Supporting maritime autonomous systems with intelligent control methods

This study analyzed the possible methods of controlling autonomous maritime transport objects in collision situations caused by an incorrect assessment of the object’s proximity or subjective decision-making. For this purpose, positional and matrix game theory, dynamic programming, and a neural network were used. The sensitivity of safe control to changes in environmental conditions was analyzed to assess the algorithms’ effectiveness. Computer simulations of the developed control algorithms and a comparison of their results confirmed that autonomous transport safety can be improved by incorporating elements of game theory and artificial intelligence.