The theoretical analysis of maritime accidents is a hot topic, but the time characteristics and dynamics of maritime accidents time series are still unclear. It is difficult to draw a clear conclusion from the cause analysis, so the accident is difficult to be predicted. To bridge this gap, this research analyzes the characteristics and evolution mechanism of maritime accidents time series from the perspective of complex network theory. The visual graph algorithm is used to model the complex network of maritime accidents data in 22 jurisdictions of the Yangtze River, map the time series into a complex network, and reveal the time characteristics and dynamics of maritime accidents time series based on the complex system theory. In the empirical analysis, degree distribution, clustering coefficient and network diameter are used to analyze the characteristics of time series. The results show that the degree distribution of maritime accidents time series network presents power-law characteristics in the macro and micro levels, which shows that the maritime accidents time series is scale-free. In addition, according to the clustering coefficient and network diameter, maritime accidents time series in the Yangtze River has the characteristics of small-world and hierarchical structure. The research of this manuscript shows that the occurrence of maritime accidents is not random events and does not follow specific patterns but presents the characteristics of complex systems, and this phenomenon is common. The analysis of maritime accidents time series by complex network theory can provide theoretical support for maritime traffic safety management.

@en

The prediction of ship location has become an increasingly popular research hotspot in the field of maritime transportation engineering, which benefits maritime safety supervision and security. Existing methods of ship location prediction based on motion characteristics have a large uncertainty and cannot guarantee trajectory prediction accuracy of the target ship. An improved method of location prediction using k-nearest neighbor (KNN) is proposed in this paper. An expanded circle area of the latest point of the target ship is first generated to find the reference points with similar movement characteristics in the constraints of distance and time intervals. Then, the top k-nearest neighbors are determined based on the degree of similarity. Relationships between the reference point of each neighbor and the latest points of the target ship are calculated. The predicted location of the target ship can then be determined by a weighted calculation of the locations of all neighbors at the predicted time and their relationships with the target ship. Experiments of ship location prediction in 10 min, 20 min, and 30 min were conducted. The correlation coefficient of the location prediction error for the three experiments was 0.992, 0.99, and 0.9875, respectively. The results show that ship location prediction with reference to multiple nearest neighbors with similar movements can provide better accuracy.@en

Conflict detection is a vital step of collision prevention at sea, determining if there is a risk of collision and when to take preventing actions. This article proposes a practical Rule-aware Time-varying Conflict Risk (R-TCR) for ship collision avoidance. Considering maritime practice, the conflict risk measure takes the ship maneuverability, the COLREGs, and good seamanship into account in the conflict risk measure. Specifically, the conflict risk is formulated as a ratio of achievable maneuvers leading to a collision to all achievable maneuvers. Simulations are carried out to show the characteristics of R-TCR. The results show that the R-TCR evaluates the entire conflict risk incorporating COLREG rules, multiple targets, different maneuverability, and varying ship domains. Finally, the proposed measure is applied to analyze the collision accident between two ships. Compared with the conventional risk indicators, the proposed R-TCR can deliver extra information to users, such as providing early warning, showing the room-for-maneuver, and suggesting evasive actions. Besides, the extra information also supports collision avoidance for autonomous ships. 

@en

Many Collision Avoidance Systems (CAS) for autonomous ships usually presume that a ship's dynamics are completely known in advance. However, precise parameters for ships in different operating conditions are, in fact, uncertain and unknown. The parameter identification of ship dynamics is challenging and time-consuming. Thus, uncertainties in the ship dynamic model are inevitable, which can lead to errors between real trajectories and predicted trajectories. These errors might result in an unexpected collision between ships. Therefore, it is necessary to consider tracking errors in the CAS, which is missing in most existing CAS. This article proposes a way to incorporate the errors in CAS. Specifically, a Velocity Obstacle (VO) algorithm is employed to find collision-free velocities with estimated tracking errors. Firstly, the ship is assumed to be a "black box" whose inputs and outputs are observable, while the internal workings are unknown. Secondly, parameters optimization of a PID controller are employed to determine the best feedback gains for tracking given trajectories; Thirdly, the maximal tracking errors for controlling the ship to arbitrary velocities are estimated. Finally, the maximal error is added to the safety distance and the VO algorithm is employed to find a collision-free solution. The proposed Unknown-Dynamics CAS (UD-CAS) can support the upgrade of existing conventional ships to Type I-III maritime autonomous surface ship.

@en