一种基于PF-RBF的ANNPID参数自整定方法  

PID Parameters Auto-tuning based on PF-RBF Identification

作　　者：苏岭东 赵成 马祥林 SU Lingdong;ZHAO Cheng;MA Xianglin(State Grid Xuzhou Power Supply Company,Xuzhou Jiangsu 221000,China;Hangzhou BONUAP Intelligent Technology Co.,Ltd.,Hangzhou Zhejiang 310012,China;Changzhou Zhike Automation Technology Co.,Ltd.,Changzhou Jiangsu 213001,China)

机构地区：[1]国网徐州供电公司,江苏徐州221000 [2]杭州邦友安派智能科技有限公司,浙江杭州310012 [3]常州致科自动化科技有限公司,江苏常州213001

出　　处：《通信技术》2021年第3期658-663,共6页Communications Technology

摘　　要：针对实际控制过程中控制对象难以建立精确模型和由于非线性非高斯噪声干扰导致控制效果难以达到预期的问题,提出了一种基于粒子滤波和RBF神经网络辨识(RBF Neural Network and Particle Filter Algrothm,PF-RBF)的单神经元PID参数自整定方法。通过PF和RBF系统辨识得到精确的系统Jacobian信息,解决神经网络PID控制由于Jacobian信息未知导致的近似计算不精确问题。仿真实例表明,相对于无PF滤波的RBF辨识控制系统,该方法能改善控制系统的性能指标和抗干扰能力,对实际控制过程具有一定的指导意义。In the actual control process,it is difficult to establish an accurate model of the controlled object,and the control effect is difficult to achieve expected due to the interference of nonlinear and non-Gaussian noise.To solve these problems,a new neural network PID controller based on PF and RBF identification is proposed.Accurate system Jacobian information is obtained through PF and RBF system identification,which can solve the problem of inaccurate approximate calculation caused by unknown Jacobian information when controlled by neural network PID.The simulation example indicates that compared with the RBF identification control system without PF filtering,this method can improve the performance and anti-interference capability of the control system,and has certain guiding significance for the actual control process.

关 键 词：粒子滤波 RBF神经网络辨识 信息 PID参数自整定 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]