﻿ 基于大数据分析的复杂环境舰船导航方法
 舰船科学技术  2022, Vol. 44 Issue (19): 142-145    DOI: 10.3404/j.issn.1672-7649.2022.19.028 PDF

Design of ship navigation method in complex environment based on big data analysis
ZHANG Zhe
School of Naval Architecture and Navigation. Wuhan Technical College of Communications, Wuhan 430065, China
Abstract: In order to improve the combat capability of ships, the design and research of ship navigation method in complex environment based on big data analysis is proposed. Use sensors carried by ships to obtain water area environment information, use grid method to build complex environment model, preprocess ship navigation data based on big data analysis technology - time normalization processing, error correction processing and outlier elimination processing, integrate GPS and IMU data, determine the current absolute position of ship navigation, and use genetic algorithm to obtain the best ship navigation path based on ship navigation positioning results, Thus, the navigation of ships in complex environment is realized. The experimental data show that the minimum value of ship navigation and positioning error of the proposed method is 0.12%, and the length of ship navigation path is shorter and smoother, which fully proves that the proposed method has better ship navigation performance.
Key words: ship navigation     information fusion technology     complex environment     big data analysis technology     path planning     ship positioning
0 引　言

1 复杂环境舰船导航方法 1.1 复杂环境模型构建

 $\left\{ {\begin{array}{*{20}{c}} {x = \bmod \left[ {{\delta _x},n} \right]}，\\ {y = ceil\left[ {{\delta _x},n} \right]}。\end{array}} \right.$ (1)

 ${\delta }_{x}=\left\{\begin{array}{cc}0，& 可通行，\\ 1，& 不可通行。\end{array}\right.$ (2)

 图 1 复杂环境栅格模型示例图 Fig. 1 Example of complex environment grid model

 图 2 栅格尺寸大小与环境模型精度关系图 Fig. 2 Relationship between grid size and environmental model accuracy
1.2 舰船导航数据预处理

 ${t'_i} = \frac{{{t_i} - {t_{\min }}}}{{{t_{\max }} - {t_{\min }}}}。$ (3)

 $\varphi ' = \varphi + \Delta \varphi。$ (4)

 $\left\{\begin{array}{c}\varepsilon \left({t}_{i}\right)=Y\left({t}_{i}\right)-\left[h\left({t}_{i}\right)X\left({t}_{i},{t}_{i-1}\right)\right]，\\ \begin{array}{cc}\left|\varepsilon \left({t}_{i}\right)\right| > {\alpha }^{\ast }，& 野值，\\ \left|\varepsilon \left({t}_{i}\right)\right|\leqslant {\alpha }^{\ast }，& 正常值。\end{array}\end{array}\right.$ (5)

1.3 舰船导航定位

 \begin{aligned}\begin{cases} {\beta = {\beta _0} + {\gamma _0}V\left( {\eta + \left( {1 - \kappa + \lambda } \right)\dfrac{{{\eta ^3}}}{6} + \left( {5 - 18\kappa + {\kappa ^2}} \right)\dfrac{{{\eta ^5}}}{{120}}} \right)}，\\ \chi = {\chi _0} + {\gamma _0}V\left( \mu + V\tan \left( \dfrac{{{\eta ^2}}}{2} + \left( {5 - \kappa + 9\lambda + 4{\lambda ^2}} \right)\dfrac{{{\eta ^4}}}{{24}} + \right.\right.\\ \quad \;\; \left.\left.\left( {61 - 58\kappa + {\kappa ^2}} \right)\dfrac{{{\eta ^2}}}{{720}} \right) \right) 。\end{cases}\end{aligned} (6)

