论文部分内容阅读
针对非线性过程,提出了一种基于自构建神经网络的内模控制方法(Internal Model Control,IMC)。采用自构建算法实现神经网络的结构学习和参数学习,在被控过程内部模型和控制器模型的辨识过程中,该网络能够根据给定的判定条件自动增加神经元节点,以满足辨识精度的要求;为了防止网络学习过拟合,基于灵敏度方法对神经网络隐层节点进行修剪删除;网络的参数学习采用梯度下降法。自构建算法可以有效地避免普通神经网络内模控制方案中网络结构难以确定的问题,仿真结果表明,该控制系统有良好的跟踪性、鲁棒性和抗干扰性。
For the nonlinear process, an internal model control (IMC) based on self-built neural network is proposed. The self-building algorithm is used to realize the structure learning and parameter learning of the neural network. During the identification of the internal model and the controller model of the controlled process, the neural network can automatically increase the neuron nodes according to the given decision conditions to meet the requirement of identification accuracy In order to prevent the network from over-fitting, the hidden nodes of the neural network are trimmed and deleted based on the sensitivity method. The gradient learning method is used to study the parameters of the network. The self-building algorithm can effectively avoid the problem that the network structure is difficult to be determined in the ordinary neural network internal model control scheme. The simulation results show that the control system has good tracking, robustness and anti-interference.