Zhu, Jie Xia Cathy2024-08-152024-08-1520032003https://hdl.handle.net/1828/20281This thesis discusses two-dimensional linear hybrid cellular automata (2D LHCA) and their use as test pattern generators in VLSI testing. It gives the transition matrix of the n-by-m LHCA. As it is desirable for the characteristic polynomial of the transition matrix to be primitive since it generates all possible non-zero test vectors, starting from any nonĀ­ zero state, two ways to speed up the calculation of the characteristic polynomial of 3-by-m LHCA are developed. And a table of minimal-cost n-by-m LHCA with maximum length cycle for small values of n and m is listed followed by some analysis. Transition properties of LHCA are important for detecting delay faults because the latter require a pair of vectors to stimulate. The thesis concentrates on analyzing 2-by-m LHCA. We discover an error in a published theorem, formulate and prove a correct result, allowing for the accurate calculation of the number of substate vectors which produced maximum numbers of the transition pairs for 2-by-m LHCA. Results show that 2-by-m two-dimensional LHCA have a larger set of transitions than one-dimensional LHCA. When applying 2-by-m LHCA as test pattern generators in testing experiments, we demonstrate that 2-by-m LHCA perform slightly better than one-dimensional LHCA for the detection of delay faults in the ISCAS85 benchmark circuits.97 pagesAvailable to the World Wide WebTwo-dimensional linear hybrid cellular automata as test pattern generators in VLSI testingThesis