基于UKF-IMM的方位-到达时间空-海机动目标的跟踪

Air-to-Sea Maneuvering Target Tracking by Bearing and Time-of-arrival Measurements Based on Interacting Multiple Model Augmented with Unscented Kalman Filtering Method

探讨了机动目标无源跟踪(PMTT)中基本的非线性估计问题,介绍了无味卡尔曼滤波(UKF)算法的设计思想与具体实现,用UKF代替原交互多模(IMM)中的扩展卡尔曼滤波(EKF)得到UKF-IMM估计算法,特别针对空—海单站只测方位—到达时间PMTT问题分别应用UKF-IMM和EKF-IMM进行了对照研究,建立了问题的离散非线性滤波估计模型,设计了典型的应用场景,给出了Monte Carlo仿真结果;表明UKF-IMM滤波算法具有更高的估计精度和收敛特性,滤波的一致性较好,更加适合解决非线性较为严重的PMTT问题。

The nonlinear filtering problem is identified firstly as the groundwork embedded in PMTT (Passively Maneuvering Target Tracking). The unscented transformation (UT) and unscented Kalman filtering (UKF) are then introduced, including their design considerations and specific algorithms. By substituting UKF for EKF, the UKF-IMM (Interacting Multiple Model) estimation algorithm is proposed to resolve the problem of PMTT. Particular attention is paid to the air-to-sea PMTT with single observer based on bearing and time-of-arrival measurements (BTO-SOPMTT). The discrete-time models are formulated and a typical scenario is depicted. The contrast results of Monte Carlo simulations between UKF-IMM and EKF-IMM have demonstrated that the former is more feasible to the severely nonlinear PMTT application by virtue of its favorable consistency with higher accuracy and stronger convergence.