Underdetermined blind source separation (UBSS) is an important issue for sparse signals, and a novel two-step approach for UBSS based on the law of large numbers and minimum intersection angle rule (LM method) is presented. In the first step, an estimation of the mixed matrix is obtained by using the law of large numbers, and the number of source signals is displayed graphically. In the second step, a method of estimating the source signals by the minimum intersection angle rule is proposed. The significance of this step is that the minimum intersection rule is better than the shortest path method, and the decomposition components can be found optimally by the former. The simulation results illustrate the effectiveness of the LM method. It has a simple principle, has good transplantation capability and may be widely applied in various fields of digital signal processing.

中文翻译：

---

基于大数定律和最小交角规则的稀疏信号欠定盲源分离

欠定盲源分离（UBSS）是稀疏信号的一个重要问题，提出了一种基于大数定律和最小相交角规则（LM方法）的新型UBSS两步法。第一步，利用大数定律得到混合矩阵的估计，并以图形方式显示源信号的数量。第二步，提出一种利用最小相交角规则估计源信号的方法。这一步的意义在于，最小相交规则优于最短路径法，而且前者可以最优地找到分解分量。仿真结果说明了LM方法的有效性。它有一个简单的原理，