基于动态径向基函数神经网络的多变量解耦控制

为提高工业控制领域中多变量、非线性、强耦合系统的解耦能力和动态特性，基于聚类结合算法和神经网络原理，提出了一种改进的基于动态径向基函数（RBF）神经网络的多变量解耦控制方法.采用聚类结合算法优化动态RBF神经网络，更好地描述了控制对象的动态行为，获得了PID参数在线调整信息，实现了多变量非线性系统的解耦控制.仿真结果表明，与基于常规RBF神经网络的PID控制方法相比，该方法具有更高的控制精度、更快的系统响应以及更好的适应性和鲁棒性，是解决多变量、非线性和强耦合问题的一种简便、有效的控制算法.

Multivariable decoupling control based on dynamic radial basis function neural network

An improved multivariable decoupling control method based on dynamic radial basis function neural network(RBFNN) was presented according to clustering algorithm and principle of neural network in order to improve the decoupling capability and dynamic property of multivariable,nonlinearity and strong coupling systems in industrial control fields.The controller's dynamic behavior was implemented by the dynamic RBF neural network to obtain the on-line tuning information of PID parameters and realize the decoupling control of multivariable nonlinear system.The simulation result indicates that compared to the PID control method based on the conventional RBF neural network,the control method has faster response,better adaptability and robustness,and is simple and effective for multivariable,nonlinearity and strong coupling systems.