一种欠定盲源分离算法通用模型
发布时间：2019-03-25 07:45
【摘要】：针对传感器数目小于源信号数目的欠定情形,研究了基于压缩感知(CS)的欠定盲源分离(UBSS)问题。从欠定盲源分离和压缩感知的数学模型入手,在源信号具有稀疏性的前提下,将其转化为CS理论中的稀疏信号重构问题。在Sparco框架下建立了CS-UBSS两步法算法通用模型,并理论证明了该模型的有限等距特性(RIP)。仿真结果说明了该算法模型针对语音信号和图像信号的可行性与适用性,拓宽了UBSS问题的解决思路,尤其是CS理论中性能优越的重构算法可以直接应用于源信号的恢复。
[Abstract]:In the case that the number of sensors is less than the number of source signals, the problem of blind source separation (UBSS) based on compressed sensing (CS) is studied. Based on the mathematical model of under-determined blind source separation and compression sensing, the sparse signal reconstruction problem in CS theory is transformed under the premise that the source signal is sparse. In this paper, a general model of CS-UBSS two-step algorithm is established under the framework of Sparco, and the finite isometric characteristic (RIP). Of the model is proved theoretically. The simulation results show the feasibility and applicability of this algorithm model for speech signal and image signal, and widen the idea of solving UBSS problem. Especially, the reconstruction algorithm with superior performance in CS theory can be directly applied to the restoration of source signal.
【作者单位】： 景德镇陶瓷大学信息工程学院;
【基金】：国家自然科学基金(51377132)
【分类号】：TN911.7
，
本文编号：2446782
[Abstract]:In the case that the number of sensors is less than the number of source signals, the problem of blind source separation (UBSS) based on compressed sensing (CS) is studied. Based on the mathematical model of under-determined blind source separation and compression sensing, the sparse signal reconstruction problem in CS theory is transformed under the premise that the source signal is sparse. In this paper, a general model of CS-UBSS two-step algorithm is established under the framework of Sparco, and the finite isometric characteristic (RIP). Of the model is proved theoretically. The simulation results show the feasibility and applicability of this algorithm model for speech signal and image signal, and widen the idea of solving UBSS problem. Especially, the reconstruction algorithm with superior performance in CS theory can be directly applied to the restoration of source signal.
【作者单位】： 景德镇陶瓷大学信息工程学院;
【基金】：国家自然科学基金(51377132)
【分类号】：TN911.7
，
本文编号：2446782
本文链接：https://www.wllwen.com/kejilunwen/xinxigongchenglunwen/2446782.html
