欢迎访问《兵工学报》官方网站,今天是 分享到:

兵工学报 ›› 2022, Vol. 43 ›› Issue (3): 556-564.doi: 10.12382/bgxb.2021.0117

• 论文 • 上一篇    下一篇

基于径向基函数神经网络的光电系统自适应控制

张通彤1, 姜湖海1, 岳巍2, 司晨1, 袁满1   

  1. (1.西南技术物理研究所, 四川 成都 610041; 2.空装驻成都地区第五军事代表室, 四川 成都 610041)
  • 上线日期:2022-04-07
  • 通讯作者: 姜湖海(1984—),男,高级工程师,硕士生导师,博士 E-mail:jhhqgj@163.com
  • 作者简介:张通彤(1988—),女,工程师,硕士。E-mail:king_ybw@163.com
  • 基金资助:
    装备预先研究项目(414XXXX104)

Adaptive Control Based on RBF Neural Network for Electro-optical System

ZHANG Tongtong1, JIANG Huhai1, YUE Wei2, SI Chen1, YUAN Man1   

  1. (1.Southwest Institute of Technical Physics, Chengdu 610041, Sichuan, China;2.The 5th Military Representative Office of Air Force in Chengdu, Chengdu 610041, Sichuan, China)
  • Online:2022-04-07

摘要: 针对光电跟踪系统对于跟踪目标的高精度需求,在硬件设计选型、装调适配完成后,通常需要伺服控制系统设计合理的算法以改善跟踪精度。为持续提高伺服控制系统的综合能力,首先分析跟踪精度的误差模型,通过理论推导以及典型数值计算仿真的方式,验证伺服控制器控制增益对于跟踪精度改善的重要性。在对比多型控制算法基础上,提出基于径向基函数神经网络的自适应控制方法,发挥神经网络控制能够自行学习优化的特点,使伺服稳定平台控制系统具有更高的跟踪精度和更好的鲁棒性。数字仿真以及半实物实验验证结果表明,与传统PID、积分分离PID、单神经元PID控制方法相比,在存在载体扰动条件下,所提方法能够实现在3 Hz带宽内时滞最小约为28 ms,幅值误差在3 Hz处约为4%,可为光电跟踪系统设计实现高精度跟踪提供一种有效设计思路。

关键词: 光电系统, 神经网络, 径向基函数, 跟踪精度, 自适应控制

Abstract: For improving the tracking accuracy,the servo control algorithm is optimized to enhance the control accuracy and rapidity of control system after hardware design and system integration. A tracking error model is proposed theoretically. According to the proposed model,the numerical computation is implemented. The result illustrates that the control algorithm plays an important role in controlling the tracking errors of electro-optical system. The radial basis function (RBF) algorithm is employed here to adaptively adjust the control system parameters for improving tracking accuracy and robustness. Both the theoretical analysis and hardware-in-the-loop simulation are practiced to verify the availability of the proposed method. It is proved that the proposed method can achieve less than 28 ms delay and lower than 4% attenuation within 3 Hz bandwidth.

Key words: electro-opticalsystem, neuralnetwork, radialbasisfunction, trackingaccuracy, adaptivecontrol

中图分类号: