Risk of accidents in maritime traffic is usually evaluated with various methods including quantitative and qualitative approaches that adopt various indicators for risk and account for a range of technical aspects of traffic. However, they lack proper reflection of causes and factors underlying the accidents at sea, including the effect of surrounding traffic and bathymetric conditions on human performance, thus probability of an accident. Moreover, the methods do not account properly for the navigation in restricted waters, namely inland and coastal traffic. A proper strategic planning of waterways by the authorities cannot be done properly without reliable algorithms accounting for relevant factors.
To bridge this gap, ASTRA develops a novel method and a tool for all parties interested in diagnostic assessment of maritime traffic. The tool will quantify the risk measured with the accident potential for a given sea area based on relevant and observable variables that can be derived directly from the traffic and bathymetric data, that are directly linked with the performance of a navigator on board a ship. Three elements are considered here as the main drivers for the human performance in accident evasive actions: (i) waterway, (ii) traffic and environment complexity, (iii) environment complexity. The proposed method is rooted in the first principles, parameterized with the use of big data analysis and machine learning methods, as well as experts’ knowledge and their understanding of the subject.
Therefore ASTRA introduces a novel, unique, scientifically sound, industry-friendly and user-intuitive solution allowing strategic assessment and mitigation of risk of accidents in maritime traffic systems (collisions, allisions, and groundings).
Coordinator
Dr Jakub Montewka, Waterborne Transport Innovation, Poland
Consortium
Waterborne Transport Innovation, Poland
Ozyegin University, Turkey
NavSim Polska Sp. z o.o., Poland
Gdynia Maritime University, Poland