基于频域低秩矩阵补全的气象雷达WTC抑制算法

Weather Radar WTC Suppression Algorithm Based on Low-Rank Matrix Completion in Frequency Domain

由于风力涡轮机杂波(WTC)的时变性严重降低气象雷达探测性能,文中根据气象信号频域低秩特性,采用矩阵补全(MC)算法高精度恢复被WTC污染的气象信号。首先,将回波信号从时域变换到多普勒频域,对被WTC污染的距离单元进行填零并构造低秩Hankel矩阵,以获得零元素随机分布的观测矩阵;其次,利用非精确增广拉格朗日乘子法(IALM)实现非凸秩函数逼近,完成矩阵元素的补全;最后,根据恢复出的Hankel矩阵得到高精度恢复的气象信号并实现WTC的抑制。在实测气象雷达数据中加入WTC,实验结果表明,与工程常用的频域多重二次插值算法相比,矩阵补全后气象信号的径向速度和谱宽的估计精度分别提高了约22.25%、31.48%。

The time-variation of wind turbine clutter(WTC) seriously reduces the detection performance of weather radar. In this paper, due to the low-rank characteristics of weather signal in the frequency domain, the matrix completion(MC) algorithm is introduced to restore the weather signal polluted by WTC precisely. First, in the Doppler frequency domain of echo signal, the range cells polluted by WTC are zero-filled to construct a low-rank Hankel matrix and an observation matrix with zero elements randomly distributed can be obtained. Secondly, the Grange multiplier method(IALM) is applied to realize the approximation of non-convex rank functions and complete the completion of matrix elements. Finally, according to the recovered Hankel matrix, the weather signal can be recovered precisely and the suppression of WTC can be realized. The experimental results show that, compared with the frequency-domain multiple quadratic interpolation algorithm commonly used in engineering, the estimation precision of radial velocity and spectral width of weather signals after matrix completion are improved by 22.25% and 31.48% respectively.