单站纯方位目标跟踪是水下目标攻击领域的一大难点，尤其是针对快速目标短时观测这一情形。针对单站纯方位目标跟踪中卡尔曼滤波类方法对初值选取比较敏感的问题，提出一种基于逆向平滑滤波的初值优化方法。通过对卡尔曼滤波的逆向平滑处理，增加对观测数据的正反向运用，对滤波初值进行优化，以达到在后续滤波中减少噪声影响、降低估计误差的效果。在不同的仿真条件下，针对无迹卡尔曼滤波和扩展卡尔曼滤波2种滤波方法分别进行多次仿真试验。结果表明，该优化方法对2种滤波方法都有缩小估计误差、提高估计精度的效果，降低了对滤波初值选取的依赖性。该方法对误差的降低率和实时性与所选取的逆向滤波步数有关，算法复杂度不高，并且在较差的初始距离误差、初始速度误差或者方位估计精度的情况下，对降低滤波的估计误差依然有效。
Single station bearings only target tracking is a major difficulty in the field of underwater target attack, especially in the case of short-term observation of fast target. In order to solve the problem that Kalman filter is sensitive to initial value selection in single station bearings only target tracking, an initial value optimization method based on inverse smoothing filter is proposed. Through the reverse smoothing of Kalman filter, the positive and negative utilization of the observation data is increased, the initial value of the filter is optimized, so as to achieve the effect of reducing noise influence and estimation error in the subsequent filtering. Under different simulation conditions, the two filtering methods of Unscented Kalman filter and Extended Kalman filter are simulated for several times. The results show that this optimization method can reduce the estimation error and improve the estimation accuracy for both filtering methods, and reduce the dependence on the selection of initial filter value. The reduction rate and real-time performance of this method are related to the selected steps of reverse filtering, and the complexity of the algorithm is not high. In the case of poor initial distance error, initial velocity error or azimuth estimation accuracy, it is still effective to reduce the estimation error of filtering.
2020,42(9): 140-147 收稿日期：2020-02-10
DOI：10.3404/j.issn.1672-7649.2020.09.027
分类号：TP391
作者简介：郑艺(1992-),女,博士研究生,研究方向为水下目标定位跟踪
参考文献：
[1] MILLER A B, MILLER B M. Underwater target tracking using bearing-only measurements[J]. Journal of Communications Technology and Electronics, 2018, 63(6): 643–649
[2] GUSTAFSSON F, HENDEBY G. Some relations between extened kalman filter and unscented kalman filter[J]. IEEE Transactions on Signal Processing, 2013, 60(2): 545–555
[3] SHALOM Y, LI X R, THIAGALINGAM K. Estimation with applications to tracking and navigation[M]. New York: Wiley, 2001: 381-394.
[4] JULIER S J, UHLMANN J K. Unscented filtering and nonlinear estimation[J]. Proc. of the IEEE, 2004, 92(3): 401–422
[5] GORDON N J, SALMOND D J, SMITH A F M. Novel approach to nonlinear/non-Gaussian Bayesian state estimate[J]. IEEE Proceeding of Radar, Sonar and Navagation, 1993, 140(2): 107–113
[6] MOHAMMAD Jannati, TOLE Sutikno, NIK Rumzi Nik Idris, et al. High performance speed control of single-phase induction motors using switching forward and backward EKF strategy[J]. International Journal of Power Electronics and Drive System (IJPEDS), 2016, 7(1): 17–27
[7] 张刚兵, 刘渝, 胥嘉佳. 基于UKF的单站无源定位于跟踪双向预测滤波算法[J]. 系统工程与电子技术, 2010, 32(7): 1415–1418
Zhang Gangbing, Liu Yu, Xu Jiajia. Unscented Kalman filter with forward-backward prediction for signal observer passive location and tracking[J]. Systems Engineering and Electronics, 2010, 32(7): 1415–1418
[8] BENEDIKT T. BÖNNINGHOFF, ROBERT M. Nickel, STEFFEN Zeiler, et al. Unsupervised classification of voiced speech and pitch tracking using forward-backward Kalman filtering[C]. ITG-Fachbericht 267: Speech Communication. Paderborn, Germany: VDE, 2016: 46-50.
[9] WU B, LIU P Y. A new cubature Kalman filter improved by backward iterative algorithm[C]// International Conference on Intelligent Computation Technology and Automation (ICICTA). Nanchang, China: IEEE press, 2015: 44-48.
[10] MARK L. Psiaki. Backward-smoothing extended Kalman filter[J]. Journal of Guidance, Control, and Dynamics, 2005, 28(5): 885–894
[11] 霍光, 李冬海. 基于后向平滑容积卡尔曼滤波的单站无源定位算法[J]. 信号处理, 2013, 29(1): 68–74
HUO Guang, LI Donghai. A single observer passive location algorithm based on backward-smoothing cubature Kalman filter[J]. Journal of Signal Processing, 2013, 29(1): 68–74
