Projects — Sunil Tyagi: vessel recognition and underwater AI

National-security AI challenge · NCIIPC AI Grand Challenge (PS12)

Underwater Acoustic Event Detection & Vessel Re-identification

End-to-end deep-learning system for underwater acoustic event detection, multi-class classification and vessel re-identification, operating directly on raw passive-sonar waveforms through a learned, multi-scale selective-kernel front end: SKANN, the Selective-Kernel Acoustic Neural Network developed by Sunil Tyagi at Oravont. Built by Oravont Systems as technical partner to an industry participant in India’s national-security AI Grand Challenge (National Critical Information Infrastructure Protection Centre, NCIIPC / Startup India), it delivered strong, well-calibrated classification across acoustically diverse classes.

SKANN: selective-kernel acoustic neural network architecture, raw waveform to classification

PS12: underwater acoustic event detection and classification work sample

SKANN selective-kernel front end (left) and the PS12 detection & classification task (right).

Oravont Systems, technical partner to an industry participant · 2025 — May 2026 · Concluded

iDEX DISC5, Challenge 10 (Indian Navy, Underwater Domain Awareness) · Ministry of Defence

Automated Vessel Recognition from Passive Sonar

Oravont Systems is technical partner to an industry participant in iDEX DISC5 Challenge 10 (Defence India Startup Challenge 5), an Indian Navy underwater domain awareness challenge, building the deep-learning core for automated recognition of vessels from passive-sonar signatures: compact acoustic fingerprints enabling match-or-flag-as-new decisions, designed for in-house, frozen-model deployment with no data egress and no retraining.

DISC5: vessel re-identification from acoustic fingerprints work sample

The enrolment-and-inference application is public on GitHub; the training code and its documentation are held privately under the programme. Enquiries welcome through the Contact page.

Oravont Systems, technical partner to an industry participant · Programme active; Oravont’s part winding down

Further reading: Re-identifying vessels from radiated noise, a working note on embeddings, galleries and enrolling new vessels without retraining.

Exploratory academic collaboration · IIT Delhi

Propagation-Robust Acoustic Vessel Re-identification

Exploratory research collaboration between Oravont Systems and IIT Delhi, extending acoustic vessel re-identification through deep metric learning and physics-based acoustic-propagation modelling, in association with a DRDO/NSTL programme (Defence Research and Development Organisation / Naval Science and Technological Laboratory). Nothing about the collaboration is finalised.

Oravont Systems with IIT Delhi · Exploratory

Generative AI · iDEX

ShaktiVarta.AI — Generative AI for Naval Design Knowledge

Large language model (LLM) + retrieval-augmented-generation (RAG) solution for knowledge extraction in naval design and construction, submitted through Oravont Systems under iDEX open challenges, with Capt (Dr) Sunil Tyagi as Principal Investigator.

Oravont Systems · Sunil Tyagi, Principal Investigator · Under evaluation

Acoustic stealth · iDEX

Active Noise Control Silencer for Submarine Exhaust (CMAC-based)

iDEX proposal, submitted through Oravont Systems, for an Active Noise Control silencer for submarine exhaust using control based on a cerebellar model articulation controller (CMAC) — a direct acoustic-stealth application.