基于神经网络的时变大滞后系统的Smith预估控制

Neural networks based smith pre-estimated control for time-varying system with large time delay

针对大滞后不确定系统提出了一种基于人工神经网络的改进型smith预估控制方案 ,设计了一个基于神经网络的补偿器来克服不确定的大延迟对控制性能的不利影响 ,解决了传统Smith预估控制鲁棒性差及需要预先知道受控对象精确数学模型的问题 .数字仿真结果表明 ,此方案可以在被控对象数学模型未知的情况下对时滞对象进行控制 ,特别是当时滞对象的特性发生变化时 ,具有较好的适应性 。

An improved Smith pre-estimated control method based on artificial neural networks is proposed for the control of an uncertain system with large time delay. A compensator based on neural networks is designed to overcome the effect resulting from the uncertain large time delay. It can be used to solve the problems of bad robustness and the prerequisite condition is that precise mathematical models must be known in advance for the conventional Smith control method. Simulation result proves that this method can control the large time delay plants when not knowing their mathematical models. Especially when the feature the plant is are varied, this new method is more adaptive.