| 一种非线性鲁棒均值预测控制器 |
| |
| 引用本文： | 王永骥,王宏.一种非线性鲁棒均值预测控制器[J].三峡大学学报(自然科学版),2002,24(5):406-408,455. |
| |
| 作者姓名： | 王永骥 王宏 |
| |
| 作者单位： | 1. 华中科技大学控制科学与工程系,武汉,430072
2. 英国曼彻斯特理工大学造纸系 |
| |
| 基金项目： | 国家自然科学基金(69974017)，河北省自然科学基金(602621)，高等学校骨干教师资助计划资助项目 |
| |
| 摘 要： | 提出了一种有界随机动态控制系统输出均值的非线性鳍棒预测控制器。在采用B－样条神经网络的平方根模型表示随机动态系统的输出概率密度函数基础上，输出分布的均值控制转换成一个非线性优化问题。为了解决这一问题，采用了L－M修正梯度搜索方法，得到预测控制器。基于Lyapunov稳定分析，得到了闭环渐近稳定的充分条件和考虑系统建模误差时的鲁棒稳定充分条件。仿真例子表明此算法正确，并获得了很好的结果。
|
| 关 键 词： | 非线性 有界随机动态系统 均值控制 B－样条神经网络 鲁棒性 |
| 文章编号： | 1007-7081(2002)05-0406-03 |
Nonlinear Robust Predictive Mean Controller of Bounded Dynamic Stochastic Systems Considering Modelling Error |
| |
| Wang Yongji Wang Hong.Nonlinear Robust Predictive Mean Controller of Bounded Dynamic Stochastic Systems Considering Modelling Error[J].Journal of China Three Gorges University(Natural Sciences),2002,24(5):406-408,455. |
| |
| Authors: | Wang Yongji Wang Hong |
| |
| Abstract: | A nonlinear one-step-ahead predictive mean controller for bounded dynamic stochastic systems is proposed. The B-spline neural network based square root model is used to represent the output probability density function, the mean control of output distribution is then addressed to a nonlinear optimization problem. To solve this problem, the L-M modification for gradient search approach is adopted;and based on the Lyapunov stability analysis, the sufficient conditions for the closed loop asymptotic stability and robust stability condition when existing modelling error are derived. Simulation examples are used to demonstrate the use of the algorithm;and the good results have been obtained. |
| |
| Keywords: | bounded dynamic stochastic systems mean control nonlinear predictive control B-spline neural networks robustness |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
