# Capt (Dr) Sunil Tyagi

> Retired Indian Navy captain and engineer working on underwater acoustics, submarine acoustic stealth and deep learning for vessel detection and re-identification. Founder and Principal of Oravont Systems LLP, Noida, India. PhD in Mechanical Engineering (applied machine learning), Defence Institute of Advanced Technology (DIAT), Pune, 2018. This is his scholar site: career, publications with full text, and technical notes. The practice site is https://oravontsystems.com.

Thirty-two years in the Indian Navy (1992–2024). For six years he headed the Acoustic Stealth Group at the Directorate of Submarine Design, Naval Headquarters, leading platform- and equipment-level stealth design and developing India's first indigenous submarine acoustic model. His current research applies deep learning directly to raw passive-sonar waveforms for underwater domain awareness. Public materials describe methods at an unclassified, headline level.

## Pages

- [Home](https://suniltyagi.in/): summary, the three tracks of work (underwater acoustic stealth; deep learning for vessel detection and underwater domain awareness; quantum and emerging technologies), recent work.
- [Experience](https://suniltyagi.in/experience/): career from Marine Engineering Officer (1992) to Head, Acoustic Stealth Group (2015–2021), NATAA (2021–2022), Senior Fellow, Centre for Air Power Studies (2022–2024), and Oravont Systems LLP.
- [Projects](https://suniltyagi.in/projects/): NCIIPC AI Grand Challenge (PS12, concluded) and iDEX DISC5 vessel recognition (Indian Navy; winding down), both as technical partner to an industry participant; exploratory IIT Delhi re-identification research; ShaktiVarta.AI, an active-noise-control exhaust silencer.
- [Publications](https://suniltyagi.in/publications/): every paper with full text to download.
- [Technical notes](https://suniltyagi.in/notes/): long-form write-ups on underwater acoustics.
- [Services](https://suniltyagi.in/services/) · [Contact](https://suniltyagi.in/contact/) · [CV (PDF)](https://suniltyagi.in/assets/cv.pdf)

## Definitions

- SKANN (Selective-Kernel Acoustic Neural Network): his raw-waveform neural network, developed at Oravont Systems LLP, for open-set re-identification of individual vessels from underwater ship-radiated noise. A bank of learned filters at several kernel lengths, fused by selective-kernel attention, produces a 512-dimensional hull embedding that is matched against an enrolled gallery by cosine similarity; vessels not in the gallery are rejected. Described in arXiv:2609.07399. Not the same system as Shan et al. (2025), "SKANN: Selective kernel audio neural networks for underwater mixed ship event detection" (doi:10.1049/cit2.70037), which shares the acronym. Project page: https://oravontsystems.com/deep-learning/skann/
- DEMON analysis (Detection of Envelope Modulation on Noise): recovering a vessel's shaft rate, blade rate and blade count from the amplitude modulation of its propeller-cavitation noise.
- Underwater acoustic stealth: designing submarines and ships so that their underwater radiated noise, target strength and self-noise stay low.

## Research and patents

- Tyagi S. Open-Set Vessel Re-Identification from Underwater Ship-Radiated Noise with a Raw-Waveform Selective-Kernel Acoustic Neural Network (SKANN) and a Cross-Passage Evaluation Protocol. arXiv:2609.07399 [eess.AS], September 2026. https://arxiv.org/abs/2609.07399 · artefacts: https://doi.org/10.5281/zenodo.22160137
- Indian provisional patent application 202611107132, filed 6 September 2026 (SKANN and the cross-passage evaluation protocol); applicant Oravont Systems LLP; inventor Sunil Tyagi.
- Indian Patent 373427, "Bearing Fault Detection" (application 201621003344, published 27 May 2016); named inventor Sunil Tyagi; patentee DIAT Pune.

