首页|一种基于相对固定增益的Kalman滤波信号跟踪算法

一种基于相对固定增益的Kalman滤波信号跟踪算法

扫码查看
研究了Kalman滤波在卫星导航终端信号跟踪中的应用.运用基于Kalman滤波形成的信号跟踪环路利用最优估计理论动态调节环路带宽,较好地解决了信号跟踪性能受信号强度和信号动态两因素制约的问题;运用信号系统中的Z变换从理论上证明了Kalman滤波稳态性能与普通锁相环路的转换联系;分析了Kalman滤波的稳态响应特点,借鉴α·β滤波的固定滤波系数思想,在保留Kalman滤波基本递归叠代过程的前提下,利用稳态效应误差恒定的性质,提出了一种基于相对固定增益的Kalman滤波信号跟踪算法.该算法采用基于SNR的查找表方式(无需计算方式)确定增益系数,以减少滤波的计算量.性能试验结果表明,该算法的鲁棒性较好,能实时跟踪信号变化,信号突变条件下的暂态响应跟踪误差略大于标准算法,稳态条件下的滤波效果与标准算法基本持平.
Kalman filter signal tracking based on relatively fixed-gain
The application of Kalman filtering to the signal tracking of satellite navigation terminals was studied.The tracking loop based on Kalman filter structure was applied to dynamical adjustment of the loop bandwidth by utilizing the optimum estimation theory to solve the restriction of signal intension and dynamic on the signal tracking performance.The relation between the Kalman filter stability and the common lock phase loop was theoretically proved by using the Z transformation in the signal system.The steady-state characteristics of the Kalman filter was analyzed, the idea of fixed filter coefficient about α · β filtering was adopted, and then, for simplifying the calculation, an algorithm for tracking of Kalman filtering signals based on the relatively fixed-gain was proposed under the condition of keeping the basic recursion process and using the principle of error constant in the steady state.The algorithm uses a SNR based seeking table (without calculation) to determine the gain coefficient to reduce the load burden.The results of the performance test show the algorithm is robust enough to track signal accurately.Its transient response error is a little bigger than that of the standard arithmetic when signals change suddenly.Its steady error is basically same as the standard arithmetic.

signal trackingKalman filterloop bandwidthfilter gain

王前、胡彩波

展开 >

北京航空航大学仪器科学与光电学院 北京100191

北京卫星导航中心 北京100094

信号跟踪 Kalman滤波 环路带宽 滤波增益

国家自然科学基金

41474027

2015

高技术通讯
中国科学技术信息研究所

高技术通讯

CSTPCDCSCD北大核心
影响因子:0.19
ISSN:1002-0470
年,卷(期):2015.25(1)
  • 2
  • 6