基于小波神经网络的深海电机系统混沌预测  

Chaotic prediction of deepwater motor system based on wavelet neural network

作　　者：孙昌志[1] 胡翠华[1] 安跃军[1] 张扬[1] 

机构地区：[1]沈阳工业大学电气工程学院,沈阳110178

出　　处：《沈阳工业大学学报》2009年第1期7-10,21,共5页Journal of Shenyang University of Technology

基　　金：辽宁省科技计划重点项目(2005216010);辽宁省教育厅科技研究项目(2005L299)

摘　　要：针对深海机器人推进电机系统易出现混沌现象,直接危及其稳定运行这一现状,提出一种有效的预测方法对混沌现象进行分析,实现对混沌的预先控制.利用相空间重构与小波神经网络相结合的方法对深海机器人推进电机系统的混沌现象进行分析研究,与以往单纯利用小波神经网络进行预测相比,通过相空间重构可以为小波神经网络的输入提供较为准确、可靠的数据样本,使得预测结果具有较强的可靠性和实用性,同时解决了小波神经网络不能进行中长期准确预测的问题,为进一步研究混沌控制和混沌抑制提供了基础条件.The chaotic phenomenon can appear in the motor system of deepwater robots and influence the stability of the system. Thus, an effective method was proposed to analyze and control the chaotic phenomenon. Through combining the wavelet neural networks with phase space reconstruction, the chaotic phenomenon appeared in the thruster motor system of deepwater robots was investigated and analyzed. Compared with performing prediction only by wavelet neural network, phase space reconstruction can offer the accurate and reliable data sample for the input of wavelet neural network. The prediction results are reliable and practical. It can solve the problem that wavelet neural network can＇t perform the accurate mid-long term prediction, and can provide the basis condition for the chaotic control and restraint.

关 键 词：小波变换 神经网路 推进电机 混沌 预测 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]