جداسازی منبع کور
Abstract
1-Introduction
2-Blind Source Separation Algorithm with Momentum Terms
3-Blind Source Separation Using An Adaptive Coupling Of Four Separation Systems Based On Variable Step Size And Variable Momentum Term
4-Simulation Experiment and Performance Analysis
5-Conclusions
6-Acknowledgement
7-References
Abstract
In order to improve the performance of the blind source separation (BSS) algorithm, this paper proposes a four-system adaptive coupling BSS algorithm based on variable momentum term and variable step size. The algorithm improves the convergence rate and steady-state error by adaptively coupling two medium systems. The first medium system is composed of two subsystems with different step sizes and the same large momentum term, and the convergence rate is faster. The second medium system is composed of two subsystems with different step sizes and the same small momentum term, and the steady-state error is smaller. Simulation results show that the convergence rate of the algorithm is faster and the steady-state error is smaller.
Introduction
In recent years, many new technologies have emerged in the signal processing field, including BSS technology. The essence of BSS is to estimate the source signal from the observed mixed signal according to the independent statistical characteristics of the source signal without any prior knowledge. BSS has become one of the hot techniques , which has been widely applied in wireless communication, seismic survey, ocean sonar detection, biomedical engineering, digital image processing, speech processing and so on [1,2]. Scholars had proposed many adaptive blind source separation algorithms to solve the problem between the convergence rate and steady-state error. In the early literature, a dual-system adaptive coupling algorithm was proposed, which adaptively selected the optimal separation system in each separation stage. Although the separation performance of the algorithm was effectively improved, the algorithm had the problem of convergence in stages [3].