摘要
针对弹道信号中存在的噪声干扰问题,尤其是没有信号的先验统计特性,因此不能使用某一固定参数的滤波器来处理该噪声干扰问题。本文基于改进的粒子群优化算法(Improved Particle Swarm Optimization,IPSO),提出一种改进的最小均方(Least Mean Square,LMS)自适应的信号滤波方法(简称IPSO-LMS方法)。传统的标准粒子群优化算法中,所有的粒子均采用相同的更新策略,导致种群多样性缺失,并且标准粒子群优化算法还特别容易陷入局部最优。本文所提的IPSO-LMS算法,通过设置不同的粒子搜索机制,智能的筛选出LMS滤波器的阶数和迭代步长的最优参数值,进一步提升了LMS自适应滤波器的性能。数值结果表明,与PSO-LMS算法相比,本文所提的IPSO-LMS算法在前2秒内瞬态响应振荡幅值大幅度减小,稳态响应收敛速度更快。此外,在不同输入信噪比(signal-to-noise ratio,SNR)下,IPSO-LMS滤波算法对于低SNR信号的降噪效果更加突出,即在一定范围内大幅提升SNR可以一定程度减小均方根误差。对于高SNR信号,虽然均方根误差进一步缩小,但SNR优化空间有限,因此本文所提的IPSO-LMS滤波算法在低SNR下滤波效果优势明显。
In response to the noise interference problem in ballistic signals,especially without prior statistical characteristics of the signal,it is not possible to use a fixed parameter filter to deal with this noise interference problem.This paper proposes an improved least mean square(LMS)adaptive signal filtering method based on the improved particle swarm optimization(IPSO)algorithm.In traditional standard particle swarm optimization(PSO)algorithm,all particles adopt the same update strategy.This traditional algorithm leads to the lack of population diversity and is prone to local optimum.The IPSO-LMS algorithm proposed in this paper intelligently filters out the optimal parameter values of the order and iteration step size of the LMS filter by setting different particle search mechanisms,and the performance of the LMS adaptive filter has been further improved.Numerical results show that compared with the PSO-LMS algorithm,the oscillation amplitude of the transient response of the IPSOLMS algorithm has been greatly reduced in the first 2 seconds.The steady-state response of this algorithm converges faster.In addition,under different input SNRs(signal-to-noise ratios),the IPSO-LMS algorithm is more prominent in the noise reduction effect of low SNR signals.That is,greatly improving the SNR within a certain range can reduce the RMSE(root mean square error)to a certain extent.For high SNR signals,although the RMSE is further reduced,the optimization space for SNR is limited.Therefore,the IPSO-LMS algorithm proposed in this paper has obvious advantages in the filtering effect of low SNR.
作者
郭祥
王利勇
冯海伟
赵桓
张灵芝
高洁宇
陈菁瑶
GUO Xiang;WANG Liyong;FENG Haiwei;ZHAO Huan;ZHANG Lingzhi;GAO Jieyu;CHEN Jingyao(Jinxi Industial Group Co.,Ltd.,Taiyuan 030025,Shanxi,China)
出处
《弹箭与制导学报》
北大核心
2025年第5期961-969,共9页
Journal of Projectiles,Rockets,Missiles and Guidance
关键词
最小均方自适应算法
粒子群优化算法
弹道信号滤波算法
信噪比
least mean square adaptive algorithm
particle swarm optimization algorithm
ballistic signal filtering algorith
signal-to-noise ratio
