Detection of discrete faults in electronics assembly

We present an optimal estimation approach to detecting discrete faults in electronics assembly. The algorithm utilizes a nonlinear fault model whose state is a vector containing the probabilities of the process in the normal operating state and all possible fault states at a given time. The model describes the time behaviour of the process state and how the process state is related to the process measurements. An extended Kalman filter is used to obtain an estimate of the process state. This approach is applied to an electronics assembly line located at the Center for Board Assembly Research at Georgia Tech.