Frequently asked questions

An observer-relative airspace-intelligence platform. It ingests aircraft positions from 60+ public ADS-B feeds, fuses and de-duplicates them, and presents a real-time, explainable view of the aircraft over a chosen point — identity, behaviour, signal integrity, and which ones warrant attention. It also publishes an open ADS-B anomaly-detection benchmark.

Open the live map, set your location, and AeroScope shows the aircraft overhead with distance, bearing, identity and an attention score for each. No app or hardware required.

No. The anomaly layer is deliberately torch-free and built on established libraries — River, scikit-learn IsolationForest, pykalman, OpenAP, stumpy and PyOD — with consensus voting (at least two detectors must agree). There is no PyTorch and no reinforcement learning; an earlier RL prototype was retired.

It validates each broadcast against its integrity fields (NIC/NACp/NACv/SIL), a Kalman normalised-innovation plausibility test, and self-consistency residuals (geometric-vs-barometric altitude, ground-speed-vs-Mach, track-vs-heading). Messages that contradict physics or their own metadata are flagged. See signal integrity.

No. It is a 0–100 attention-ranker that sorts a busy list, computed from eight transparent weighted factors, all shown. It tells you where to look first; it does not pass judgement. See threat scoring.

It tracks drones that broadcast (FAA Remote ID and larger UAS) and flags low-and-slow behavioural candidates, but a fully silent consumer drone with no transponder cannot be seen by ADS-B alone. See drone detection for the honest breakdown.

It surfaces military and state aircraft that broadcast ADS-B, using ICAO hex ranges, callsign patterns and pattern-of-life. It cannot see aircraft with the transponder off or under emissions control. See military aircraft tracking.

Public ADS-B broadcasts on 1090 MHz republished by volunteer networks (adsb.fi, adsb.lol, airplanes.live, OpenSky) plus a satellite-ADS-B merge — 60+ feeds, de-duplicated by ICAO 24-bit address. See multi-source fusion.

Yes — the AeroScope ADS-B Anomaly Benchmark v1, CC-BY 4.0, with 38 documented columns pairing real traffic with injected attacks and an IsolationForest baseline (ROC-AUC ≈ 0.87). See research & dataset.

The live platform is free in a browser, and there is a documented REST/WebSocket API plus CSV/JSON/GeoJSON exports. See the API docs.

It only aggregates publicly broadcast ADS-B, which is what many open networks already do; follow your local laws. It does not transmit, jam or interfere with any aircraft. See our privacy policy.