基于修正SUKF的固定单站无源定位跟踪算法

Fixed Single Observer Passive Localization and Tracking Algorithm Based on Improved SUKF

针对基于超球面单形采样策略的SUKF在初始误差和观测误差较大时滤波性能下降的问题,以固定单站无源定位目标跟踪为应用背景,引入后向平滑处理,并结合系统模型特点对算法作进一步的优化,提出一种修正的SUKF算法。理论分析和仿真实验表明,文中所提算法相对于传统UKF和SUKF算法,在运算开销增加不大的前提下,有效提高了算法的稳定性和滤波精度,具有一定的工程实用意义。

Filtering performance of sequential unscented Kalman filtering （SUKF） algorithm based on hyper - spheres simplex sampling would degrade due to large initial error or observation error. For fixed single observer pas- sive localization and tracking application, in order to achieve further optimization of the algorithm, by introducing backward smoothing processing, an improved SUKF algorithm is proposed based on features of system model. Theo- retical analysis and simulation results indicate that, comparing with the traditional unscented Kalman filtering （UKF） and SUKF algorithms, the proposed algorithm can effectively improve stability and filtering accuracy and only with a little computation load increasing; it has some engineering value for practical application.