[1]郭凌飞,张林波.通过CWLS-DL优化St-OMP算法的盲信号重构[J].应用科技,2019,46(03):40-45,50.[doi:10.11991/yykj.201809021]
 GUO Lingfei,ZHANG Linbo.Blind signal reconstruction of St-OMP algorithm optimized by CWLS-DL[J].Applied science and technology,2019,46(03):40-45,50.[doi:10.11991/yykj.201809021]
点击复制

通过CWLS-DL优化St-OMP算法的盲信号重构(/HTML)
分享到:

《应用科技》[ISSN:1009-671X/CN:23-1191/U]

卷:
第46卷
期数:
2019年03期
页码:
40-45,50
栏目:
现代电子技术
出版日期:
2019-04-29

文章信息/Info

Title:
Blind signal reconstruction of St-OMP algorithm optimized by CWLS-DL
作者:
郭凌飞 张林波
哈尔滨工程大学 信息与通信工程学院, 黑龙江 哈尔滨 150001
Author(s):
GUO Lingfei ZHANG Linbo
College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China
关键词:
压缩感知盲信号重构信号重构精度计算复杂度稀疏成分分析加权最小二乘字典学习正交匹配追踪算法
Keywords:
compressed sensingblind signal reconstructionsignal reconstruction accuracycomputational complexitysparse component analysisweighted least squaresdictionary learningorthogonal matching pursuit algorithm
分类号:
TN911.7
DOI:
10.11991/yykj.201809021
文献标志码:
A
摘要:
针对稀疏成分分析理论的“两步法”中的源信号重构算法改进,提出一种由相关性加权最小二乘字典学习法与分段正交匹配追踪算法组合的算法,能够解决带权重信号误差的F-范数最小化问题,并通过增加单次迭代的原子数改变算法复杂度。将此组合算法用于语音信号的盲源分离仿真实验,完成源信号重构。实验结果表明,用该组合算法重构的信号,能在保证提高重构精度的同时,与算法复杂度存在良好的折中。无噪声环境下该组合算法的性能为最佳,有噪声环境下可达到信号重构要求的最小信噪比约为17~18 dB。
Abstract:
Considering the source signal reconstruction algorithm, which is improved in the "two-step method" of sparse component analysis theory, this paper proposes an algorithm combining correlation weighted least squares dictionary learning method and stagewise orthogonal matching pursuit algorithm. It solves the F-norm minimization problem that has weighted signal errors and changes the complexity of the algorithm by increasing the number of atoms in a single iteration. This combination algorithm is used to simulate the blind source separation of speech signals and complete source signal reconstruction. The experimental results show that the signal reconstructed by the combined algorithm can guarantee the improvement of reconstruction accuracy and the compromise of algorithm complexity. The performance of this combined algorithm is optimal in a noiseless environment. The minimum signal-to-noise ratio that can achieve signal reconstruction in a noisy environment is about 17-18 dB.

参考文献/References:

[1] 许华杰, 何敬禄, 胡小明. 基于块稀疏度估计的压缩感知自适应重构算法[J]. 计算机应用研究, 2018, 35(1):305-308, 320
[2] BARNICH O, VAN DROOGENBROECK M. ViBe:a universal background subtraction algorithm for video sequences[J]. IEEE transactions on image processing, 2011, 20(6):1709-1724.
[3] ELAD M. Optimized projections for compressed sensing[J]. IEEE transactions on signal processing, 2007, 55(12):5695-5702.
[4] 何继爱, 刘向阳. 压缩感知理论及其在盲源分离中的应用[J]. 测控技术, 2016, 35(11):149-152
[5] 王粒宾, 崔琛, 李莹军. 基于加权最小二乘的字典学习算法[J]. 系统工程与电子技术, 2011, 33(8):1896-1900
[6] 张红丽. 基于稀疏性增强的欠定盲源分离算法研究[D]. 秦皇岛:燕山大学, 2016:30.
[7] 徐健, 常志国, 张小丹. 采用交替K-奇异值分解字典训练的图像超分辨率算法[J]. 武汉大学学报(信息科学版), 2017, 42(8):1137-1143
[8] FU Weihong, CHEN Jiehu, YANG Bo. Source recovery of underdetermined blind source separation based on SCMP algorithm[J]. IET signal processing, 2017, 11(7):877-883.
[9] 叶娅兰, 何文文, 程云飞, 等. 面向压缩感知的基于相关性字典学习算法[J]. 电子科技大学学报, 2017, 46(5):703-708
[10] 练秋生, 王小娜, 石保顺, 等. 基于多重解析字典学习和观测矩阵优化的压缩感知[J]. 计算机学报, 2015, 38(6):1162-1171
[11] 汪浩然, 夏克文, 牛文佳. 分段正交匹配追踪(StOMP)算法改进研究[J]. 计算机工程与应用, 2017, 53(16):55-61
[12] CHEN P, PENG D H, ZHEN L L, et al. Underdetermined blind separation by combining Sparsity and independence of sources[J]. IEEE access, 2017, 5:21731-21742.

相似文献/References:

[1]安澄全,彭军伟.基于混合优化的平滑l0 压缩感知重构算法[J].应用科技,2013,40(05):23.[doi:10.3969/j.issn.1009-671X.201211020]
 AN Chengquan,PENG Junwei.Sparse recovery using smoothed l0 based on hybrid optimization algorithm[J].Applied science and technology,2013,40(03):23.[doi:10.3969/j.issn.1009-671X.201211020]

备注/Memo

备注/Memo:
收稿日期:2018-09-25。
作者简介:郭凌飞,男,硕士研究生;张林波,女,副教授
通讯作者:郭凌飞,E-mail:guolingfei@hrbeu.edu.cn
更新日期/Last Update: 2019-04-29