基于BP神经网络的船体变形测量研究  

Measurement of Ship Deformation Based on BP Neural Network

作　　者：朱昀炤[1] 谢永成[1] 

机构地区：[1]装甲兵工程学院,100072

出　　处：《微计算机信息》2009年第16期251-252,314,共3页Control & Automation

基　　金：海军武器装备预研基金;基金申请人:朱昀炤;项目名称:惯性测量匹配法测定船体变形技术研究;基金颁发部门:海军武器装备预研基金委(编号不公开)

摘　　要：分析了船体变形的惯性测量原理,推导了测量系统的状态方程和观测方程,详细讨论了BP神经网络在测量船体变形时的应用情况,具体进行了BP神经网络结构的设计,训练样本的获得以及训练的方法,并对BP神经网和卡尔曼滤波测量船体变形角的误差进行了仿真对比,说明采用BP神经网络方法可以缩短变形角的辨识时间,更加快速的建立舰船统一坐标基准。The principle of inertial measurement of ship deformation is analyzed. The state equation and measurement equation of ship deformation measurement system is deduced. The situation is discussed in detail that the application of of ship deformation measurement based on Back Propagation network. The Back Propagation network is designed. The acquirement of training sample and the training method is also discussed. The simulation results show that the measurement precision based on Back Propagation network and that based on kalman filter is almost the same, but the time used by the neural network is lesser and the ship uniform coordinates reference is established faster.

关 键 词：船体变形 惯性测量 神经网络 卡尔曼滤波 

分 类 号：U666[交通运输工程—船舶及航道工程]