MATLAB环境下的单神经元自适应实时控制系统

Real-time Single Neuron Adaptive Control System Based on Matlab Environment

利用神经元模型和自学习功能构成自适应PID控制器,在Matlab实时开发环境xPCTarget下建立液位实时控制系统。给出了单神经元自适应PID控制器的结构和控制算法,介绍了基于xPCTarget的快速原型设计方法,即通过拖拉Simulink模块搭建被控对象的实时原型,同时以神经元自适应PID控制器完成系统的实时控制。为检验控制效果还采用了PID控制器进行液位实时控制。最终结果表明:单神经元自适应控制能克服传统PID控制器不稳定的缺点,使系统具有较好的稳定性。

Real-time level process control system was devised on the basis of xPC Target software environment of Matlab.Meanwhile an adaptive PID controller were designed by means of neuron model and self-learning algorithm.The architecture and algorithm of the single neuron adaptive PID controller were discussed.The rapid prototyping process which created block diagrams in Simulink using simple drag and drop operations was introduced in detail.The real-time control system adopted adaptive PID controller was created through Simulink model with xPC Target.In order to verify the performance of two approaches,the PID algorithm was also used to control the level system.The results indicate that the proposed method can overcome the shortcoming of the traditional PID controller and achieve satisfactory stability.