In this paper, we report on the development of an intelligent system for air quality monitoring and early detection and diagnosis of air contaminants. Optimal identification of contaminants is based upon the use of an Implicit Kalman Filter that uses both experimental data and a theoretical model to obtain optimal estimates. We have developed a three-dimensional unsteady-state model of contaminant transport, which uses a flow field generated numerically for the cabin using a finite element mesh. The optimal contaminant estimates are used as the basis for the detection of a contamination event. The algorithm is shown to distinguish between sensor faults and process faults.