Employing the Empirical Mode Decomposition To Denoise the Random Telegraph Noise

Random telegraph noise (rtn) is a stochastic phenomenon which leads to characteristic variations in electronic devices. finding features of this signal may result in its modeling and eventually removing the noise in the device. measuring this signal is accompanied by some noise and therefore we require a method to improve the signal to noise ratio (snr). as a result, the extraction of an accurate rtn is a remarkable challenge. empirical mode decomposition (emd) as a fully adaptive and signal dependent method, with no dependency to the specific function, can be an appropriate solution. in this paper, we evaluate the most recent methods and compare them with our proposed approach for the artificial and actual rtn signals. the results show the higher accuracy and efficiency by about 54%, 61% and 39% improvement in snr, mean square error (mse) and percent root mean square difference (prd) respectively for the optimized wited method. finally, an indicator to evaluate the reliability in digital circuits is introduced.