Depth-aided prediction of a vessel’s longitude using a neural network

A depth-based method for determining vessel longitude using a neural network is proposed. The input of the network is a vector containing depth values measured at a certain spatial interval using an echo sounder. The output of the network is the longitude corresponding to the position of the last sounding. The network has several hidden layers, varying from one to ten layers. The hidden neurons use hyperbolic tangent activation functions, while the single neuron of the output layer employs a linear activation function. Algorithms for training data generation, network training, and testing are defined. Together, they form the basis for creating a neural network-based system for vessel longitude prediction, implemented in the Python programming language. The TensorFlow library is used for working with neural networks. An official electronic navigational chart is chosen as the source of bathymetric data. Based on the extracted sounding layer, a regular grid is formed, where depth values at the grid nodes are calculated using linear interpolation. The procedure for forming training and test data includes pseudorandom variations of sea level, which may result from both real fluctuations and measurement errors. For the test set, the accuracy of the network is acceptable for navigation, and the results depend on the number of hidden layers. The best accuracy, in terms of minimal maximum absolute longitude prediction error, is achieved with the network having the greatest number of hidden layers. It is also noted that it is necessary to test the developed neural networkbased system using vessel motion modelling.

1. Klyueva, S. F. and V. V. Zav’yalov. Sintez algoritmov batimetricheskikh sistem navigatsii Vladivostok: Mor. gos. un-t, 2013: 132.

2. Stepanov, O. A. Metody otsenki potentsial’noy tochnosti v korrelyatsionno-ekstremal’nykh navigatsionnykh sistemakh Analiticheskiy obzor SPb: TsNII “Elektropribor”, 1993: 84.

3. Ma, D., Y. Ben, et al. “A robust fusion terrain-aided navigation method with a single beam echo sounder.” Ocean Engineering 286 (2023): 115610. DOI: 10.1016/j.oceaneng.2023.115610.

4. Golden, J. P. “Terrain Contour Matching (TERCOM): a cruise missile guidance aid.” Image processing for missile guidance 238 (1980): 10–18.

5. Savit, Carl H. Patent No. 3,212,189. Marine navigation procedure. U. S. Publ. 19 Oct. 1965.

6. Deryabin, Viсtor V. “Neural network-based method for determining vessel position by seabed relief.” Vestnik Gosudarstvennogo universiteta morskogo i rechnogo flota imeni admirala S. O. Makarova 15.5 (2023): 723–734. DOI: 10.21821/2309-5180-2023-15-5-723-734.

7. Deryabin, Viсtor V. “Depth-based vessel position fixing by means of a neural network.” Vestnik Gosudarstvennogo universiteta morskogo i rechnogo flota imeni admirala S. O. Makarova 16.1 (2024): 7–16. DOI: 10.21821/2309-5180-2024-16-1-7-16.

8. Deryabin, Viсtor V. “Seabed relief-based vessel position fixing with a neural network.” Vestnik Gosudarstvennogo universiteta morskogo i rechnogo flota imeni admirala S. O. Makarova 15.2 (2023): 172–179. DOI: 10.21821/2309-5180-2023-15-2-172-179.

9. Deryabin, Viсtor V., Sazonov Anatoly E. “Depth-aided prediction of vessel latitude based on a neural network.” Vestnik Gosudarstvennogo universiteta morskogo i rechnogo flota imeni admirala S. O. Makarova 17.1 (2025): 94–104. DOI: 10.21821/2309-5180-2025-17-1-94-104.

10. Wang, M–X. and Y. Qu. “Approximation capabilities of neural networks on unbounded domains.” Neural Networks 145 (2022): 56–67. DOI: 10.1016/j.neunet.2021.10.001.

11. Gorban’, A. N. “Obobschennaya approksimatsionnaya teorema i vychislitel’nye vozmozhnosti neyronnykh setey.” Siberian Journal Of Numerical Mathematics 1.1 (1998): 11–24.

12. Kingma, Diederik P., and Ba Jimmy. “Adam: A method for stochastic optimization.” 3rd International Conference on Learning Representations. 2015. DOI: 10.48550/arXiv.1412.6980.