MIMO system invertibility and decoupling control strategies based on ANN /spl alpha/th-order inversion
MIMO system invertibility and decoupling control strategies based on ANN /spl alpha/th-order inversion
复制标题
DOI：
10.1049/ip-cta:20010283
复制
发表时间：
2001-03
期刊：
影响因子：
--
通讯作者：
X. Dai;D. He;X. Zhang;T. Zhang
中科院分区：
文献类型：
--
作者：
X. Dai;D. He;X. Zhang;T. Zhang
研究了逆系统理论。建立并证明了由输入输出微分方程描述的MIMO非线性连续系统可逆的充分条件，并给出了相应的逆系统的解析表达式.为了克服逆系统实际实现的困难，提出了两种ANN /spl α/th阶逆系统。每一个神经网络由一个静态神经网络和若干个积分器组成，构成一种连续的动态神经网络结构。这两种方法都可以在原非线性连续系统模型不完善或几乎未知的情况下构造出来。两种形式的神经网络逆的建议提供了一个实用的方法，线性化和解耦控制的真实的对象。
The theory of inverse systems is studied. A sufficient invertibility condition for the MIMO nonlinear continuous system described by input-output differential equations is established and demonstrated, and then an analytical expression of the corresponding inverse system is given. To overcome the difficulty of the practical implementation of inverse systems, two kinds of ANN /spl alpha/th-order inverse systems are proposed. Each of them consists of a static ANN and several integrators and forms a kind of continuous dynamic ANN structure. Both of them can be constructed even if the original nonlinear continuous system is poorly modelled or almost unknown. The proposal of two forms of ANN inversions provides a practical approach for the linearisation and decoupling control of real plants.