion, and explicitly exercise concurrent second-receiver isolation on the required SATCOM1 settings. Sustained live weather-product completion, offline reopen, open/hidden-view performance, and the ten-minute receive interval are technically validated.
  • [ ] Add optional product-aware map overlays and supported multispectral enhancements without converting Internet products into received evidence.
  • [ ] Implement and independently qualify additional approved HRIT/LRIT RF service profiles; architecture reuse alone is not a support claim.