围绕智能航行研究，本期精选发表于《中国舰船研究》的5篇双语文章，欢迎阅读！
精选文章/Selected Articles
01
面向船舶智能航行测试的变稳船控制系统设计
Design of variable stability ship control system for ship intelligent navigation test
【摘要】 [目的]智能航行控制系统作为智能船舶的大脑和中枢，其控制性能的好坏直接决定船舶航行的安全性和经济性，因此需要对船舶智能航行控制系统进行验证。面向智能船舶智能航行控制系统的验证问题，提出一种通用型验证平台——变稳船。[方法]首先，基于模型跟随原理提出一种变稳船系统架构，分析智能船舶三自由度上的运动特性，然后根据三自由度变稳目标定义位置变稳误差，采用滑模控制技术设计变稳控制器，最后，在Matlab软件中进行仿真验证。[结果]结果显示，提出的变稳控制方法能够有效跟踪模拟待模拟船，实现变稳性能。[结论]所做研究可为智能航行控制系统的验证提供新的思路。
【Abstract】 [Objectives] The navigation control system is the brain and center of an intelligent ship, and the safety and economy of the ship's navigation are directly determined by its control performance. Therefore, it is necessary to validate the ship intelligent navigation control system. This paper proposes a general verification platform, the variable stabilized ship, to simulate the sailing state of the target ship at different scales and with different hydrodynamic characteristics. [Methods] First, the structure of the variable stability ship is proposed on the basis of the model-following principle, and an analysis is made of the three degrees of freedom movement characteristics. The steady error between the variable stability ship and the target ship is determined by their position errors, and the variable stability controller is built on the basis of sliding mode control.Finally, Matlab simulations are carried which verify that the proposed method can maneuver the variable stability ship to follow the states of the target ship. [Results] Simulation results show that the proposed method maneuver the variable stability ship to follow the states of the target ship. [Conclusions] The results of this study can provide references for the validation and verification of navigational control systems for intelligent ships.
02
基于深度强化学习的智能船舶航迹跟踪控制
Tracking control of intelligent ship based on deep reinforcement learning
【摘要】 [目的]智能船舶的航迹跟踪控制问题往往面临着控制环境复杂、控制器稳定性不高以及大量的算法计算等问题。为实现对航迹跟踪的精准控制，提出一种引入深度强化学习技术的航向控制器。[方法]首先，结合视线（LOS）算法制导，以船舶的操纵特性和控制要求为基础，将航迹跟踪问题建模成马尔可夫决策过程，设计其状态空间、动作空间、奖励函数；然后，使用深度确定性策略梯度（DDPG）算法作为控制器的实现，采用离线学习方法对控制器进行训练；最后，将训练完成的控制器与BP-PID控制器进行对比研究，分析控制效果。[结果]仿真结果表明，设计的深度强化学习控制器可以从训练学习过程中快速收敛达到控制要求，训练后的网络与BP-PID控制器相比跟踪迅速，具有偏航误差小、舵角变化频率小等优点。[结论]研究成果可为智能船舶航迹跟踪控制提供参考。
【Abstract】 [Objectives] The tracking control of intelligent ships often faces the problem of low controller stability in complex control environments and manual algorithmic computing. In order to achieve precise tracking control, this paper proposes a controller based on deep reinforcement learning (DRL). [Methods] Guided by the line-of-sight (LOS) algorithm and based on the maneuvering characteristics and control requirements of ships, this paper formulates a path of Markov decision processes by following the control problem, and designs its state space, action space, and reward by applying a deep deterministic policy gradient (DDPG) algorithm to implement the controller. An off-line learning method is used to train the controller. After the training, a comparison is made with BP-PID control to analyze the control effects. [Results] Simulation results show that the deep reinforcement learning (DRL) controller can rapidly converge from the training process to meet the control requirements, with the advantages of small yaw error, and a visible reduction in the frequency of changes of the rudder angle. [Conclusions] The study results can provide a reference for the tracking control of intelligent ships.
