论文部分内容阅读
尽管模糊PID控制器具有良好的控制品质,但存在计算复杂和实时性差的问题,为了解决这个问题,利用RBF神经网络逼近能力重构模糊PID控制器,由于重构的RBF神经网络的并行计算能力,这简化了计算复杂性并提高实时性.通过选择不同的给定信号,比较模糊PID控制器和重构的RBF神经网络的控制性能,得到两者的控制效果是相当的.说明重构的RBF神经网络可以取代模糊PID控制器,从而减少了计算复杂性,避免维度灾难并改善控制实时性.
Although the fuzzy PID controller has good control quality, it has the problems of complex computation and poor real-time performance. In order to solve this problem, the RBF neural network approximation ability is used to reconstruct the fuzzy PID controller. Because of the parallel computing ability of the reconstructed RBF neural network, , Which simplifies the computational complexity and enhances the real-time performance.Comparing the control performance of the fuzzy PID controller and the reconstructed RBF neural network by selecting different given signals, the control effect of the two is equivalent. RBF neural network can replace the fuzzy PID controller, thus reducing the computational complexity, avoiding dimension disasters and improving the real-time control.