基于忆阻的自适应单神经元多变量解耦PID控制器

Adaptive Single-Neuron Multivariate Decoupling PID Controller Based on Memristor

为了解决强耦合问题和控制过程中参数难于调节的问题,提出了基于忆阻的自适应单神经元多变量解耦PID控制器,根据传统控制器的结构和原理,构建了两变量解耦控制器的结构。分析了两变量解耦控制器的性能,给出了实验仿真结果。通过实验结果,发现这个新的两变量解耦控制器具有较高的控制可靠性、较好的适应性,比传统的控制器学习速度快。最后得到了一个理想的、自适应、自组织的多变量解耦控制器,并获得了较好的控制效果。

Memristor is a kind of nonlinear circuit element with memory function when power off.Its resistance value changes automatically with its charge changing and flux changing which can be used as a controllable parameter in the modern control systems.The proportional-integral-derivative(PID) is still the major control mechanism in the majority of manufacturing industries.In recent years,many researchers have proposed a lot of decoupling PID control systems for solving the strong coupling problem and the problem that the control parameters are difficult to adjust in the control process.In this paper,we present a adaptive single-neuron multivariate decoupling PID controller based on memristor.Based on the structure and principle of the traditional controller,the structure and principle of the bivariate decoupling controller with memristor is established.Then the performance of the bivariate decoupling controller is analyzed and the simulation results are shown.According to experimental simulation results,we find out that the new bivariate decoupling controller has higher control reliability and better adaptability and acts more rapidly than the traditional controller.Finally we get the ideal,adaptive and self-organizing multivariate decoupling PID controller and the better control effect of the new controller is achieved.