针对机动目标跟踪过程中量测噪声统计特性不确定的问题，提出了一种模糊自适应容积卡尔曼滤波（FACKF）算法。通过在线判断实际残差与理论残差的一致程度，利用模糊推理系统实时调整容积卡尔曼滤波的量测噪声协方差阵权值，从而修正量测噪声协方差阵，使其逐步接近真实噪声值，进而提高目标跟踪算法的自适应能力。使用角测量跟踪模型及主动雷达跟踪模型对算法效果进行仿真验证，实验结果表明，在观测噪声异常的情况下，FACKF算法比传统容积卡尔曼滤波和无迹卡尔曼滤波有更高的滤波精度与稳定性。

To deal with the uncertain statistics of measurement noise in maneuvering target tracking, a Fuzzy Adaptive Cubature Kalman Filter（FACKF） is proposed based on fuzzy inference system.By on-line judging the degree of compatibility between actual residual and theoretical residual, the measurement noise covariance of cubature Kalman filtering is adjusted in real time by using the fuzzy inference system to make it closer to the real measurement covariance gradually.Accordingly, the adaptability of the tracking algorithm is improved.Simulations using bearing-only tracking and active radar tracking model demonstrate that, compared with regular cubature Kalman filter and unscented Kalman filter, the proposed algorithm provides better filtering accuracy and stability when the observation noise is abnormal.