基于改进BP神经网络的船舶轨迹识别方法  

Ship trajectory identification method based on improved BP neural network

作　　者：凡甲甲 祁云嵩 葛霓琳 FAN Jia-jia;QI Yun-song;GE Ni-lin(School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang 212000,China)

机构地区：[1]江苏科技大学计算机学院

出　　处：《计算机工程与设计》2019年第12期3639-3644,共6页Computer Engineering and Design

基　　金：国家自然科学基金项目(61471182);2018年江苏省研究生科研创新计划基金项目(KYCX18_2335)

摘　　要：针对船舶航行轨迹识别,为提高识别率,进行深入研究,提出一种采用附加动量法和自适应学习速率法的改进BP神经网络方法。采用附加动量不断修正BP神经网络的权重,加快网络收敛速度,在迭代过程中进行学习率自适应调整,减少迭代次数。运用改进BP神经网络和传统BP神经网络对船舶自动识别系统(automatic identification system,AIS)信息进行训练,分别建立分类识别模型。以安徽巢湖水域为例进行实验,实验结果表明,改进BP神经网络对船舶轨迹识别具有更高的准确率。To improve the recognition rate for ship navigation trajectory identification,an improved BP neural network method using additional momentum method and adaptive learning rate method was proposed.Additional momentum was used to continuously correct the weight of BP neural,to speed up the network convergence.The learning rate was adaptively adjusted in the iterative process,to reduce the number of iterations.The improved BP neural network and traditional BP neural network were used to train the automatic identification system(AIS)information of ships to establish the classification recognition model.Taking the Chaohu waters in Anhui as an example,the results show that the improved BP neural network has higher accuracy for ship trajectory recognition.

关 键 词：轨迹识别 BP神经网络 船舶自动识别系统 附加动量 自适应学习 

分 类 号：TP398.1[自动化与计算机技术—计算机应用技术]