Abstract-In this paper, an improved L 1 -SVD algorithm based on noise subspace is presented for direction of arrival (DOA) estimation using reweighted L 1 norm constraint minimization. In the proposed method, the weighted vector is obtained by utilizing the orthogonality between noise subspace and signal subspace spanned by the array manifold matrix. The presented algorithm banishes the nonzero entries whose indices are inside of the row support of the jointly sparse signals by smaller weights and the other entries whose indices are more likely to be outside of the row support of the jointly sparse signals by larger weights. Therefore, the sparsity at the real signal locations can be enhanced by using the presented method. The proposed approach offers a good deal of merits over other DOA techniques. It not only increases robustness to noise, but also enhances resolution in DOA estimation. Furthermore, it is not very sensitive to the incorrect determination of the number of signals and can primely suppress spurious peak in DOA estimation. Simulation results are shown that the presented algorithm has better performance than the existing algorithms, such as MUSIC, L 1 -SVD algorithm.

A new speech denoising method that aims for processing corrupted speech signal which is based on the sparse representation theory of speech signal. In this paper, we train a composite dictionary consisting of the concatenation of the speech dictionary and the noise dictionary by using the K-SVD algorithm. Noise is divided into structured and unstructured noise in this paper. For structured noise, we train speech and noise dictionary firstly, and then according to the different coherence between speech and noise, we use LARC algorithm with a suitably chosen residual coherence threshold to realize the separation of the speech and the noise. For unstructured noise, we only need speech dictionary to extract the clean speech from corrupted speech. Experiments indicate that the proposed method gives better enhancement results in terms of quality measures of speech. The proposed method outperforms the universal dictionary speech enhancement algorithm.

No abstract

On the basic of multiple populations of immune algorithm and clonal selectionalgorithm, the purpose of this paper is to further improve the detectionefficiency and reducing the false alarm rate. This paper uses the kddcup99 dataset as the experimental data set, and chooses four types of attack data groupof experiment data set as initial population of multiple populations of clonalselection algorithm, through the algorithm to create the optimal model. Basedon the principle of normal data larger than the abnormal data, in turn,experimental data set matched with the normal data set and the optimal model bythe improved R matching algorithm. The results of this paper show that thedetection rate increased significantly