Sdr satellite

SDR & Satellite Reception

How to hear satellites with a $30 dongle, decode the bits, predict the next pass, and (if we want) wire it into Life Agent OS or Opsis as a real-world-event source.

What this skill covers

The full open-source satellite reception stack, by layer:

Layer	What it does	Pick
L0 Hardware abstraction	One API across dongles	SoapySDR; pure-Rust rtl-sdr-rs
L1 DSP framework	Demod, filter, sync	GNU Radio (standard), liquid-dsp (lean C), rustradio (Rust), radiorust (Tokio)
L2 Receiver UI	Watch the spectrum, demod live	SDR++, SDRangel, Gqrx, CubicSDR, SigDigger
L3 Satellite decoders	TLE-aware, multi-protocol	SatDump (90+ sats)
L4 Ground-station orchestration	Schedule, rotate, Doppler	SatNOGS, sgoudelis/ground-station (2026), Hamlib
L5 Orbit mechanics	TLE → pass times → az/el	sgp4 (Python), neuromorphicsystems/sgp4 (Rust), satkit (full astro)

Detail: references/tools-by-layer.md.

Quick start — your first weather image

About a 1-afternoon project, $50 total.

# 1. Hardware: RTL-SDR v4 ($30) + V-dipole antenna ($20 wire or 3D-printed QFH)

# 2. Software (macOS)
brew install --cask satdump
brew install gpredict     # pass prediction GUI

# 3. Find next NOAA 18/19 pass overhead
gpredict      # add NOAA 18, NOAA 19 from CelesTrak TLEs; watch for AOS

# 4. Decode (SatDump live mode, frequency 137.9125 MHz for NOAA 19)
satdump live_processing noaa_apt_demod baseband.raw \
  --source rtlsdr --frequency 137912500 --samplerate 2400000

Output: PNG weather imagery of wherever you are, georeferenced. SatDump handles APT, LRPT, HRPT, HRIT, Inmarsat Aero, and ~85 other satellite formats automatically.

Detail: references/recipes.md.

Quick start — no hardware (use someone else's)

# Visit https://network.satnogs.org/observations/new/
# Pick a station that can see the satellite you want
# Schedule an observation; the station records IQ + decodes for you
# Download the resulting waterfall, demod, decoded data via the SatNOGS API

SatNOGS is ~600 crowd-sourced ground stations worldwide. Free to schedule, public results.

Quick start — Rust pipeline (no GUI)

For embedding satellite reception into a Rust service — relevant for Life Agent OS, Opsis world-state events, or arcan sensorium:

# Cargo.toml
[dependencies]
rtl-sdr-rs = "0.x"          # hardware
rustradio = "0.x"           # DSP block graph (or radiorust for Tokio-native)
sgp4 = "1.x"                # neuromorphicsystems/sgp4 — no_std capable
satkit = "0.16"             # full astrodynamics if needed

Shape:

• rtl-sdr-rs → raw IQ stream
• rustradio flowgraph → APT/LRPT demod → frames
• sgp4 / satkit → georeference each frame against TLE
• Emit as OpsisEvent { kind: "sat_pass", sat_id, lat, lon, alt, timestamp, payload }

Detail: references/composition.md.

What you can actually hear

Satellite class	Band	What you get	Difficulty
NOAA 15/18/19 (APT)	137 MHz	Live weather imagery, ~15-min passes	★ easy
Meteor-M N2/N2-2 (LRPT)	137 MHz	Higher-res digital weather imagery	★★
ISS SSTV / APRS	145 MHz	Slow-scan images, packet radio	★★
GOES-16/18 (HRIT)	1.7 GHz	Full-disk geostationary Earth	★★★ (needs small dish)
Inmarsat Aero (STD-C, EGC)	L-band	Aviation messages, maritime safety	★★★
Iridium pager telemetry	L-band	Pager metadata; voice is encrypted	★★★
GPS L1 / Galileo E1	L-band	Your own position from raw signals	★★★★

Architectural threads (why this lives in the workspace)

Two non-forced connections to existing Broomva work:

SatNOGS ↔ Spaces. SatNOGS's distributed ground-station network has the same pattern as core/life/spaces: global registry of nodes, a scheduler that assigns work to nodes with the right capability, observations flow back to a central DB. If we ever want "Life node, but with antennas," SatNOGS already solved the rotator/Doppler/TLE-sync coordination problem on top of an open scheduling API.

Satellite passes ↔ Opsis. Opsis is already an AI-native world-state engine with OpsisEvent protocol and a CesiumJS globe. Live satellite passes are literally georeferenced world-state events. satkit (Rust SGP4 + ITRF/Geodetic transforms) → OpsisEvent { kind: "sat_pass", ... } → CesiumJS globe already renders it. Zero forced abstraction; the pieces match.

Detail: references/composition.md.

Repository layout (if we build the pipeline)

Speculative — no code shipped yet, this is the skill's compositional seed:

~/broomva/experiments/sdr-satellite/         # (proposed)
├── crates/
│   ├── sat-sdr/             # rtl-sdr-rs wrapper, hardware enumeration
│   ├── sat-decode/          # rustradio flowgraphs for APT / LRPT / SSTV
│   ├── sat-orbit/           # satkit-based pass prediction + Doppler
│   └── sat-opsis-bridge/    # decoded frames → OpsisEvent emitter
└── scripts/
    ├── first-image.sh       # The "your first NOAA image" walkthrough
    └── schedule-satnogs.sh  # Submit observation to SatNOGS network

References

• tools-by-layer.md — full L0-L5 stack survey with comparison tables
• recipes.md — concrete recipes: first NOAA APT image, SatNOGS scheduling, GOES dish setup, Meteor-M decode
• composition.md — Rust pipeline composition for Life / Opsis / arcan; SatNOGS ↔ Spaces pattern analysis

Source list

• SatNOGS — global ground-station network · Wiki · Network
• SatDump — the decode engine
• Ground Station orchestrator (sgoudelis/ground-station) · RTL-SDR.com writeup
• RTL-SDR supported software (the canonical big list)
• 15 Best SDR Software 2026 (OneSDR)
• rtl-sdr-rs crate · rustradio · radiorust
• neuromorphicsystems/sgp4 (Rust, no_std) · satkit (Rust + PyO3 astrodynamics) · python-sgp4 · heyoka.py differentiable SGP4