中文摘要：

独立分量分析（ICA）是盲分离的核心技术,是信号处理领域的一种新的发展。FastICA是独立分量分析中收敛速度较快的算法,因为它的收敛速度快且要求内存空间小而备受关注,但存在步长μ选取不当可能导致算法收敛速度减慢甚至不收敛的问题。为了克服其缺点,在基于峭度的FastICA算法的基础上增加精确线性搜索优化技术来求μ,使改进后的算法收敛速度更快且不需要手动来选择步长参数。编制相应的matlab程序,将改进的算法用于语音信号分离,验证了它的高效性。

英文摘要：

Independent component analysis（ICA）is the kernel technology of blind signal processing and a new development of signal processing technology.FastICA is a fast algorithm of ICA.As its convergence is very fast and the required memory is small,it has attracted broad attraction,but if step-size μ was chose incorrectly,the algorithm may be having slow convergence even no convergence.To overcome the drawbacks,exact line search was imposed on this fundament for μ.The improved algorithm not only has high convergence speed and the step size has no need to choose by people.It makes the correspond matlab programs that is used to separate the audio signal,the experimental results showthat it is an efficient method.