纯滞后系统的单神经元Smith预测控制

Single Neuron with Smith Predictor Control of the Dead-time Systems

该文针对纯滞后系统的特点 ,给出了一种新的单神经元Smith预测控制算法 ,即由单神经元控制器和Smith预估器组合的复合控制 ,并将其应用于以电炉为被控对象的计算机温度控制系统。实验结果表明 ,这种控制方式能够充分利用神经元自学习的能力及补偿控制的优点 。

The general methods to control the long dead-time systems are mentioned. A single neuron controller with smith predictor is also presented. This controller is a multiplex controller including a single neuron and a Smith predictor. The control algorithm was applied to the temperature control system. The simulation and practical results proved the good adaptability and disturb-rejection of this controller.