03
基于CSSOA的多船智能避碰决策研究
Multi-vessel intelligent collision avoidance decision-making based on CSSOA
【摘要】 [目的]智能避碰决策作为船舶安全航行的关键技术之一，对智能船舶的发展具有重要意义。针对多船会遇下的智能避碰决策问题，提出一种基于高斯变异和Tent混沌的改进麻雀搜索优化算法（CSSOA）。[方法]算法采用Tent混沌映射初始化麻雀原始种群，提高其多样性，并对适应能力差和搜索停滞的麻雀个体进行混沌映射，利用高斯变异提升局部搜索能力和鲁棒性，改进方案优化启发式算法收敛速度慢和易陷入局部最优的问题。综合考虑船舶间船速比、最小会遇距离、相对距离、最小会遇时间、相对方位等因素，利用模糊隶属度函数建立船舶碰撞风险模型，并通过多船典型会遇场景进行实例验证。[结果]实验结果显示，改进算法的平均迭代次数较粒子群算法和原麻雀算法分别减少了77.97%和53.57%。[结论]改进后的麻雀优化算法能以更优的收敛速度寻到安全经济的避碰路径，为船舶驾驶员提供避碰决策参考。
【Abstract】 [Objective] As one of the key technologies for the safe navigation of ships, intelligent collision avoidance decision-making is of great significance for the development of intelligent ships. Aiming at the intelligent collision avoidance decision-making problem under multi-vessel encounters, an improved chaos sparrow search optimization algorithm (CSSOA) based on Gaussian variation and Tent chaos is proposed. [Methods] The algorithm uses Tent chaotic mapping to initialize the original sparrow population and improve its diversity, chaotic mapping is applied to sparrows with poor adaptability and stagnant search ability, and Gaussian mutation is used to improve the local search ability and robustness. The improved scheme optimizes the problems of heuristic algorithms such as slow convergence speed and tendency to fall into the local optimum. A collision risk model is established using the fuzzy membership function with the comprehensive consideration of the ship-to-ship speed ratio, minimum encounter distance, relative distance, minimum encounter time and relative orientation. [Results] In a typical encounter scenario involving multiple ships, the experimental results demonstrate that the average number of iterations for the improved algorithm is reduced by 77.97% and 53.57% compared to particle swarm optimization and the original sparrow algorithm respectively. [Conclusion] The improved CSSOA can achieve a safer and more efficient collision avoidance path at a superior convergence speed, providing valuable guidance for ship navigators in making collision avoidance decisions.
04
基于驾驶实践的无人船智能避碰决策方法
Intelligent collision avoidance decision-making method for unmanned ships based on driving practice
【摘要】 [目的]为实现沿海无人驾驶船舶自主航行，充分考虑无人驾驶船舶智能避碰决策的合理性和实时性后，提出并建立一种基于驾驶实践的无人船智能避碰决策方法。[方法]首先，以本体论为基础，设计无人驾驶船舶航行态势本体概念模型，并结合《国际海上避碰规则》及良好的船艺将船舶航行态势量化划分为12种会遇场景；然后，从驾驶实践的角度改进影响碰撞危险度因子的模糊隶属度函数，提出一种多元碰撞危险度评估模型，实现船舶碰撞危险度的精确计算；最后，以船舶避碰总路径最短为目标函数，提出一种基于驾驶员视角（BOP）的智能避碰决策模型，在船舶操纵性、舵角限幅等约束下求解最优避碰策略，并在典型的会遇场景下进行仿真实验。[结果]结果表明，该方法可以准确判断驾驶航行态势，给出合理的转向策略，实现典型会遇场景下的有效避碰。[结论]所做研究可为实现船舶自主航行提供理论基础和方法参考。
【Abstract】 [Objective] In order to realize the intelligent navigation and autonomous collision avoidance of unmanned ships in coastal areas, we propose an intelligent collision avoidance decision-making method based on driving practice. [Method] First, the real-time rationality and uniqueness of the intelligent collision avoidance decision-making process for unmanned ships are analyzed. The ontological conceptual model of the navigation situation is then designed and combined with the International Regulations for Preventing Collisions at Sea (COLREGS) and good seamanship practices, and the ship encounter scenarios are quantitatively divided into 12 types. An improved composite assessment model of collision risk index is then proposed from the perspective of piloting practice to reflect collision risk more accurately. Finally, an intelligent collision avoidance decision-making model based on the operator’s perspective (BOP) is built, and the optimal collision avoidance strategy is solved by taking the shortest total collision avoidance path as the objective function under the constraints of ship maneuverability and rudder angle amplitude limit. Simulation experiments are then conducted in different obstacle environments. [Results] The simulation results show that this method can accurately determine the piloting situation, provide a reasonable steering strategy and achieve effective collision avoidance in different obstacle environments. [Conclusion] This study provides a theoretical basis and method for realizing the intelligent collision avoidance decision-making and dynamic local collision avoidance path planning of ships.
