摘要
误差配准是雷达组网融合跟踪系统的重要环节,而可观测度又是决定配准效果的关键因素。现实中,一般难以定量描述系统的可观测度,因此本文首先提出了一种间接的可观测度检测方法,用于衡量未知可观测度条件下系统误差估计的精度。通过分析可观测度较低的情况,得出在低可观测度条件下随机误差是导致系统误差估计大幅波动的主要原因,为此,利用一维离散小波变换的低通滤波特性,有效抑制原始观测数据中的随机误差,再基于滤波后的数据集,采用扩展卡尔曼滤波方法实现系统误差估计。仿真实验表明,本文算法能够有效检测系统的可观测度,并且在低可观测度下,对系统误差的估计精度明显优于已有算法。
Registration is an important process for data fusion and target tracking in radar networking system, and observ-ability is the key factor which determines the registration effect. Actually, itls hard to describe the observability quantitatively, so an indirect observability detection method is proposed in this paper,which focuses on how to measure the precision of systematic error estimation in unknown observability condition. According to the analysis of the result in low observability condition, it is established that random error is the main cause of large variety of systematic error estimation. Therefore, the original measurements are processed by lowpass filtering which is realized by one-dimensional discrete wavelet transformation,and it can suppress the additive measurement noise; then the extended Kalman filtering is used to estimate the systematic error based on the new filtered data set. The simulation results show that the proposed method can evaluate the observability of the system efficiently and has higher registration accuracy than the old methods in low observability condition.
出处
《现代电子技术》
2008年第9期57-60,共4页
Modern Electronics Technique
关键词
系统误差估计
可观测度
检测
低通滤波
扩展卡尔曼滤波
systematic error estimation
observability
detection
lowpass filtering
extended Kalman filtering
