传统的阵列干扰抑制算法假设天线阵列与干扰之间相对静止,完成干扰抑制,当阵列与干扰之间存在相对运动时,存在协方差矩阵估计快拍数不足的问题,从而导致阵列滤波性能下降。针对上述运动干扰稳健滤波问题,提出了一种算法对现有零陷展宽技术抗运动干扰方法进行改进。一方面,将零陷展宽技术与导数约束方法相结合,用得到的新协方差矩阵求自适应权值;另一方面,为了提高自适应波束形成对噪声和误差的稳健性,引入了对角加载技术,加载量通过寻找协方差矩阵的特征值曲线拐点得到,有效避免了加载量过大导致的干扰信号被淹没以及加载量过小导致噪声分量引起性能损失的问题。该算法对零陷展宽系数选取依赖性较低,且对角加载量不需要设置固定的经验值,有利于稳健运动干扰滤波,适用于实际工程中先验信息不足条件下的自适应阵列抗运动干扰情况。

Traditional interference suppression methods assume that the array antenna is relatively stationary to the interference. However, when the relative movement between the array and the interference is not valid, the covariance matrix estimation cannot be achieved in real time, which leads to a decrease in filtering performance. In view of the robust filtering problem of the abovementioned problem, the presented method improves the existing interference suppression method. On the one hand, the new covariance matrix obtained by combining zero shift extension technique with the derivative constraint method is used to obtain the adaptive weights. On the other hand, in order to improve the beamforming robustness to noise and error, the diagonal loading technique whose load value is obtained by the characteristic curve inflection point of the covariance matrix is employed, which avoiding signal submerging caused by big loading value and performance loss caused by small loading value. The proposed algorithm has low dependence on the zero broadening coefficient, meanwhile, the value of diagonal loading does not need to set a fixed value, which is conducive to the steady motion interference filter and suitable for the case of few prior information.

TN957.52

国家自然科学基金资助课题(61701395)