潜艇报警鱼雷制导方式的模式识别方法  

Discriminance of Submarine Alarm Torpedo Guidance Way Based on Pattern Recognition

作　　者：郑文恩[1] 刘剑[1] 肖明彦[1] 

机构地区：[1]海军潜艇学院软件中心,山东青岛266071

出　　处：《计算机仿真》2015年第12期6-9,14,共5页Computer Simulation

摘　　要：潜艇防御鱼雷中通常仅用鱼雷的初始报警方位作为决策依据,而忽略了来袭鱼雷的弹道特征信息。因此,由于对鱼雷报警信息利用不充分,对鱼雷制导方式的判别方法,存在主观性较强和结果不可靠等缺陷。分析了声自导和线导鱼雷弹道模型,提出了基于模式识别的鱼雷制导方式识别的方法。建立了两种制导方式的鱼雷弹道模式识别特征模型,通过仿真产生了模式识别的训练集和测试集,应用Matlab程序实现了基于BP神经网络的分类器的训练和测试,实验结果验证了上述方法的有效性。In the paper, the trajectory models of acoustic homing torpedo and wire -guided torpedo were ana- lysed, and a discriminance of torpedo guidance way based on BP Neural Network was put forward. The feature mod- els for the two kinds of torpedo＇s trajectory were built, and the training set and testing set for recognition were pro- duced through simulation. Developed a classifier based on BP Neural Network, and implemented its training and testing through MATLAB programming, the experiment verified the usefulness of that method as a result.

关 键 词：潜艇 防御鱼雷 鱼雷制导方式 模式识别 神经网络 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]