电子科技 ›› 2021, Vol. 34 ›› Issue (11): 26-30.doi: 10.16180/j.cnki.issn1007-7820.2021.11.004

• • 上一篇    下一篇

基于正则化模型的RLS算法改进

孙帅,刘子龙,万伟   

  1. 上海理工大学 光电信息与计算机工程学院,上海 200082
  • 收稿日期:2020-07-04 出版日期:2021-11-15 发布日期:2021-11-16
  • 作者简介:孙帅(1993-),男,硕士研究生。研究方向:非接触性监测。|刘子龙(1972-),男,副教授。研究方向:控制科学与控制理论、机器人控制。
  • 基金资助:
    国家自然科学基金(61603255)

Improvement of RLS Algorithm Based on Regularization Model

SUN Shuai,LIU Zilong,WAN Wei   

  1. School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200082,China
  • Received:2020-07-04 Online:2021-11-15 Published:2021-11-16
  • Supported by:
    National Natural Science Foundation of China(61603255)

摘要:

对于天线信号项目中存在的杂波信号干扰,通常采用RLS算法进行滤波处理,但是常规RLS算法的遗忘因子通常是固定的,不能同时满足跟踪速度快与稳定误差小的要求。针对上述问题,文中提出了一种基于正则化模型的改进RLS算法。通过增加改进函数使遗忘因子增加可变性,并改进自相关矩阵逆矩阵方程。增加相关限制条件,将杂波信号降噪为原始信号,保证降噪效果存在且唯一,较大程度上减少了杂波的干扰。实验波形图显示,与常规RLS算法相比,改进算法具有更强的跟踪能力和稳定性。

关键词: 正则化模型, RLS算法, 改进因子, 自相关矩阵, 目标函数, 参数估值, 收敛速度, 跟踪能力

Abstract:

For the clutter signal interference existing in the research of antenna signals, the RLS algorithm is commonly used for filtering processing. However, the forgetting factor of conventional RLS algorithm is usually fixed, which cannot meet the requirements of fast tracking speed and small stability error at the same time. To solve the above problems, this study proposes an improved RLS algorithm based on regularization model. By adding an improved function, the forgetting factor can increase variability, and the inverse matrix equation of the autocorrelation matrix can be improved. The added correlation restriction condition can reduce the noise of the clutter signal to the original signal, and the existence and uniqueness of the noise reduction effect are ensured, thus greatly reducing the interference of the clutter. The waveform diagram of this experiment show that compared with the conventional RLS algorithm, the improved algorithm has stronger tracking capability and stability.

Key words: regularization model, RLS algorithm, improvement factor, autocorrelation matrix, objective function, parameter estimation, convergence speed, tracking ability

中图分类号: 

  • TP391
Baidu
map