AN APPLICATION OF TCRBF NEURAL NETWORK IN MULTI-NODE FAULT DIAGNOSIS METHOD

Zbigniew Czaja, Michal Kowalewski

Abstract

This paper presents the new self-testing method for diagnosis of analog parts in mixed-signal embedded systems controlled by microcontrollers. The tested analog part is stimulated by a sinus-wave supplied by the onboard generator and its responses are sampled in selected nodes by microcontrollers ADC. The measurement space is represented by differences between values of selected node voltages. Fault detection and localization is performed by a Two-Center Radial Basis Function (TCRBF) Neural Network. The diagnosis procedure was implemented and simulated in a PC.

Keywords

fault diagnosis, neural network, BIST

Download

IMEKO-WC-2009-TC4-340.pdf

DOI

-

IMEKO TC

TC4 - Measurement of Electrical Quantities

Back to the proceedings