基于BP神经网络优化的某破障武器自抗扰控制研究  

Research on active disturbance rejection control of barrier-breaking weapon based on BP neural network optimization

作　　者：龚永昌 高强[1] 王嘉良 张亮伟 严来福 GONG Yongchang;GAO Qiang;WANG Jialiang;ZHANG Liangwei;YAN Laifu(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China)

机构地区：[1]南京理工大学机械工程学院,南京210094

出　　处：《自动化与仪器仪表》2024年第7期76-79,共4页Automation & Instrumentation

基　　金：中央高校基本科研业务费专项资金资助(2023101001)。

摘　　要：针对某破障武器位置随动系统工作时,受到各种非线性扰动导致系统的响应速度和跟瞄精度受到影响的问题,设计了一种改进型自抗扰控制器(ADRC)。利用BP神经网络(BPNN)适应性好和自学能力强的优点,对ADRC的相关参数在线寻优,并采用人工鱼群算法(AFSA)对神经网络的权值进行优化,来进一步提升控制器性能。利用MATLAB软件对控制器进行仿真验证,仿真结果表明:该控制方法能有效提高随动系统的抗干扰能力和跟踪精度。A modified Active Disturbance Rejection Controller(ADRC)is designed to address the issue of various nonlinear disturbances affecting the response speed and tracking accuracy of a certain obstacle breaking weapon position servo system during operation.By utilizing the advantages of good adaptability and self-learning ability of BP neural network(BPNN),the relevant parameters of ADRC are optimized online,and the artificial fish swarm algorithm(AFSA)is used to optimize the weights of the neural network,in order to further improve the performance of the controller.Using MATLAB software to simulate and verify the controller,the simulation results show that this control method can improve the anti-interference ability and tracking accuracy of the servo system.

关 键 词：随动系统 破障武器 BP神经网络 人工鱼群算法 自抗扰控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]