A new method for approximation of physiological signals

A new robust method for the approximation of physiological signals is proposed. The noise must satisfy only one condition: that the probabilities for positive and negative values are equal. The authors' results illustrate that the proposed method provides a more accurate approximation of signals than either the least square error or the least absolute deviation method in cases that include unknown noise density     

Estimation of slowly changing components of physiological signals

A method for the estimation of slowly changing components of physiological signals is presented in this communication. The method is based on a sequential approximation of slowly changing components by a low-order polynomial function. The polynomial coefficients are obtained by minimizing the distance between the expected zero crossing density (ZCD) value of the fast components of the physiological signal and the estimated ZCD value of these components. The method has been tested and preliminary results were satisfactory     

Split Burg and covariance lattice algorithms

The number of multiplications in split lattice algorithms, related to the autocorrelation matrix, is reduced by 30 percent in comparison with classical lattice algorithms. Split version of the Burg method (BM) and the covariance lattice method (CLM) are presented. As in previously published split algorithms the number of multiplications in the proposed algorithms is reduced