Detection Pipeline
Four stages transform raw radio signals into verified detection candidates.
KLT Signal Detection
Eigenvalue decomposition of the signal's covariance matrix. If a narrowband signal is present, the first eigenvalue dominates—the eigengap. The dominant eigenvector is then used to reconstruct a denoised signal, which is accumulated into a waterfall for drift analysis in the next stage.
Drift Analysis
The waterfall built from KLT-reconstructed spectra is analyzed by Radon transform to detect linear drift. Drift rate, stability, and physical plausibility are evaluated as classification indicators.
Classification
Cascade logic integrates KLT and drift results. Conservative thresholds minimize false positives.
turboSETI Verification
ETI candidates are automatically verified using Breakthrough Listen's turboSETI. Same tool, same format—independently reproduced.
Storage & Sharing
Detections archived locally (SigMF + FilterBank) and recorded to PostgreSQL. Stations gossip signed observations directly to peer Stations over a P2P overlay — no central server, no privileged operator.
P2P Network
No central authority. Stations sign their own observations and gossip directly to peers. Anyone can join as an equal, anonymously.
Each Station holds a self-generated Ed25519 identity. Observations are signed and propagated Station-to-Station via gossip — no registry, no privileged operator, no server that can be silenced to stop the listening. Tor-onion transport keeps operator IPs private. ETI-candidate observations are auto-mirrored network-wide so the evidence outlives any single Station. Browser viewers connect through voluntary aggregator Nodes that expose the gossip stream; if one aggregator falls, another can take its place.
Open spec: network-spec.md.
Solar System SETI
If an extraterrestrial intelligence has sent probes to our solar system, where would they park? We search for their signals.
The Hypothesis
An advanced civilization might explore the galaxy not by broadcasting radio signals, but by sending autonomous probes to nearby star systems. A Bracewell probe would park in a stable orbit and wait—potentially for millions of years—until a technological civilization emerges.
Self-replicating probes could explore the entire galaxy within a few million years by building copies of themselves from local materials, populating every star system in a timespan far shorter than the galaxy's age.
The Lurker hypothesis [1] asks: if such probes exist in our solar system, where would they be? The most likely locations are dynamically stable niches near Earth—horseshoe orbits, quasi-satellites, and Earth Trojans at the L4/L5 Lagrange points. These are gravitational niches where an artifact could persist indefinitely without expending energy.
Starglider implements this search. We compute the full N-body gravitational dynamics of the solar system, identify stable co-orbital objects from the MPC asteroid catalog, classify their resonant behavior (horseshoe, quasi-satellite, or Trojan via the Namouni–Morais method), and point our antennas at the most promising candidates. We also target planets, the Moon, and near-Earth asteroids with stable orbits (MEGNO ≈ 2)—any location where an artifact might have been placed.
Pipeline
Catalog & Classification
Load 1.5M asteroids from the MPC catalog. Filter by target class (NEA, co-orbital, planet). For co-orbitals, classify resonant type from N-body relative longitude analysis (horseshoe, quasi-satellite, Trojan L4/L5).
Stability & Observability
Full N-body MEGNO integration (Yoshida4, JPL GM, 8 planets + Moon) evaluates orbital stability. Observable candidates are filtered by solar elongation and elevation at the observer's site. N-body ephemeris verifies precise RA/Dec.
Targeted Observation
The selected target enters Starglider's tracking mode. N-body ephemeris provides predicted RA/Dec. Solar system targets have very low Doppler drift (co-orbital objects nearly co-move with Earth), so detection uses ON/OFF switching rather than drift search: observe the target (ON), slew away (OFF), then difference to isolate any signal present only at the target position.
References
C. Maccone, "The KLT (Karhunen-Loève Transform) to extend SETI searches to broad-band and extremely feeble signals",
Acta Astronautica, 67(11–12), 1427–1439, 2010
M. Trudu et al., "Performance analysis of the Karhunen-Loève Transform for artificial and astrophysical transmissions: denoizing and detection",
MNRAS, 494(1), 69–83, 2020
— arXiv:2003.04243
S. Sheikh et al., "Choosing a Maximum Drift Rate in a SETI Search: Astrophysical Considerations",
ApJ, 884(1), 14, 2019
— arXiv:1910.01148
J.-L. Margot et al., "A Search for Technosignatures from 14 Planetary Systems in the Kepler Field with the Green Bank Telescope at 1.15–1.73 GHz",
AJ, 155(5), 209, 2018
— arXiv:1802.01081
M. Lebofsky et al., "The Breakthrough Listen Search for Intelligent Life: Public Data, Formats, Reduction, and Archiving",
PASP, 131, 124505, 2019
— arXiv:1906.07391
M. L. Brady, "A Fast Discrete Approximation Algorithm for the Radon Transform",
SIAM J. Comput., 27(1), 107–119, 1998
G. Nir, B. Zackay, E. O. Ofek, "Optimal and Efficient Streak Detection in Astronomical Images",
AJ, 156(5), 229, 2018
— arXiv:1806.04204
G. Beylkin, "Discrete Radon Transform",
IEEE Trans. Acoust., Speech, Signal Process., 35(2), 162–172, 1987
Bracewell Probes & Lurker Hypothesis
R. Bracewell, "Communications from Superior Galactic Communities",
Nature, 186, 670–671, 1960
R. Bracewell, "The Galactic Club: Intelligent Life in Outer Space",
W. H. Freeman, 1975
J. von Neumann, "Theory of Self-Reproducing Automata",
ed. A. Burks, Univ. of Illinois Press, 1966
R. Freitas, "The Search for Extraterrestrial Artifacts (SETA)",
JBIS, 36, 501–506, 1983
M. Papagiannis, "Are We All Alone, or Could They Be in the Asteroid Belt?",
QJRAS, 19, 277–281, 1978
[1] J. Benford, "Looking for Lurkers: Co-orbiters as SETI Observables",
AJ, 158(4), 150, 2019
— arXiv:1903.09582
Co-orbital Dynamics & Classification
F. Namouni, "Secular Interactions of Coorbiting Objects",
Icarus, 137(2), 293–314, 1999
M. H. M. Morais & A. Morbidelli, "The Population of Near-Earth Asteroids in Coorbital Motion with the Earth",
Icarus, 160(1), 1–9, 2002
C. de la Fuente Marcos & R. de la Fuente Marcos, "Asteroid 469219 2016 HO3, the smallest and closest Earth quasi-satellite",
MNRAS, 462(4), 3441–3456, 2016
Celestial Mechanics & N-body Methods
P. Cincotta & C. Simó, "Simple tools to study global dynamics in non-axisymmetric galactic potentials — I",
A&A Suppl., 147(2), 205–228, 2000 (MEGNO)
H. Rein & S.-F. Liu, "REBOUND: an open-source multi-purpose N-body code for collisional dynamics",
A&A, 537, A128, 2012
H. Rein & D. Tamayo, "WHFAST: a fast and unbiased implementation of the Wisdom–Holman integrator",
MNRAS, 452(1), 376–388, 2015
H. Yoshida, "Construction of higher order symplectic integrators",
Physics Letters A, 150(5–7), 262–268, 1990
R. Park et al., "The JPL Planetary and Lunar Ephemerides DE440 and DE441",
AJ, 161(3), 105, 2021