Discussion on: “A Fault Detection Filter Including an Adaptive Noise Cancellation Strategy”

Abstract

The paper by J. Y. Keller presents a fault detection and isolation (FDI) system that is capable of detecting and isolating a single deterministic actuator fault, the effect of which is modeled as an input vector (with constant direction) to the dynamics of the plant. The method is illustrated with a linear model of the longitudinal motion of an aircraft. There are several interesting aspects in the method presented. First, this FDI system does not use the traditional approach of employing a bank of separate Kalman filters each designed for a separate fault scenario [1–3]. Rather, the system is composed of a unique Kalman filter. This has the significant advantage of reducing the complexity of the filter implementation and the computational load. Second, contrary to standard fault detection filters that rely on the direct sequence of output residuals