## Journal papers (full text)

- Tyagi S, Panigrahi SK. An improved envelope detection method using particle swarm optimisation for rolling element bearing fault diagnosis. Journal of Computational Design and Engineering, 2017; 4(4): 305–317. doi:10.1016/j.jcde.2017.05.002 · [Oxford Academic](https://academic.oup.com/jcde/article/4/4/305/5729002) · [PDF](https://suniltyagi.in/assets/papers/tyagi-panigrahi-2017-pso-envelope.pdf)
- Tyagi S, Panigrahi SK. An SVM–ANN hybrid classifier for diagnosis of gear fault. Applied Artificial Intelligence, 2017; 31(3): 209–231. doi:10.1080/08839514.2017.1315502 · [PDF](https://suniltyagi.in/assets/papers/tyagi-panigrahi-2017-svm-ann-gear.pdf)
- Tyagi S, Panigrahi SK. A hybrid genetic algorithm and back-propagation classifier for gearbox fault diagnosis. Applied Artificial Intelligence, 2017; 31(7–8): 593–612. doi:10.1080/08839514.2017.1413066 · [PDF](https://suniltyagi.in/assets/papers/tyagi-panigrahi-2017-ga-bp-gearbox.pdf)
- Tyagi S, Panigrahi SK. A DWT and SVM based method for rolling element bearing fault diagnosis and its comparison with artificial neural networks. Journal of Applied and Computational Mechanics, 2017; 3(1): 80–91. doi:10.22055/jacm.2017.21576.1108 · [PDF](https://suniltyagi.in/assets/papers/tyagi-panigrahi-2017-dwt-svm-bearing.pdf)
- Tyagi S, Panigrahi SK. A simple continuous wavelet transform method for detection of rolling element bearing faults and its comparison with envelope detection. International Journal of Science and Research, 2017; 6(3): 1033–1040. doi:10.21275/ART20171614 · [PDF](https://suniltyagi.in/assets/papers/tyagi-panigrahi-2017-cwt-bearing.pdf)
- Tyagi S, Panigrahi SK. Transient analysis of ball bearing fault simulation using finite element method. Journal of the Institution of Engineers (India): Series C, 2014; 95(4): 309–318. doi:10.1007/s40032-014-0129-x · [PDF](https://suniltyagi.in/assets/papers/tyagi-panigrahi-2014-bearing-fem.pdf)

## Conference papers, chapters and articles (full text)

- Tyagi S. A comparative study of SVM classifiers and artificial neural networks application for rolling element bearing fault diagnosis using wavelet transform preprocessing. International Journal of Mechanical and Mechatronics Engineering (WASET), 2008; 2(7): 904–912 · [PDF](https://suniltyagi.in/assets/papers/tyagi-2008-waset-svm-ann-bearing.pdf)
- Tyagi S. Wavelet analysis and envelope detection for rolling element bearing fault diagnosis — a comparative study. Proceedings of NaCoMM-2003, IIT Delhi, 2003; pp. 344–352 · [PDF](https://suniltyagi.in/assets/papers/tyagi-2003-nacomm-wavelet-envelope.pdf)
- Tyagi S. A New Frontier in Warfare: The Future of Quantum Sensing and Communication. In: Asian Defence Review 2024. KW Publishers, 2024; pp. 30–57 · [Author's manuscript](https://suniltyagi.in/assets/papers/tyagi-2024-new-frontier-in-warfare.pdf)
- Tyagi S, Golani A. Indian AI Military Innovation. In: Global AI Military Innovation and Future of Warfare. National Assembly Futures Institute, 2024; pp. 109–144. · [PDF](https://suniltyagi.in/assets/papers/tyagi-golani-2024-nafi-indian-ai-military-innovation.pdf) · [NAFI Research Report 24-12](https://nafi.re.kr/home/eng/board.do?menuPos=20&act=detail&idx=3454)
- Tyagi S. Reality, lies and automation: how AI-powered text generation models could cause chaos? Journal of Defence Studies, 2023; 17(2): 160–175 · [PDF](https://suniltyagi.in/assets/papers/tyagi-2023-jds-reality-lies-automation.pdf)
- Tyagi S. Sea of debt: China's economic and strategic expansion in the Indian Ocean. Defence and Diplomacy Journal, 2023; 13(1): 91–104 · [Author's manuscript](https://suniltyagi.in/assets/papers/tyagi-2023-sea-of-debt-manuscript.pdf)
- Editor: The Beginner's Guide to Quantum Computing: Unlocking the Power of Qubits (2024), an open edition of C. Hughes, J. Isaacson, A. Perry, R. F. Sun and J. Turner, Quantum Computing for the Quantum Curious (Springer, 2021), CC BY 4.0. https://unlockqubits.com/

