时滞对象的神经网络Smith预测控制
The Smith predictive control of NN to the delayed object
-
摘要: 针对系统的时滞问题提出一种新的控制系统,将神经网络Smith预估器与神经元控制器通过预测误差有效结合在一起,起到预测控制目的.在神经网络Smith预估器中采用优于BP网络的IRN网络,仿真的结果表明系统的实时控制速度有了明显地提高.Abstract: Considering the delayed problem of the system, a new control system was introduced, which can realize the predictive control by combining the Smith predictor of the NN and the neuron controller through the predictive errors effectively. We use the NN of IRN in the Smith predictor, which is better than the NN of BP.The result of the simulation has indicated that this system can improve the real time control speed efficiently.