Identification of faulty locations in digital circuits using SVM classifier

During the manufacturing process or after manufacturing, a circuit may behave as faulty. These faults can be of any type and can make the circuit to behave abnormally. So, for analyzing and enhancement of the circuit, locating and detecting these faults becomes necessary. These faults are very hard to detect especially soft errors. Here in this paper, by using support vector machine, faults that are present in the nets in the circuits are identified. This method mainly focuses on single event transient faults which are not visible to the human eye. For identification of these faults, radiation scan test is done which consumes more time and a greater number of engineers are also involved. Hence this paper proposes a method which is easy and effective compared to the traditional methods that are used to detect these faults.