基于QDRNN-ADRC的重力稳定平台控制研究

针对稳定平台系统存在系统模型不够精确或者参数变化,或者外部干扰未知等现象,以及采用自抗扰控制器存在参数众多且难以整理的问题,提出了一种基于准对角递归神经网络—自抗扰控制器(QDRNN—ADRC)的重力稳定平台控制算法.通过自抗扰控制器对系统的"总扰动"进行估计并补偿,同时引入神经网络的辨识功能对自抗扰控制器部分参数进行在线整定,基于自抗扰控制技术核心架构设计了QDRNN—ADRC.仿真结果表明:有效解决了重力稳定平台利用神经网络的辨识功能对自抗扰控制器部分参数进行在线整定外扰动的干扰以及参数自适应整定问题,相对于传统控制方法,其在稳定精度、快速性及抗干扰性方面均具有一定优势.

Research of gravity stabilized platform control based on QDRNN-ADRC

Model for stable platform system is not accurate or parameters change,or unknown external disturbances,and the active disturbance rejection controller(ADRC)parameters are numerous and difficult to sort out. In view of the above problems,a control algorithm for gravity stabilized platform based on quasi-diagonal recurrent neural network(QDRNN)-ADRC is proposed. The estimation and compensation of ADRC for the system of disturbance,and using the identification function of the neural network,the ADRC parameters are on-line tuned;based on ADRC technology core architecture,design QDRNN-ADRC controller,and Simulink simulation. The simulation results show that this method can effectively solve the gravity stable platform online tune outside interference and adaptive parameter tuning problem,compared with the traditional control method,it has certain advantages in quickness and stable precision,anti-jamming.