摘 要

随着船舶信息化的不断发展，智能船舶被认为是船舶领域未来发展的大方向，世界各国对于智能船舶技术的发展尤为关注，对智能船舶开展了大量的研究。智能航行是智能船舶的一项重要功能，对于船舶实现无人化和提高航行安全性至关重要。船舶智能决策是智能航行的关键技术，该技术解决的主要问题为智能船舶在航行过程中的路径规划问题。路径规划分为全局路径规划和局部路径规划，其中局部路径规划，通常也称为避障，它是根据传感器实时采集环境信息进行障碍物建模，并规划出一条从当前位置到目标地的最优路径。

船舶在复杂的海洋环境中航行时，不可避免地会遇到许多障碍物，例如海岛、其他船只等。但是船舶在规避障碍物的过程中，大多依赖船员的经验知识，缺乏模式化的避碰决策体系。因此，深入研究智能船舶局部路径规划技术具有重要意义。

为了提高船舶在航行时的安全性，使船舶正确地规避障碍物并到达目标点，本文针对上述问题，以智能船舶为研究对象，以避碰策略为主题，开展了如下工作：

（1）通过仿真实现基于人工势场法的避碰决策算法。首先建立人工势场模型，根据目标点、障碍物位置建立人工势场，根据势场中各点的势能大小规划一条从当前位置到目标点的路径，并合理选择斥力势场大小，使船舶在航行中同时兼顾安全性和经济性。

（2）针对人工势场法规划路径不平滑以及海事规则难以数字化的问题，本文首先采用了一种对船舶的角加速度设限幅的方法，通过限制船舶每一个周期速度方向的角度变化，使得到的路径没有“凸点”，更加光顺；采用设置船舶转向可行域的方法，船舶在与其它船舶会遇时，只能向海上避碰规则允许的方向航行，使得到的路径点满足海事规则。仿真结果表明，该方法可以满足船舶在航行中的操纵性和合法性。

（3）针对船舶规避障碍物时，缺乏模式化的决策避碰体系问题，本文在人工势场法的基础上，制定了船舶对单障碍物以及对多障碍物的避碰策略。船舶在单障碍物避碰时，通过判断船舶之间的会遇态势，结合海事规则，规划安全路径；船舶在多障碍物避碰时，通过船舶之间的碰撞危险度，确定避碰重点船，在单障碍物避碰策略的基础上，规划安全路径。仿真结果表明，智能船舶在单障碍物和多障碍物的会遇态势中，均能按照海事规则的要求，规划出一条使船舶安全通过的路径，验证了避碰策略的合理性。

关键词：人工势场法；海事规则；避碰策略；单障碍物；多障碍物

Abstract

With the development of the information of ship, intelligent ships are considered to be the general direction of the future development of the ship field. Intelligent navigation is an important function of intelligent ships. It is of great significance to realize unmanned ships and improve navigation safety. Intelligent ship decision is the key technology of intelligent navigation. The main problem solved by this technology is the path planning of intelligent ships during navigation. Path planning is divided into global path planning and local path planning. Among them, partial path planning is usually also called obstacle avoidance, which collects environmental information from sensors in real time to model obstacles and plans an optimal path from the current location to the target ground.

When sailing in a complicated marine environment, ships inevitably encounter many obstacles, such as islands and other ships.

However, in the process of avoiding obstacles, ships mostly rely on the empirical knowledge of crew and lack a model of a collision avoidance decision-making system.

In order to improve the safety of the ship while sailing, so that the ship can avoid obstacles and reach the target point correctly, this paper focuses on the above problems, takes intelligent ships as the research object, and takes the intelligent ship partial obstacle avoidance strategy as the theme, and carries out the following work:

(1) The algorithm of collision avoidance decision system based on artificial potential field method is realized through simulation. First, build an artificial potential field model and build an artificial potential field according to the target and the position of the obstacle. Then plan a path from the current position to the target point according to the potential energy of each point in the potential field, and choose the size of the repulsive potential field reasonably so that the ship can be both safe and economy.

(2)Aiming at the problems of unsmooth planning path of artificial potential field method and difficult to digitize maritime rules, this paper firstly proposes a method of limiting the acceleration of the ship. By limiting the changing angle of speed of the ship, the resulting path has no "bumps" and is smoother. The method of setting the ship to the feasible area is adopted. When meeting with other ships, the ship can only sail in the direction allowed by the collision avoidance rules at sea, so that the waypoints meet the maritime rule.

(3) Aiming at the problem of lacking of a model-based decision-making collision avoidance system when ship avoids obstacles, based on the artificial potential field method, this paper formulates the collision avoidance strategy for ships against single obstacles and multiple obstacles. When the ship avoids collision with a single obstacle, it will plan the safe path by judging the situation between the ships and combine the maritime rules. When the ship avoids collision with multiple obstacles, the key collision avoidance ship is selected through the collision risk between the ships. And based on a single obstacle avoidance strategy, the system plans a safe path. Through simulation experiments, it is found that in the situation of single obstacles and multiple obstacles, the intelligent ship can plan a path to pass safely according to the requirements of the maritime regulations, which verifies the rationality of the collision avoidance strategy.

Keywords: artificial potential field method; maritime rules; collision avoidance strategy; single obstacle; multiple obstacles

目录

第1章 绪论 1

1.1 研究背景和意义 1

1.2 国内外研究现状 1

1.3 研究目标 3

1.4 论文研究内容及章节安排 4

第2章 船舶航行中的海事规则 5

2.1 引言 5

2.2 船舶领域 5

2.3 船舶会遇态势 5

2.4 船舶避让规则 6

2.5 避让责任 7

2.6 本章小结 8

第3章 船舶避碰流程及避碰评价指标 9

3.1 引言 9

3.2 危险度评估模型 9

3.3 避碰流程 9

3.4 避碰评价指标 10

3.4.1 避碰时机评价模型 11

3.4.2 避让幅度评价模型 11

3.4.3 避碰方式评价模型 11

3.4.4 安全通过距离评价模型 11

3.4.5 最终评价值计算 12

3.5 本章小结 13

第4章 基于人工势场法的智能船舶局部路径规划策略 14

4.1 引言 14

4.2 人工势场法 14

4.2.1 人工势场法原理 14

4.2.2 人工势场法的局限性 15

4.3单障碍物避碰策略 15

4.3.1 基于人工势场法的避碰策略 15

4.3.2 基于船舶操纵性的避碰策略 16

4.3.3 基于海事规则的避碰策略 16

4.4 多障碍物避碰策略 17

4.4.1 多障碍物会遇态势 17

4.4.2 单艇与静态障碍物避碰策略 17

4.4.3 多艇避碰策略 18

4.5 避碰策略流程图 18

4.6 本章小结 20

第5章 基于人工势场法的智能船舶局部路径规划仿真 21

5.1 引言 21

5.2 单障碍物避碰仿真及评价 21

5.2.1 对遇态势仿真 21

5.2.2 追越态势仿真 23

5.2.3 交叉相遇态势仿真 24

5.3 多障碍物避碰仿真及评价 26