IJMSMT · Original Research Article Vol. 1, No. 1 (October 2026), pp. 30–39
Maritime Technology and Digital Innovation

EVALUATING THE ACCURACY OF ARTIFICIAL INTELLIGENCE SYSTEMS IN CELESTIAL NAVIGATION CALCULATION

Prince Joseph Okporu* — Nigeria Maritime University, Okerenkoko, Delta State, Nigeria
* Corresponding author: prince.okporu@nmu.edu.ng
Published: 22 October 2026 •Received: 20 Aug 2026 •Accepted: 21 Sep 2026 •DOI: 10.5281/zenodo.23559474 • Licence: CC BY 4.0

Abstract

As Artificial Intelligence (AI) increasingly transforms maritime education, its reliability in high-precision technical tasks such as celestial navigation remains largely unexplored. This study evaluates the accuracy and reliability of three prominent Large Language Models (LLMs); ChatGPT, Gemini, and DeepSeek in performing celestial navigation calculations. Employing an experimental comparative research design, the models were tested against standardised problems involving Difference of Latitude (DLAT), Difference of Longitude (DLONG), Azimuth, Compass Error, and Latitude by Ex-meridian. The outputs were analysed against manually calculated solutions derived from established maritime references. The results reveal a significant bifurcation in performance: while all models achieved 100% accuracy in basic linear tasks (DLAT/DLONG), accuracy declined sharply in complex spherical trigonometry and multi-step calculations. Gemini demonstrated the highest overall accuracy (64%), followed by DeepSeek (56%) and ChatGPT (44%). Statistical analysis using the Friedman test confirmed that Gemini is significantly more reliable than ChatGPT (p = 0.014). Error classification identified computational mistakes (41.7%) and magnitude errors (37.5%) as the most prevalent failure modes, posing substantial risks to navigational safety if left unverified. The study concludes that current AI tools are unsuitable as standalone resources for specialised maritime competencies. A hybrid instructional approach is recommended, where students first master traditional foundational skills before using AI as a supplementary tool for practice under expert supervision.

Keywords: Celestial Navigation, Generative AI, Accuracy Assessment, Maritime Education, ChatGPT, Gemini, DeepSeek.

Keywords: Celestial NavigationGenerative AIAccuracy AssessmentMaritime EducationChatGPTGeminiDeepSeek.
Full text

The complete article is available as a PDF, free to read and download under CC BY 4.0.

Download PDF

References

  1. Acheampong, K. B., Owusu Bandoh, S., Dimaweh, T. G. A., & Akweittey, E. (2025). AI usage in mathematics
  2. teaching and mathematics performance: The role of students interest. Turkish Journal of Mathematics Education, 6(2), 41–50.
  3. Burch, D. (2015). Celestial navigation: A complete home study course (2nd ed.). Starpath.
  4. Critchley-Marrows, J. J. R., & Mortari, D. (2023). A return to the sextant - Maritime navigation using celestial bodies and the horizon. Sensors, 23(10), 4869. https://doi.org/10.3390/s23104869
  5. Dachev, Y., & Panov, A. (2017). 21st century celestial navigation systems. In 18th Annual General Assembly of
  6. the International Association of Maritime Universities (IAMU).
  7. Joshi, S. (2025). A comprehensive review of DeepSeek: Performance, architecture and capabilities. Preprints.org. https://doi.org/10.20944/preprints202503.1887 .v1
  8. Liubarets, I. (2024). Celestial navigation as the emergency GNSS backup: Enhancing navigational reliability.
  9. Scientific Journal of Gdynia Maritime University, 129, 7–21. https://doi.org/10.26408/129.01
  10. Lo, C. K. (2023). What is the impact of ChatGPT on education? A rapid review of the literature. Education
  11. Sciences, 13(4), 410. https://doi.org/10.3390/educsci13040410
  12. Maity, S., & Deroy, A. (2024). Generative AI and its impact on personalised intelligent tutoring systems.
  13. arXiv. https://arxiv.org/abs/2410.10650
  14. Memarian, B., & Doleck, T. (2023). ChatGPT in education: Methods, potentials, and limitations. Computers in
  15. Human Behavior: Artificial Humans, 1(2023), 100022. https://doi.org/10.1016/j.chbah.2023.100022
  16. Mohamed, M. Z., Hidayat, R., Suhaizi, N. N., Mat Sabri, N., Mahmud, M. K. H., & Baharuddin, S. N. (2022).
  17. Artificial intelligence in mathematics education: A systematic literature review. International Electronic Journal of Mathematics Education, 17(3), em0694. https://doi.org/10.29333/iejme/12132
  18. Pinčetić, T., Lušić, Z., & Esko Krančević, P. (2024). The role of celestial navigation in modern day and future
  19. navigation. MT’24: 10th International Conference on Maritime Transport. https://doi.org/10.5821/mt.12734
  20. Pursnani, V., Sermet, Y., Kurt, M., & Demir, I. (2023). Performance of ChatGPT on the US fundamentals of
  21. engineering exam: Comprehensive assessment of proficiency and potential implications for professional environmental engineering practice. Computers and Education: Artificial Intelligence, 5, 100183. https://doi.org/10.1016/j.caeai.2023.100183
  22. Putro, B. S., Selasdini, V., & Saputra, I. R. (2023). Analysis the importance of applying celestial navigation to
  23. avoid the deficiency during ship inspections at LPG/C Navigator Aries. AIP Conference Proceedings, 2675(1), 040005. https://doi.org/10.1063/5.0117928
  24. Sajja, R., Sermet, Y., Cikmaz, M., Cwiertny, D., & Demir, I. (2024). Artificial intelligence-enabled intelligent
  25. assistant for personalised and adaptive learning in higher education. Information, 15(10), 596. https://doi.org/10.3390/info15100596
  26. Subramaniam, H. (2004). Practical navigation: Nutshell series book 1 (4th ed.). Vijaya Publications. (Reprinted
  27. 2010)
  28. Wang, D., Bian, C., & Chen, G. (2024). Using explainable AI to unravel classroom dialogue analysis: Effects of
  29. explanations on teachers' trust, technology acceptance and cognitive load. British Journal of Educational Technology. Advance online publication. https://doi.org/10.1111/bjet.13466
  30. Zary, A., & Zary, N. (2025). Artificial intelligence in technical and vocational education and training: Empirical
  31. evidence, implementation challenges, and future directions. Preprints.org. https://doi.org/10.20944/preprints202503.1147 .v1