The concept of ship domain has been widely used in risk assessment, collision avoidance and plays an important role in marine traffic engineering as a principal criterion for navigational safety. After several decades of research and development, there are still issues on ship domain that need to be improved such as how the ship domain is calibrated (i.e., size and shape), what factors influence the domain and how it is applied in navigation. The ability to reflect continuously risk level is still lack in most of existing ship domains.
Moreover, ship domain is likely to be dynamic in nature, fuzzy in usage and interactive in application. The objective of this research is to develop a novel ship domain that accounts for dynamic aspect and can be effectively applied to collision risk evaluation in navigation by series of boundaries in both open sea and restricted areas.
Following a critical review of literature on the concept of ship domain and how it has been represented, measured and modeled, a new approach of modeling ship domain around the ship, named Potential Risk Ship Domain, is proposed. The formulated mathematical model is then calibrated using Gaussian influence function to express the impact of the ship to neighborhood. The potential risk of the neighborhood is taking the characteristics of the ship and the actual situation of navigation. The set of ship location and movement data associated are used to calibrate the ship domain characterized by ship attributes, i.e., overall length, speed, and heading. The size of ship domain depends on a selection of the parameters. In order to determine the size of ship domain, the approximation method is used to determine the value of parameter by comparison with previous ship domains.
The ship domain, consists of three boundaries that divide the area around the ship into four zones, named Collision Assessment Zone, with different potential risk levels. Based on the value of potential collision risk index, the degree of risk level around the ship is evaluated in real time and in both restricted and open sea areas. Improvement of boundaries also overcome the deterministic problem of existing models. Furthermore, the overlapped function when two ships are coming closer helps to predict the potential collision area. The effectiveness of proposed framework is validified by applying in simulations of real accident scenarios.
Obviously, the results showed that the proposed ship domain has ability to evaluate the potential risk of collision and to identify the dangerous area around the ship.