Blind Source Separation Algorithm And Its Application

Blind Source Separation (BSS) is a new research field of the signal processing and Neural Network. Because it can recover the source signals from the observed signals without any prior knowledge of the mixing system and source signals. So it has very important theory significance and utility value in audio signal processing, wireless signal processing, biomedical signal processing, earthquake signal processing, image enhancement and so on.In this dissertation, we investigate the linear instantaneous mixture model and give two new algorithms, and discuss BSS apply in real world audio separation and blind multiuser detection. Research works in this dissertation are as follows:1. Analyzing information maximisation algorithm and extended information maximisation algorithm, aim at problem of low convergent rate of extended information maximisation algorithm, proposed an improved extended information maximisation algorithm.2. Aim at the problem that the choice of non-linear function and step influence convergent rate and stability of most of BSS algorithms. Proposed an algorithm of blind source separation based on maximum signal to noise ratio. The merit of this new algorithm is very low computational complexity and without any iterative.3. Investigation characteristic of real world audio, combine stationary for short time-scale and non-stationary for longer time-scales, proposed a time frequency domain blind source separation algorithm.4. Introduce some algorithms for blind multiuser and giving result of Computer simulation.