“… Similarly, the above result is also true for U (1, 2) − z −1 U (1, 3) and U ( i , j ) − z −1 U ( i , j + 1) per i = 1, 2, …, L + 1 and j = 1, 2, …, L , in which L = M /2. Therefore, a matrix of lower rank from U , that is, T L × ( L + 1), can be formed as follows 2 In order to restore signal components in the adaptive process, the gain of the sub‐array formed by the L + 1 elements in line with θ s should be proved.…”