Robust Fault Detection and Isolation applied to Indoor Localization

In this paper, a passive method for robust fault detection and isolation is elaborated in the context of set-membership approach. The basic idea behind the proposed method is to compute a feasible set containing the states of a nonlinear system states using an interval constraints propagation method. A consistency test is performed using the q-relaxed intersection technique which consists in tolerating a given number q of outliers from m measurements. The robust feasible set, compatible with m-q observations, is then obtained and faulty sets are used for the isolation procedure. An application, within the framework of robot/human localization, is given to verify the efficiency of the method.