Design of integral sliding mode observers with application to fault and unknown input reconstruction

This paper presents a new technique for fault diagnosis and estimation of unknown inputs in a class of nonlinear systems. The novelty of this approach is governed by the design of two nonlinear observers which incorporate a combination of sliding and integral control actions. The observers are networked together. The first is used for fault diagnosis and the second is used for the reconstruction of unknown inputs. The two observers exchange their respective reconstructed signals online and in real time. It is shown that under certain conditions, the proposed observer is able to reduce chattering without compromising estimation accuracy. Another significant advantage of the proposed approach is that the network of 2-interconnected integral sliding mode observers permits the relaxation of the fault isolability from the unknown inputs.