机载雷达在进行空面目标探测时回波会受到强度不同的地面杂波影响,导致角度测量常常出现异常偏差,在实际应用中,其量测误差难以用一个准确的测量协方差矩阵来描述,降低了滤波效果。此外, 考虑到直角坐标系下的雷达探测属于非线性观测,文中在传统线性抗差滤波模型的基础上,提出了一种改进的自适应抗差UKF算法,以解决非线性抗差滤波问题,同时,研究了当目标发生机动时的抗差滤波工程应用问题,提出了一种合理的抗差权系数设置方法,使滤波器在具备抗差滤波的同时仍然维持对机动目标的滤波效果。实验仿真表明,通过比较不同条件下各滤波器性能,证明了该算法的有效性。

The echo from ground target detected by airborne radar will be influenced by variety intension clutter,result in variety angle measure deviation which is difficult to describe as an accuracy covariance matrix. Considering radar observation is non-linear in the rectangular coordinate system, an improved adaptively UKF algorithm for solving non-linear robust filtering problems is proposed in this paper. Moreover, the studying on target maneuver model demonstrates this algorithm can maintain filtering effect while target maneuvering. Finally, simulations by comparing respective filter performance under different scenarios demonstrate this algorithm is effective and reliable.