Exploiting Temporal Structures of Cyclostationary Signals for Data-Driven Single-Channel Source Separation
Exploiting Temporal Structures of Cyclostationary Signals for Data-Driven Single-Channel Source Separation
复制标题
DOI：
10.1109/mlsp55214.2022.9943311
复制
发表时间：
2022-08
期刊：
影响因子：
--
通讯作者：
Gary C. F. Lee;A. Weiss;A. Lancho;Jennifer Tang;Yuheng Bu;Yury Polyanskiy;G. Wornell
中科院分区：
文献类型：
--
作者：
Gary C. F. Lee;A. Weiss;A. Lancho;Jennifer Tang;Yuheng Bu;Yury Polyanskiy;G. Wornell
本文研究了单通道信号源分离问题，重点研究了循环平稳信号，这类信号特别适用于多种应用领域。与经典的SCSS方法不同，我们考虑的设置只有源的例子，而不是他们的模型，激发了数据驱动的方法。对于具有基本循环平稳高斯成分的源模型，我们为任何基于模型或数据驱动的分离方法建立了可达到的均方误差（MSE）的下限。我们的分析进一步揭示了最佳分离的操作和相关的实施挑战。作为一种在计算上有吸引力的替代方案，我们提出了一种使用U-Net架构的深度学习方法，该方法与最小MSE估计器具有竞争力。我们在模拟中证明，适当的域知情的架构选择，我们的U-Net方法可以接近最佳性能，大大降低了计算负担。
We study the problem of single-channel source separation (SCSS), and focus on cyclostationary signals, which are particularly suitable in a variety of application domains. Unlike classical SCSS approaches, we consider a setting where only examples of the sources are available rather than their models, inspiring a data-driven approach. For source models with underlying cyclostationary Gaussian constituents, we establish a lower bound on the attainable mean-square-error (MSE) for any separation method, model-based or data-driven. Our analysis further reveals the operation for optimal separation and the associated implementation challenges. As a computationally attractive alternative, we propose a deep learning approach using a U-Net architecture, which is competitive with the minimum MSE estimator. We demonstrate in simulation that, with suitable domain-informed architectural choices, our U-Net method can approach the optimal performance with substantially reduced computational burden.