EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 2.5

Hybrid Sonar-Vision Approach for Path Estimation and Real-Time Classification, Tracking of Underwater Mine like Target using AI Edge-Enabled Tethered ROV

Jinka Venkata Aravind, Shanthi Prince

Defence Science Journal · 27.10.2026 · 76 (6) · 942-955

DOI: 10.14429/dsj.21676 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Defence Science Journal, online erschienen am 27.10.2026. Verfasst von Jinka Venkata Aravind und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 2.5).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Due to ocean currents, naval mine like targets may drift from predefined mine patterns or grids, making it essential to estimate their path for effective neutralization. In this study, the underwater target is tracked using active sonar, and the target path was estimated with the Extended Kalman Filter (EKF). The mathematical modelling of target motion analysis is presented, and various simulation scenarios are analysed to evaluate the Root Mean Square (RMS) error for different speeds and course angles of both the observer and the target. For a fixed course of 75° between the target and observer, with both moving at lower speeds (1 m/s), the RMS error approached zero, demonstrating the best performance. Furthermore, a YOLOv5 based target detection algorithm was employed for real-time classification, detection, and tracking of target using an vision based AI Edge enabled tethered underwater ROV. Pool experiments were conducted, achieving 94 % and 92 % confidence scores in real-time classification and detection of the target.

Zitieren

APA

Aravind, J. V., & Prince, S. (2026). Hybrid Sonar-Vision Approach for Path Estimation and Real-Time Classification, Tracking of Underwater Mine like Target using AI Edge-Enabled Tethered ROV. Defence Science Journal, 76(6), 942-955. https://doi.org/10.14429/dsj.21676

Vancouver

Aravind JV, Prince S. Hybrid Sonar-Vision Approach for Path Estimation and Real-Time Classification, Tracking of Underwater Mine like Target using AI Edge-Enabled Tethered ROV. Defence Science Journal. 2026;76(6):942-955. doi:10.14429/dsj.21676

BibTeX
@article{Aravind2026,
  author = {Aravind, Jinka Venkata and Prince, Shanthi},
  title = {Hybrid Sonar-Vision Approach for Path Estimation and Real-Time Classification, Tracking of Underwater Mine like Target using AI Edge-Enabled Tethered ROV},
  journal = {Defence Science Journal},
  year = {2026},
  volume = {76},
  number = {6},
  pages = {942--955},
  doi = {10.14429/dsj.21676}
}
Kostenlos starten →