| Summary: | 碩士 === 國立清華大學 === 通訊工程研究所 === 103 === Compared with traditional data acquisition, compressive sensing can reconstruct signal from far fewer samples than the traditional method. One of the critical problem is to reconstruct the original signal from the compressed measurement. There are many popular reconstruction algorithms such as Basis pursuit(BP), Matching Pursuit(MP), Orthogonal Matching Pursuit(OMP), Compressive Sampling Matching Pursuit(CoSaMP); Despite their good performance, these algorithms require the sparse level as a prior information for reconstruction. This thesis demonstrates a recursively updating function based on Subspace Pursuit algorithm, which is a famous algorithm for signal recovery. Our proposed algorithm is called Sparsity Updating Subspace Pursuit(SUSP). The main contribution is that the proposed SUSP can solves the sparse signal problem when the sparsity is not available. This property makes the algorithm to solve many practical compressive sensing problem when the sparse level, the number of non-zero elements of a signal is not given. While SUSP can deal with unknown sparsity problem, results show that it outperforms many existing famous greedy algorithm when the sparsity is pre-known.
|
|---|
