Improved FxGMCC algorithm based on the sigmoid-sinh segmentation function
To resolve the conflict between fast convergence and low steady state error of the filter-x gen-eralised maximum entropy criterion(FxGMCC)algorithm in dealing with noisy active control,a varied step-size FxGMCC algorithm based on the sigmoid function(SFxGMCC)and a varied step-size FxGMCC algorithm based on the sinh function were respectively presented by using the sigmoid function and the sinh function(ShFxGMCC).Then a sigmoid-sinh segmented function SSFxGMCC algorithm is proposed by analysing the variable step size property of SFxGMCC algorithm and ShFxGMCC algorithm.In order to verify the property of the SSFxGMCC algorithm,the property simulation tests of the SSFxGMCC algorithm with the FxGMCC algo-rithm and the SFxGMCC algorithm with the ShFxGMCC algorithm are carried out by using the alpha-stable dis-tribution noise and the mixed noise as the input signals,respectively.The outcomes reveal that,in contrast to the FxGMCC algorithm,using the SSFxGMCC algorithm in dealing with the alpha stable distribution noise and the mixed noise,the average noise reduction is improved by 16.5％and 17.4％,respectively,on the premise that the convergence speed has been enhanced.It can be seen that the SSFxGMCC algorithm effectively solves the problem that fast converse and low steady state error cannot be satisfied simultaneously by the FxGMCC al-gorithm when dealing with noise active control.
万方数据
维普
