摘要
针对机动目标跟踪中固定区间平滑估计算法对噪声相关性考虑不完全的问题,提出了一种具有一般相关过程噪声与量测噪声的离散线性系统最优固定区间平滑估计算法.该算法通过将固定区间内全部量测进行集中式扩维,并对误差传递进行分析,从而精确给出了误差间的相关性,在线性无偏最小方差意义下对系统状态进行递推估计.与不考虑相关性的卡尔曼平滑算法以及仅考虑量测相关性的正、逆向滤波融合平滑估计算法相比,新算法在噪声的高斯分布假设下是最优的,且随噪声相关性增强其优越性越明显.仿真结果表明,在相关系数为0.36时,新算法的位置跟踪均方根误差比不考虑相关性和仅考虑量测相关性的平滑估计算法可降低38%.
In view of on the incomplete consideration of noise correlation in the fixed-interval smoothing algorithm for maneuvering target tracking, an optimal fixed-interval smoothing algo- rithm is proposed for discrete-time linear system with general correlated measurement noises and process noises. Based on the linear unbiased minimum variance estimation theory, the new algorithm estimates the system states recursively by using the centralized expanding-dimension method with all measurements in the fixed interval, and calculates the correlations between the errors precisely using analysis of the error transfer property. Compared with the uncorrelated Kalman smoothing algorithm and the forward-backward filtering based fusion-smoothing algorithm in which only the measurement noises correlation is considered, the new algorithm is the best one under the hypothesis of Gauss distribution. Bigger the correlation coefficient is, more obvious the superiority of the new algorithm is. Simulation results show that when the correlation coefficient is 0.86, the root mean square error of position tracking of the new algorithm is decreased 38% or more compared with the uncorrelated and measurement correlated Kalman smoothing algorithm.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2008年第8期954-957,1039,共5页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目(60602025)
关键词
目标跟踪
相关噪声
定区间平滑
target tracking
correlated noise
fixed-interval smoothing
