Optimisation and implementation of a nonlinear identification procedure for unknown Ics

Abstract

The analysis of unknown integrated circuits (ICs) has attracted significant interest over the last years. In particular, the problem of non-invasive strategies were discussed to avoid the destruction of the device under test. Here, the focus is on powerful analysis tools to identify today's complex ICs. Despite the fact that these algorithms are able to analyse very complex unknown ICs in a short time, their efficiency can be further improved which will be demonstrated in this paper. Here, one research field is the optimisation of the overall nonlinear analysis procedure. This paper presents the optimisation and implementation of the non-invasive analysis procedure. The overall identification time of the nonlinear algorithm can be improved by up to 90% compared to the traditional procedure. The identification procedure optimised was simulated and fully tested on IEEE ISCAS benchmark models as well as user defined models of real ICs to validate the results. For every circuit analysed the function has been successfully determined by the proposed optimised identification procedure.