The estimation of average lifetime has been considered as a serious reliability concern for VLSI components. However, with the tremendous downscaling of MOSFET devices, it becomes very difficult to predict such measure with satisfactory precision due to the amplified parasitic effects at the deep scale level. The present work aims at investigating the efficiency of adaptive neuro-fuzzy inference system approach as a predictor of junctionless gate all around MOSFET device lifetime. Our predictions are compared versus the numerical results obtained by ATLAS simulator, where a good agreement is obtained.

Timgad Basin is a part of the North-East Algerian Saharan Atlas. It is located at the northern of the Aurès Mountain. It is bordered in the North by Bou Arif, Amrane and Tagratine mountains, while in the South it is limited by Chelia, Rdam and Nerdi Bouhmar mountains. It is shaped as a synclinal gutter oriented East-West and covers an area of about 1000 Km2 . The climate is semi-arid (cold winter and hot summer) with average annual rainfall not exceeding 350 mm. The basin is composed by varied lithology, showing large folded structures down sloping towards the center, which favor groundwater storage. The Miocene sandstone aquifer is the most important. It is characterized by a wavy shape, which form series of shale-marl filling depressions. Recent wells drilled in the basin edges and center, put in evidence its artesian character. The piezometric survey shows a groundwater convergent flow oriented towards the basin east (Bou el Freis), likewise the sandstone aquifer is marked by the presence of a West-East drainage axis which separates the basin northern part characterized by a low reservoirs hydraulic capacity from the southern part distinguished by the groundwater relative abundance.

Electrocardiogram (ECG) signal with artifact might lead to wrong analysis. In this paper, discrete cosine transform (DCT) denoising method for artifacts removal from ECG signals is proposed. Based on the energy properties of DCT transform is located in the low frequency coefficients and baseline wander (BW) noise is also a low frequency signal; therefore we deduce that the firsts coefficients of the DCT of the ECG signal is corresponds to the BW artifact in it. In our approach the BW is estimated from the low coefficients of the DCT transform, and the corrected ECG signal is obtained by its subtraction from noisy ECG signal in the time domain. The simulation results are done based on real ECG signals taken from MIT-BIH arrhythmia database whose corrupted with the BW noise from MIT-BIH noise stress test database. A comparison analysis of the median filtering (MF), discrete wavelet transform (DWT) decomposition and DCT transform for ECG signal denoising is done. The obtained results showed that our proposed method based on DCT transform gives better performances compared to MF and DWT methods in terms of Signal to Noise Ratio (SNR), and Pearson correlation parameter.