实验表明，GBCD 算法能够很好地解决 DOA 中点源很近或者快照很少的问题。但是 GBCD 在噪声干扰下的性能没有明确。论文证明在噪声‖ E ‖2尘ε的情况下，当感知矩阵 A 满足δK ＋1＜1K ＋1，并且 X 非零元素的满足mini ∈Ωmin l ∈｛1… t｝｜Xil ｜＞2（1＋δK ＋1）ε（1－ K ＋1δK ＋1），那么 GBCD 方法能够精确恢复 K‐稀疏信号 X 的支撑集。%Experiments show that GBCD has good performance when the point sources are intensive ,or the number of snapshots is small ,but the function of GBCD under the interfere of noise is not illustrated .In this paper ,when ‖ E ‖ 2 尘 ε , if the sensing matrix A satisfies the RIP with δK + 1 < 1K + 1 ,then the GBCD can exactly recover the support of X in K itera‐tions ,under the constraint on the minimum magnitude of the nonzero elements of X .
展开▼
