独立分量分析在自动机振动信号处理中的应用
Application of Independent Component Analysis in Automata Vibration Signal Process

自动机工作中的冲击响应多处在低信噪比和被噪声干扰的复杂振动信号中，寻求一种能在多干扰、噪声强的复杂振动信号中去除干扰信号和噪声的方法，可以提高速射武器自动机故障诊断准确率。研究了独立分量分析(independent component analysis,简称ICA)的基本理论，采用基于改进粒子群的独立分量分析算法模拟生成了仿真信号，获得了比较理想的分离效果。经实际射击数据验证了该方法的可行性，这种基于改进粒子群的独立分量分析算法在自动机结构振动信号处理方面具有较好的效果。
The fault diagnosis accuracy of a working automaton can be improved when the interfering signals and noise are removed from the complex vibration signal under multi-interference and heavy noise. The basic theory of ICA (independent component analysis) is studied, and ICA based on improved particle swarm optimization is introduced into the simulation. A comparatively satisfactory separation effect is obtained, and practical data shows that the approach is feasible.