State estimation and fault reconstruction from an adaptive descriptor sliding mode observer

An adaptive descriptor sliding mode observer (SMO) is proposed to estimate system states and reconstruct actuator fault for a class of uncertain system with both actuator fault and output disturbance. The system model is firstly transformed and augmented to be a descriptor system model. In this way, the actuator fault and output disturbance are decoupled, and the output disturbance turns into new states of the transformed system. Secondly, inspired by the idea of unknown input observer and sliding mode observer, an adaptive descriptor SMO is developed where the strict constraint in the dimension of outputs is relaxed. Meanwhile, with the help of adaptive technique, the requirements of known boundary in disturbance and fault are liberalized. At last, actuator fault is reconstructed based on the idea of equivalent output error injection. Case studies of a reusable launch vehicle during re-entry flights are conducted to demonstrate the effectiveness of the proposed approach.