The efficient and structured analysis of unknown CMOS integrated circuits (ICs) has become a topic of great relevance in recent years. Up until now, different invasive [1], [2] and non-invasive [3], [4] strategies have been developed for procedure of analysis. However, invasive procedures always lead to the destruction of system under investigation. The non-invasive approaches published so far have the disadvantage that ICs are analysed by using complex algorithms. Here, no subdivision exists to avoid extensive analysis times in the case that only simple structures are investigated. Moreover, traditional procedures cannot automatically distinguish between input and output pin types, which is usually required in the investigation of real unknown integrated circuits. This paper presents an efficient non-invasive procedure to determine binary multi-input multi-output (MIMO) ICs by its input-output behaviour. It was implemented into analysis environment described in [5] and classifies unknown ICs by means of automata theory. A novel separation procedure is proposed in this paper to further minimise the IC analysis. All sections of the classification procedure are simulated and fully tested on ISCAS-85, ISCAS-89 and ISCAS-99 benchmark models of real ICs [6], [7] and the results are presented in this paper.
[1]
A. Th. Schwarzbacher,et al.
Determination of Pin Types and Minimisation of Test Vectors in Unknown CMOS Integrated Circuits
,
2006
.
[2]
Niraj K. Jha,et al.
Switching and Finite Automata Theory
,
2010
.
[3]
John P. Hayes,et al.
Unveiling the ISCAS-85 Benchmarks: A Case Study in Reverse Engineering
,
1999,
IEEE Des. Test Comput..
[4]
Y. Kuroe.
Learning and identifying finite state automata with recurrent high-order neural networks
,
2004,
SICE 2004 Annual Conference.
[5]
David Lee,et al.
Principles and methods of testing finite state machines-a survey
,
1996,
Proc. IEEE.
[6]
Beatrice Fraboni,et al.
Layout reconstruction of complex silicon chips
,
1993
.