BLIND SEPARATION OF MECHANICAL SOURCE SIGNALS BASED ON THE ESTIMATION OF SOURCE NUMBERS
Zhi Li · 2011
In the traditional blind separation method of mechanical sources,the number of mechanical sources is always assumed to be known.However,the mechanical sources do not suffice to this condition.In order to overcome this deficiency,combining of Bayesian inferring and automatic relevance determination (ARD),a new estimation method of mechanical sources based on Bayesian independent component analysis is proposed.In the proposed method,the optimal number of mechanical sources is determined by the evidence comparison of different models.The simulation and experiment results show that the proposed method is very effective and has good robustness.