Iterative logic arrays (ILAs) are combinational modules connected in geometrically regular patterns. Fault detection of one-dimensional ILAs has been extensively studied. The inputs to and outputs from ILAs are the only observations available to designers. Thus the problem is to select the proper input vector so that the corresponding output vector gives the designer ideas about the malfunctioning cell. Testing an ILA completely for all detectable faults will involve an impractically large number of input vectors.
The author has tried to devise methods to test two- or higher-dimensional ILAs, by considering many common functions, most of which are arithmetic. He demonstrates the single functional fault testing of one-dimensional ILAs. Then, based on the idea of the inverse of the output function (in order to propagate error in a cell to the boundary of the ILA), he shows how to derive tests for fault detection of a single cell using graph procedures.
The paper addresses an important area in fault detection. It is well written and demonstrates the result with examples. Although it appears that the paper may be of limited scope, it is sure to open up a new line of inquiry.