Investigation of Extreme Learning Machine-Based Fault Diagnosis to Identify Faulty Components in Analog Circuits

Due to the growing complexities in electronic circuits, it is important to find the faults in a circuit and also diagnose since it is a crucial part during integrated circuit design process. In the whole process, it takes a lot of manual effort to extract and select features. Here we have investigat...

Full description

Saved in:
Bibliographic Details
Published in:Circuits, systems, and signal processing systems, and signal processing, 2024-02, Vol.43 (2), p.711-728
Main Authors: Biswas, Suman, Mahanti, Gautam Kumar, Chattaraj, Nilanjan
Format: Article
Language:English
Subjects:
Accuracy
Algorithms
Analog circuits
Approximation
Artificial neural networks
Back propagation
Bandpass filters
Case studies
Circuit design
Circuits
Circuits and Systems
Classification
Electrical Engineering
Electronic circuits
Electronics and Microelectronics
Engineering
Fault diagnosis
Feature selection
Fourier transforms
Fuzzy logic
High pass filters
Instrumentation
Integrated circuits
Machine learning
Neural networks
Optimization
Principal components analysis
Signal processing
Signal,Image and Speech Processing
Support vector machines
Wavelet transforms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Due to the growing complexities in electronic circuits, it is important to find the faults in a circuit and also diagnose since it is a crucial part during integrated circuit design process. In the whole process, it takes a lot of manual effort to extract and select features. Here we have investigated the scope of the extreme learning machine (ELM)-based fault diagnosis technique in the identification of the faulty component in the analog signal conditioning circuits. The fault diagnosis has been done without feature selection and extraction ELM method. As a case study, we have considered a Sallen–Key bandpass filter and a circuit with four-opamp biquad high-pass filter to investigate the proposed methodology. We have used a single pulse as input and collected the raw data for training and testing purpose. The result from the computation experiment gave 100% and 99.82% average accuracy.
ISSN:0278-081X
1531-5878
DOI:10.1007/s00034-023-02526-9