Masking Internal Node Faults and Trojan Circuits in Logical Circuits

A combinational circuit C is considered. Masking of internal node logical faults with using the subcircuit that outputs are connected with circuit C internal nodes that are fed by fault nodes is suggested. The sub-circuit inputs are connected with either circuit C inputs or with internal nodes of circuit C that precede the fault nodes. Masking is based on applying of incompletely specified Boolean functions of internal nodes. Algorithms of deriving incompletely specified Boolean function for some internal node v are described. One of them gets the incompletely specified function that depends on input variables of circuit C, another gets the incompletely specified Boolean function that depends on internal variables of circuit C that corresponds to internal nodes preceding fault nodes. Using these algorithms for several fault nodes we obtain the system of incompletely specified Boolean functions that is implemented by masking (patch) circuit. This approach may be also applied for masking Trojan Circuits (TCs). It is supposed that TC output is injected into a line of combinational circuit C. Experimental results are given. They demonstrate possibilities of essential cutting overhead when using patch function in comparison with duplication.