The object of this study is the processes of automated management of maritime safety by analyzing the manifestations of the human factor of sea navigators.
The task solved is justified by the need for formal and logical analysis and intelligent identification of mental motivational states (MMS) of marine navigators whose actions can cause dangerous situations during the control of the ship’s movement. High accident rates due to the fault of the navigators, in the absence of automated means of monitoring their condition, cause a contradiction between the existing means of safety control in controlling the movement of the vessel and the modern requirements of navigation, which needs to be resolved.
A safety management approach was devised that takes into account the specificity of navigational tasks and the p-adic classification of dangerous MMS for navigators. This has made it possible to create three security modes that are activated depending on the detected state of the navigator’s MMS.
Features of the results are the combination of analysis by means of p-adic systems and intelligent methods of data processing. As a result, sufficient identification accuracy was obtained for more than 75 % of MMS through neural network training.
Experimental data collected during the navigation watch, as well as on the Navi Trainer 5000 navigation simulator (Wärtsilä Corporation, Finland), became the basis for simulation by means of neural networks. In turn, the training of neural networks made it possible to obtain sufficient identification accuracy by performing up to 3000 iterations. Overall, the learning rate of the neural network was 0.98, which indicates a high level of identification.
From a practical point of view, the results could be used for the automated management of shipping safety, as well as for evaluating the level of adaptation of the navigator to dynamically changing conditions. The proposed approach provides opportunities for the application of modern intelligent technologies in the field of maritime transport safety, namely artificial neural network tools that determine notification modes or activation of automatic ship traffic control modules.
The specified contradiction requires the design of specialized systems for automated safety management of ship traffic control based on the identified states of navigators