系统仿真学报 ›› 2018, Vol. 30 ›› Issue (10): 3770-3780.doi: 10.16182/j.issn1004731x.joss.201810022

• 仿真应用工程 • 上一篇    下一篇

预测小波神经网络智能控制系统仿真研究

刘经纬1, 周瑞2, 赵辉3, 朱敏玲4, 孟祥花5, 张宇豪1,6   

  1. 1. 首都经济贸易大学信息学院,北京 100070;
    2. 北京中医药大学中药学院,北京 100029;
    3. 清华大学信息技术研究院,北京 100084;
    4. 北京信息科技大学计算机学院,北京 100101;
    5. 北京信息科技大学理学院,北京 100192;
    6. 首都经济贸易大学计算交通研究中心,北京 100070
  • 收稿日期:2016-02-17 修回日期:2016-07-30 出版日期:2018-10-10 发布日期:2019-01-04
  • 作者简介:刘经纬(1982-),男,北京,博士,副教授,研究方向为人工智能与智能控制。
  • 基金资助:
    国家自然科学基金(71371128, 11402006),北京社科基金研究基地(16JDYJB028),首经贸学术骨干培养计划(00791754840263)

Simulation Research on Predictive Wavelet Neural Network Intelligent Control System

Liu Jingwei1, Zhou Rui2, Zhao Hui3, Zhu Minling4, Meng Xianghua5, Zhang Yuhao1,6   

  1. 1. Information College, Capital University of Economics and Business, Beijing 100070, China;
    2. School of Chinese Materia, Beijing University of Chinese Medicine, Beijing 100029, China;
    3. Research Institute of Information Technology, Tsinghua University, Beijing 100084, China;
    4. School of Computer Science, Beijing Information Science & Technology University, Beijing 100101, China;
    5. School of Applied Science, Beijing Information Science and Technology University, Beijing 100192, China;
    6. Computational Transportation Science Center, Capital University of Economics and Business, Beijing 100070, China;
  • Received:2016-02-17 Revised:2016-07-30 Online:2018-10-10 Published:2019-01-04

摘要: 针对高危险高污染环境中控制系统参数在线整定和优化的需求,提出了一种结合向量时间序列预测方法和小波神经网络方法的控制参数在线整定智能控制系统,即通过增设小波神经网络作为控制系统的智能整定机制和自回归移动平均向量时间序列算法预测输出替代根据实时输出进行计算。采用理论分析和对多个方法进行计算机仿真对比实验等方式,验证了上述方法和智能控制系统具有可行性、工程应用稳定性、更好的快速性、更低的稳态误差等特性,给出了稳定性保证方法

关键词: 向量自回归移动平均, 向量时间序列预测控制, 控制参数在线整定, 小波神经网络, 智能控制系统仿真

Abstract: In order to realize control parameters online tuning and optimizing applications in the high risk and high pollution environment, vector time series prediction based intelligent control methods are proposed, which are combined autoregressive moving average vector time series predictive method and wavelet neural network method. Based on theoretical and comparative research: feasibility, stability, better dynamic characteristics, less steady-state error of the proposed methods are verified. A comparative simulation platform for the above intelligent control system is provided. Stability guarantee method of the above new methods is given.

Key words: VARMA, vector time series predictive control, control parameters online tuning, wavelet neural network, intelligent control system simulation

中图分类号: