Novel solution for blind deconvolution based on independent component analysis

Blind deconvolution based on independent component analysis (ICA) has become the focus of intensive research due to its potential in many applications. However there exists the question that the number of sensors is usually less than the number of source signals. In this paper, by using convolution operation to the input signal, a new algorithm based on nonlinear ICA is proposed. This algorithm is applied to extract the filter in blind deconvolution. Computer simulations show the algorithm can be employed to obtain more reliable and better estimated signals for transient impulse signal extraction.