Single stuck-at-faults detection using test generation vector and deep stacked-sparse-autoencoder

Abstract

This paper proposed a new method for testing digital circuits without hardware implementation. This data-based method detects hundreds of single stuck-at faults in the ALU circuits, utilizing deep stacked-sparse-autoencoder (SSAE). ATALANTA software is one of the free automatic test pattern generation tools which cover faults in high accuracy. Test vectors which are extracted from bench circuits via ATALANTA software are the key point of the paper. Fault detection is introduced as a two-class problem. SSAE network is trained using the test vectors. Dimension reduction is done automatically in SSAE. Network performance is tested by changing sparse coefficients, number of stacked autoencoder and data augmentation. The results of this step are compared with the traditional multilayer perceptron classification. In this method, unlike SSAE, a manual method of reducing the dimension and extracting the feature is used. Fault coverage of ATALANTA software is over than 94%. Finally, the results obtained from the deep neural network show its significant performance in the circuit faults detection automatically.

Fruit category classification by fractional Fourier entropy with rotation angle vector grid and stacked sparse autoencoder

“…N R FRFE features were sent to a custom deep SSAE, which is one of the deep neural network models. The SSAE has been successfully applied to electricity theft detection (Huang & Xu, 2021), underwater heterogeneous data processing (Wang et al, 2020), stuck‐at‐faults detection (Malihi & Malihi, 2020), and so forth.…”

A Simple Stuck-at-faults Detection Method in Digital Combinational Circuits

“…In this article we will consider single damages of the stuck-at-fault (0/1) type, that mainly occur in digital combinational devices . To detect such faults, various methods and algorithms [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16] are known, for example: Truth Table and Fault Vatrix method based on the Boolean derivatives (Boolean difference), the Path Sensitization method and the Searching D-Algorithm based on the test patterns (PODEM), Brach-and-Bount method, etc . However, the vast majority of them are based on an analytical approach and are characterized by their complexity and cumbersomeness of practical implementation, that grows with the increase with amount of variables .…”