本文提出了一种在非线性量测环境下对机动目标的解耦自适应滤波算法。在视角直角坐标系中实现了一种修正的输入估计方法。由于所提出的这种滤波器有解耦结构,因此在不降低性能的前提下可以大幅度减少计算量。同时,由于伪量测噪声之间的交叉相关性较小,因此可以提高机动探测的精度。 This paper presents a decoupled adaptive filtering algorithm for maneuvering targets in a nonlinear measurement environment. A modified input estimation method is implemented in the Cartesian coordinate system. Due to the proposed decoupling structure of this filter, the amount of computation can be drastically reduced without sacrificing performance. At the same time, the accuracy of maneuvering detection can be improved due to the small cross-correlation between the pseudo-measurement noise.