摘要
针对多模贝叶斯“滤波—辨识联合算法”(JEI)多次迭代后存在的滤波器发散问题,提出了一种基于数据融合的多模自反馈被动角跟踪算法(MSPT DF)。特点是利用数据融合理论中的自反馈方法,将基于最小方差准则的反馈机制引入到原有算法中,使得误差较小的“融合”滤波结果能够实时地反馈给子模型,帮助其修正滤波误差,并且根据各模型的方差大小调整模型权重;改进算法大大改善了原有算法稳定性差和跟踪精度不高的缺点,计算机仿真结果证明了文中结论。
A multimodel self-feedback bearings-only passive tracking filter based on data fusion theory is presented to solve the collapse problem of "Joint-Identification algorithm" (JEI) when it works iteratively many times. Its advantage is that it introduces self-feedback methods into JEI based on data fusion theory which feed the "fusion output" back to the sub-models to correct their filtering errors. The modified algorithm greatly improves the original algorithm's intrinsic drawbacks such as poor stability and low tracking accuracy. Simulation results prove this conclusion.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2006年第5期689-692,共4页
Systems Engineering and Electronics
关键词
角被动跟踪
数据融合
自反馈
bearings-only passive tracking
data fusion
self-feedback