摘要
在分析实验室船模轴频电场和海洋实测船舶轴频电场信号与环境电场数据的基础上,利用径向基神经网络预测模型对轴频电场信号进行控制.先对信号进行测量,在合理设置训练集与目标集的基础上建立预测模型,得到预测值后在下一时刻利用电极在场源处反向输出该预测值,以达到减弱特征信号的目的.通过对实验室与海洋实测船舶轴频电场数据进行处理,结果表明:用该方法能对船舶轴频电场信号特征进行有效控制,使幅值减小到原信号的30%以下.
Based on fully analyzing of shaft rate （SR） electric field and noise got in the lab and at sea, a method of controlling SR electric field through radical-basis-function （RBF） neural network was presented. The signal was got first to form the predicting model as the training data and the target data was set properly. The same value of the prediction result was exported reversely from the source u sing electrodes after it was obtained so as to weaken the characteristic signal. Through processing the SR electric field and noise got in the lab and at sea, it can be seen that this method can effectively con trol the SR signal of ship and can reduce the signal amplitude to less than 30%.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2012年第4期84-87,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
海军工程大学水中军用目标国防重点实验室基金资助项目(5144407105JB11)
关键词
轴频电场
径向基神经网络
信号预测
时间序列预测
预测模型
信号特征控制
SR electric field
RBF neural network
signal prediction
time series prediction
predic tion model
signal controlling
