Xiaojia Liu’s scientific contributions

The risks for ferries in the Yangtze River are relatively high, as they frequently cross the main traffic flows, leading to more intersections with other upwards and downwards ships. Although some studies have developed many models to assess collision risks in the Yangtze River, collision warning studies on ferries are scant. Meanwhile, most of the current collision studies evaluate risk based on AIS data, which are incapable of providing real-time ship information as they are discrete-time series data. In this work, fused data combining radar and AIS data are applied in a real-time ship collision warning model to assess the dynamic risk for ferries in the Yangtze River. Firstly, data fusion technology is proposed to acquire refined ship trajectories from AIS and radar data. Then, a widely used geometric collision model is enhanced to assess the real-time collision risk for ferries. And lastly, to illustrate the model, a real case of a ferry crossing through the Yangtze River is studied. The real-time risk values of the ferries are calculated based on fused data inputs, and the output results indicate that the use of fused data provides more accurate and continuous real-time ship risks. Thus, the proposed approach is evidenced to support the development of smart maritime surveillance.

The currently used grading system for navigational environment evaluation is inflexible to deal with the fuzziness and randomness of evaluation indexes and the weighing of influencing factors is complicated and subjective in a degree. The comprehensive risk evaluation model based on the combination weighing grey cloud clustering is proposed to solve the problems. The normal evaluation indexes are modified with grey cloud whitenized weight ones, which solves the ambiguity of the index evaluation. The combination weighing combines the objective and subjective information according to the principle of additive integration. The subjective weights are determined with cloud theory while the objective weights are determined according to the principal components by means of factor analysis. An example of the practical application is analyzed and the results show that the model can effectively solve the problems of subjectivity, randomness and ambiguity of boundary indices in existing evaluation, and gives the objective risk degree evaluation for port navigational environment.