05
基于智能控制的船舶水动力导数敏感性分析方法
Sensitivity analysis method of ship hydrodynamic derivatives based on intelligent control
【摘要】 [目的]为了获得用于智能控制的船舶运动简化数学模型，以Mariner船为研究对象，提出结合标准操纵性试验与比例-积分-微分（PID）航向控制试验的敏感性分析方法。[方法]将控制指标、操纵性指标及整个时历过程典型运动状态变量平方损失进行复合分析，得到包含多维敏感性系数的数据集；引入K-means机器学习算法对该数据集进行聚类分析，完成水动力导数敏感性强弱的自动划分，进而对模型进行简化，并对所提简化模型、前人简化模型和完整模型的航向控制与航迹控制进行仿真试验。[结果]试验结果验证了所提敏感性分析方法的有效性，显示所提模型具有更高的操控预报精度。[结论]研究表明所提方法对指导基于智能控制的船舶运动建模具有一定的意义。
【Abstract】 [Objectives] In order to obtain a simplified mathematical model of ship motion for intelligent control, this paper takes a Mariner-class vessel as the research object and proposes a sensitivity analysis method combining the standard maneuverability test and proportion-integral-differential (PID) heading control test. [Methods] Compound analysis of the control index, maneuverability index, and squared loss of typical motion state variables throughout the entire process is performed to obtain a dataset containing multi-dimensional sensitivity coefficients. A machine learning-based K-means algorithm is introduced to perform cluster analysis on the dataset. The automatic sensitivity division of hydrodynamic derivatives is completed and the model is simplified. [Results] Contrastive simulation tests of heading control and track control are carried out among the simplified model, former simplified model, and complete model, and the results show that the sensitivity analysis method proposed in this paper is effective and the model proposed in this paper has higher accuracy in control prediction. [Conclusions] The method proposed in this paper has certain significance for guiding ship motion modeling for intelligent control.
期刊推荐/Recommended Journal
《中国舰船研究》创刊于2006，由中国船舶重工集团有限公司主管、中国舰船研究设计中心主办。该刊以登载舰船及相关专业新的理论方法、技术手段、设计概念及科技成果为已任，高举创新旗帜，加强技术交流，开展学术争鸣，推动理论创新与技术进步，促进舰船事业发展和海军装备建设现代化。
联系我们/Contact Us
中国知网“中文精品学术期刊外文版数字出版工程”（简称JTP）自2015年启动，已与400余种学术期刊合作出版了5万余篇双语对照论文，积累了丰富的学术翻译/英语加工/学术推广经验。形成了集双语出版、主题电子书出版、双语讲座视频制作、期刊英文内容编校加工、资讯编译、海外推广为一体的全方位服务体系，全面助力期刊提升国际影响力。