UKF算法在纯方位目标运动分析中的应用

Application of UKF in Bearings-only Target Motion Analysis

针对传统算法在解决纯方位目标运动分析时存在的有偏、收敛速度慢或发散等不足,该文将无味卡尔曼滤波(UKF)算法应用到纯方位目标运动分析中。由于UKF在处理非线性问题时表现良好,以及不需要计算Jacobian矩阵或Hessian矩阵,实现起来比较方便。根据无味变换的基本原理给出了滤波过程的具体计算步骤并进行了仿真计算。理论分析和仿真结果表明,UKF的性能相当于二阶高斯滤波器,它在纯方位目标运动分析中的滤波精度、稳定性和收敛时间都优于传统算法。

The traditional algorithms applied in bearings-only target motion analysis （BD-TMA） have some shortages or disadvantages such as biased, slow convergence or divergence. To solve the problem, unscented Kalman filter （UKF） is applied in bearings-only target motion analysis. The realization of UKF is comparatively simple and convenient because UKF is feasible in processing nonlinear problems and doesn＇ t compute the Jacobian Matrix or Hessian Matrix. In this paper, the UKF is applied in BO-TMA through bringing forward the filtering steps of algorithm based on the principle of unscented transformation （UT）. Theoretical analysis and simulation result indicate that the UKF has the same performance to 2-rank Gauss filter and has better performance than traditional algorithms in precision, stability and convergence time when it is applied in BO-TMA.