Normalized Cross-cumulants for 1/f Spectrum Estimation in Gaussian and/or non-Gaussian Noise

A normalized four-order cross-cumulants (NFCC) method is proposed to suppress the effect of unavoidable Gaussian and non-Gaussian noise on decreasing the accuracy of transistor 1/f spectrum estimation. Further more, a NFCC-based parameter estimation algorithm of ARMA model for a-stable process in noisy environments is developed. Then, we also propose a cross-spectrum method effective in noisy environments based on sampled normalized four-order cross-cumulants (SNFCC) function. The effectiveness of the proposed NFCC method for 1/f spectrum estimation is demonstrated through simulation results and diodes output noise spectrum estimation results.