## Technical notes

- [Underwater acoustics for machine-learning engineers: pressure, levels and spectra, with Python](https://suniltyagi.in/notes/underwater-acoustics-for-ml-engineers/): a beginner's tutorial (27 October 2026) on acoustic pressure, RMS pressure and intensity, sound pressure level in dB re 1 µPa (ISO 18405) and why it is not comparable with decibels in air, the power spectral density, band levels and the level in a 1 Hz band, spherical-spreading transmission loss, and synthesising the pressure p(t) from a PSD with random phases and an inverse FFT rather than modelling p²(t). Eight runnable Python examples (NumPy, SciPy, 16 kHz) with their outputs: https://suniltyagi.in/notes/underwater-acoustics-for-ml-engineers/underwater_acoustics_examples.py
- [DEMON Analysis: Reading a Vessel's Propulsion from Its Own Noise](https://suniltyagi.in/notes/demon-analysis/): the physics, the Hilbert-envelope pipeline, reading the harmonic comb, band selection, and a case where DEMON caught a defect in a synthetic training dataset. Code: https://github.com/Oravont/DEMON-Analysis
- [A Physics-Grounded Synthetic Underwater Acoustic Dataset for Underwater Domain Awareness](https://suniltyagi.in/notes/synthetic-dataset/): 12,000 clips across four vessel classes and ambient sea noise, generated full-factorially from first-principle acoustic models (CC BY 4.0). Dataset: https://huggingface.co/datasets/suniltyagi/underwater-acoustic-synthetic-dataset (DOI 10.57967/hf/11191) · source: https://github.com/suniltyagialtair/Underwater-Acoustic-Synthetic-Dataset
- [SSR-Sim: what a submarine's sonar echo is made of](https://oravontsystems.com/research/ssr-sim-submarine-sonar-echo/): an Oravont working note (19 October 2026, hosted on oravontsystems.com) on target strength. On a generic 66 m double-hull submarine, SSR-Sim puts the beam echo at 43 dB at 5 kHz, more than half of it from the flat sides of the sail; the note covers aspect, frequency, range (the near-field effect on long hulls) and what a coating buys. PDF: https://oravontsystems.com/assets/SSR-Sim_note_sonar_echo_2026-10-19.pdf. SSR-Sim is target-strength software developed in-house at Oravont Systems LLP by Sunil Tyagi, and used in Oravont's acoustic stealth engineering services: https://oravontsystems.com/stealth/ssr-sim/ Series introduction: https://oravontsystems.com/notes/target-strength/what-an-echo-can-tell-a-sonar/

- [What an echo can tell a sonar, and three assumptions that make it computable](https://oravontsystems.com/notes/target-strength/what-an-echo-can-tell-a-sonar/): article 1 of the Oravont series "Submarine target strength, explained" (22 October 2026, hosted on oravontsystems.com). Video (3 min, YouTube): https://www.youtube.com/watch?v=y5b6QSZY8ow. Oravont's working notes and series: https://oravontsystems.com/notes/

## Profiles

- ORCID: https://orcid.org/0000-0001-5897-7955 · Scopus Author ID 57194035292 · Google Scholar: https://scholar.google.com/citations?user=aKkPxGcAAAAJ · Wikidata: https://www.wikidata.org/wiki/Q141676937
- LinkedIn: https://www.linkedin.com/in/sunil-tyagi-01 · GitHub: https://github.com/suniltyagi
- Practice: Oravont Systems LLP, https://oravontsystems.com · info@oravontsystems.com