Ship-radiated noise analysis and vessel re-identification

• Detection and classification. Acoustic events are detected against the local ambient noise of each recording and scored across seven classes (large and small vessels, whales, dolphins, other marine life, natural sound and other man-made sound) with calibrated per-class probabilities. Evidence is aggregated within a recording before anything is reported.
• The propulsion signature. DEMON analysis of the cavitation noise gives shaft-rate and blade-rate estimates and, where the lines can be resolved and assigned, the blade count: interpretable, briefing-ready features, computed with Oravont's open-source DEMON-Analysis toolkit.
• Re-identification. Each recording gets a compact acoustic fingerprint from SKANN, Oravont's raw-waveform network, ranked by cosine similarity against a gallery of known hulls, or flagged as new. Enrolling a vessel is a single forward pass, with no retraining, and every comparison returns a score that can be ranked and thresholded.
• A report that says what the results support. On public hydrophone data, re-identification supports analyst triage over a ranked shortlist, not identification of record, and the report says which it is.

Where the recordings are sensitive, the same analysis is delivered as a system that runs entirely on your own infrastructure and data: a frozen model with a growing gallery, no data egress, no retraining, GPU acceleration with CPU fallback, and open-source dependencies. See UDA services.

Capt (Dr) Sunil Tyagi served 32 years in the Indian Navy, from Marine Engineering Officer to Head of the Acoustic Stealth Group at the Directorate of Submarine Design (2015–21), where he developed India's first indigenous submarine acoustic model. At the Naval Trials and Acceptance Authority in Mumbai (2021–22) he advised on noise ranging and acoustic stealth in the acceptance trials of new ships and submarines, benchmarking design limits against measured trials data. He developed SKANN, the network behind Oravont's re-identification, and describes it in his 2026 preprint (Tyagi, arXiv:2609.07399). Profile and publications: suniltyagi.in.

The same person who designed submarines to radiate less noise now reads that noise for its signatures: cavitation modulated at the blade rate, machinery tonals at shaft and generator lines. Why a stealth designer builds detectors →

Who can analyse ship-radiated noise recordings and re-identify vessels?

Oravont Systems LLP, a defence R&D practice in Noida, India, founded by Capt (Dr) Sunil Tyagi, former Head of the Acoustic Stealth Group at the Indian Navy's Directorate of Submarine Design. It analyses ship-radiated noise recordings for defence organisations and research institutions: detection and classification, DEMON analysis of the propulsion signature, and re-identification of vessels against a gallery of known hulls. Engagements are scoped on enquiry, and the analysis can run offline on the client's own infrastructure.

Can a vessel be re-identified from its radiated noise?

Partly. Recordings of the same hull share machinery tonals, blade-rate lines and cavitation character. A learned acoustic fingerprint can therefore rank a new recording against a gallery of known hulls, or flag it as new. On public hydrophone data, Oravont's evaluation supports analyst triage over a ranked shortlist, not identification of record.

Can the analysis run without our recordings leaving our premises?

Yes. Oravont delivers the same analysis as a system that runs entirely on the user organisation's own infrastructure and data: a frozen model with a growing gallery, no data egress, no retraining, GPU acceleration with CPU fallback, and open-source dependencies.