 $\begin{split}\begin{cases} {V = \dfrac{1}{{\sqrt {1 - {e^2}{{\sin }^2}\upsilon } }}}，\\ {\eta = \left( {\tau - {\tau _0}} \right)\cos \upsilon }，\\ {\kappa = {{\tan }^2}\upsilon }，\\ {\lambda = {{\cos }^2}\upsilon \times \dfrac{{{e^2}}}{{1 - {e^2}}}}，\\ {\mu = \sin 6\upsilon \times \dfrac{{35{e^6}}}{{3072}}} 。\end{cases}\end{split}$ (7)

 图 3 舰船导航位置确定框架图 Fig. 3 Ship navigation position determination framework
1.4 舰船路径规划

1) 构建复杂环境栅格模型，并获取障碍物位置、属性等信息；

2) 初始化信息熵遗传算法，制定迭代终止条件；

3) 构造适应度函数，为舰船路径寻优做准备，表达式为：

 $\left\{ {\begin{array}{*{20}{c}} {F\left( g \right) = \dfrac{{1000}}{{10 \times d\left( g \right) + \dfrac{{\varPsi \left( g \right)}}{\psi } \times m \times 2 + {\sigma _\theta }}}}，\\ {H\left( g \right) = F\left( g \right) - \omega * {N_p}}。\end{array}} \right.$ (8)

4) 通过交叉算子、变异算子对个体进行更新，并重新计算适应度数值。当未满足迭代终止条件时，转至步骤3进行重复迭代；当满足迭代终止条件时，输出个体对应的舰船路径。

2 仿真实验

2.1 实验环境段

 图 4 实验水域复杂环境栅格模型示意图 Fig. 4 Schematic diagram of grid model of complex environment in experimental waters

 $\left\{ {\begin{array}{*{20}{c}} {J = {V_N} + Q\cos \vartheta }，\\ {W = {V_E} + Q\sin \vartheta } 。\end{array}} \right.$ (9)

 $\left\{ {\begin{array}{*{20}{c}} {\left( {J + \Delta J} \right) = \left( {{V_N} + \Delta {V_N}} \right) + \left( {Q + \Delta Q} \right)\cos \left( {\vartheta + \Delta \vartheta } \right)} ，\\ {\left( {W + \Delta W} \right) = \left( {{V_E} + \Delta {V_E}} \right) + \left( {Q + \Delta Q} \right)\sin \left( {\vartheta + \Delta \vartheta } \right)}。\end{array}} \right.$ (10)

2.2 实验结果分析

 图 5 舰船导航路径示意图 Fig. 5 Schematic Diagram of Ship Navigation Path
3 结　语

 [1] 程建华, 刘佳鑫, 赵琳. 极区航海导航与定位保障技术发展综述[J]. 中国舰船研究, 2021, 16(5): 16-29. [2] 张国庆, 姚明启, 杨婷婷, 等. 考虑事件触发输入的船舶自适应动力定位控制[J]. 控制理论与应用, 2021, 38(10): 1597–1606. [3] 庞玺斌, 梁成程, 张闯. 一种船舶组合导航系统混合误差模型及应用[J]. 船海工程, 2020, 49(5): 127-132. DOI:10.3963/j.issn.1671-7953.2020.05.030 [4] 高嵩, 国岩, 李新飞, 等. 动力定位船矢量推进器回转动力学响应分析[J]. 北京理工大学学报, 2020, 40(11): 1182–1189. [5] 史博, 陈琳, 戴宪邦, 等. 某型AUV运载器的定位方法与误差分析[J]. 船舶工程, 2020, 42(6): 98–104. [6] 张啸尘, 赵建森, 王胜正, 等. 基于YOLOv3算法的船舶双目视觉检测与定位方法[J]. 上海海事大学学报, 2021, 42(1): 26–32. [7] 杨浩琪, 姚红革, 王诚, 等. 复杂场景下遥感船舶的检测与分割定位[J]. 计算机辅助设计与图形学学报, 2020, 32(3): 472–485. [8] 王国宇, 刘克中, 陈默子, 等. 基于信号到达角的船载WiFi室内定位影响因素分析[J]. 大连海事大学学报, 2020, 46(4): 24–31.