摘要
分析了强跟踪滤波器中新息方差近似计算方法的不足,提出了一种基于衰减记忆思想来近似计算新息方差的改进强跟踪滤波器。然后,对于过程噪声水平未知的目标状态估计问题,在Sage-Husa过程噪声水平自适应估计算法的基础上,一旦通过基于新息的滤波器发散判据检测到可能出现的发散现象,提出用改进的强跟踪滤波器进行抑制,极大地提高了滤波算法的鲁棒性。对三种典型的目标机动形式进行的Monte-Carlo仿真结果进一步验证了新提出算法的有效性。
Based on the analysis of the shortcomings of the fading-memory approximate calculation algorithm for the innovative variance in strong tracking filter, a modified strong tracking filter is proposed in this paper. Then in order to solve target state estimation problem with unknown process noise statistics, based on the Sage-Husa process noise statistics adaptive estimation algorithm, once the likely divergence of the tracking filter is detected by the innovation based filter divergence criterion, the modified strong tracking filter is used to restrain the likely divergence. Thus robustness of the tracking filter can be improved greatly. Validity of the new proposed algorithm is verified by means of Monte-Carlo simulations in three typical target maneuver scenarios further.
出处
《系统仿真学报》
CAS
CSCD
2004年第5期1020-1023,共4页
Journal of System Simulation
基金
国家重点基础研究发展规划(973)项目(2001CB309403)
关键词
强跟踪滤波器
自适应滤波
机动目标跟踪
发散抑制
噪声辨识
strong tracking filter
adaptive filtering
maneuvering target tracking
divergence restraint
noise identification
