摘要
讨论了具有自适应功能的基于改进Elman网络的动态系统辨识建模与预测方法,并依据特定海情下实测的舰船航态数据,利用该法训练舰船航态的预测模型。该法对具有时变特性的非线性系统的建模与预测有一定的自适应能力。仿真结果表明,按该法进行建模与预测,具有动态性好、逼近速度快,精度高等特点,并且自适应性与抗噪性有较大改善,说明改进的Elman网络是一种新颖、可靠的负荷预测方法。
Presents an adaptive dynamic system identification modeling and prediction method based on improved Elman network.This method has been used to train prediction model according to actual measured data of a warship's motion attitude. This method is adaptive to modeling and prediction of time variable and nonlinear system. The prediction accuracy of improved Elman is wonderful.The experiments show that not only the improved Elman neural is efficiency and accuracy but also its adaptive and antinoise abilities have been obviously improved, so the improved Elman neural network is a new and reliable method for predicting warship's motion attitude.
出处
《华东船舶工业学院学报》
2004年第5期19-22,共4页
Journal of East China Shipbuilding Institute(Natural Science Edition)
关键词
ELMAN网络
舰船航态
自适应建模
预测
Elman neural network
warship's navigation state
adaptive modeling method
prediction
