Extraction of impacting signals using blind deconvolution

Two algorithms, the objective function method (OFM) and the eigenvector alg orithm (EVA), of the inverse filter estimation are applied to extract the i mpulsive impacting signals. Both algorithms maximize the estimation of the cumulants with observed data only. The resolution of the reconstructed sign als are improved from the observed (measured) signals. Therefore, the "clea n" impacting signals can be compared to the known impacting phenomena to de termine their origin. The performance of both algorithms are investigated a nd these appear to perform well. (C) 2000 Academic Press.