非线性滤波算法在无源双基地雷达目标跟踪中的比较研究

针对无源双基地雷达目标跟踪问题,仿真分析了EKF、UKF、CDF等几种非线性滤波算法的状态估计性能。同时,基于后向平滑估计原理,利用当前观测数据平滑估计前时刻状态变量的均值和方差,提出了一种改进的UKF(CDF)滤波算法-BSUKF/CDF。仿真结果表明,在理想高斯白噪声情况下,UKF/CDF及BSUKF/CDF的跟踪性能相近,但均明显优于EKF;但若考虑角闪烁噪声,BSUKF/CDF的跟踪性能则优于UKF/CDF及EKF。

Comparison of Nonlinear Filtering for Passive Bistatic Radar Target Tracking

For passive bistatic radar target tracking problem, the performances of several nonlinear filtering algorithms such as EKF, UKF and CDF were simulated and analyzed. Also, a new nonlinear filtering algorithm called BSUKF/CDF based on backward-smoothing principle was proposed. In BSUKF/CDF algorithm, the current observation was used to smoothly estimate the previous mean and covariance of the state variable. The simulation results show that in Gaussian environment, BSUKF/CDF and UKF/CDF have almost the same tracking performance, and both perform better than EKF; however in angle glint noise environment, BSUKF/CDF perform much better than UKF/CDF and EKF.