Efficient and reliable path planning is crucial for smart ships when avoiding collisions with static and dynamic obstacles in complex marine environments. This research proposes a novel path planning method based on the fast marching method to specifically assist with safe navigation for autonomous ships. At the very beginning, a unified representation is specially produced to describe the path planning space based on the parametric fast marching speed function. In addition, the spatial–temporal interaction effects of dynamic obstacles are considered and integrated into the construction of planning space. Subsequently, a path optimization strategy is put forward based on the trajectory prediction of dynamic objects. Particularly, the effectiveness of the method has been validated and evaluated through a number of simulations, which proves that such a method is practical in narrow and crowded waterways.
The USV(Unmanned Surface Vessel) different from land robot is that the obstacles has uncertainty when the USV autonomous navigation. The core factors of autonomous navigation include risk measurement and route planning. Proposed a route planning method for the USV based on the probability of collision and A* algorithm. First, the core risk objects in the channel have been selected, such as channel, beacon and pier. The probability distribution of collision is described by Artificial Potential Field (APF) model. Then, the track distribution in the different key sections has been acquired by the history AIS data. The parameter values in the APF model are solved by non-linear optimization on the history AIS data. After the Potential Field distribution is obtained, the optimal route is solved by A* algorithm. To solve the problem that the route may be hard to handle, we proposed a smoothing algorithm based on regression analysis. That algorithm for the probability distribution is based on the actual vehicle routes, therefore it is more reliable. The planning routes take the manoeuvrability of the vehicle into consideration. In the simulation, the path of the USV is closed to the real vessel and it performs well.
In the context of global carbon emission reduction, as an important source of carbon emission, the accurate calculation of carbon emission from ships has been concerned. AIS, as a navigational guidance and AIDS equipment that is forcibly installed on ships, provides a good data basis for calculating ship carbon emission. At present, most studies tend to use “bottom-up” Ship carbon emission accounting methods to deal with a large amount of AIS data aiming to obtain carbon emission on a wide spatio-temporal scale. On the one hand, the massive AIS data make the refined AIS data completion and ship navigation state recognition complicated and difficult to achieve. On the other hand, the accuracy of the calculation results of ship carbon emission based on AIS data is very dependent on the accurate recognition of ship navigation state. This makes further calculation of ship carbon emission difficult. Therefore, this paper proposes a calculation method of ship carbon emission that focuses on ship interaction behavior and is applicable to massive AIS data. First, this study proposes a new recognition method of ship interaction, which realizes efficient recognition of ship interaction and obtaining spatial and temporal distribution characteristics. Secondly, this study improves a ship navigation recognition method, with which combining ship interaction recognition method can improve the calculation speed. Finally, this paper builds a simple visualization system of carbon emissions from ships, which can calculate carbon emissions and display them in the form of heat map. To verify the feasibility of this study, this paper conducts four experiments, respectively showing the performance of improved ship navigation state recognition algorithm, the relationship between ship navigation state and ship interaction, improvement of calculation speed using method in this paper and carbon emission calculation result with description of accuracy.
The scale of global trade continues to expand, and the total carbon emissions of container shipping, as the core carrier of international logistics, are rising year by year. The proposal of the dual carbon goal puts forward strict requirements for energy conservation and emission reduction in the shipping industry. Carbon emission of container ships is influenced by multiple factors such as ship parameters, navigation conditions, energy types, policy environment, etc. It is difficult for traditional statistical analysis methods to accurately capture the complex correlation between various variables. Machine learning algorithm shows significant advantages in processing high-dimensional and nonlinear complex data, and can automatically tap the potential correlation between characteristic variables such as ship deadweight tonnage, sailing speed, fuel type and carbon emissions, breaking through the limitations of traditional empirical formulas. Aiming at the insufficient comprehensiveness of feature extraction and weak capability of time series information capture in traditional prediction models, this paper proposes HLOA-CNN-BIGRU regression algorithm. The research firstly carries out correlation analysis and violin chart analysis, and compares and verifies with various machine learning algorithms. The results of regression performance evaluation show that the algorithm has outstanding advantages in prediction accuracy and fitting effect. The mean square error is 718.577, the root mean square error is 26.806, the mean absolute error is 18.635, the mean absolute percentage error is 14.176, and the coefficient of determination is 0.943. This study provides reliable technical support for accurate prediction and scientific control of carbon emissions of container shipping, and has important practical value in promoting the green and low-carbon transformation of shipping.