Embedding FDI in launcher attitude controllers

Abstract

dynamics of the weightings (if any) can be accommodated in the YOULA parameter. Furthermore, the closed-loop poles partition of the resulting control system should be chosen with care, since the deterministic separation principle relies on reduced sets of state-feedback poles, state-estimator poles, and remaining YOULA parameter poles (or a static one), as it will briefly reviewed in this work.Observer-based realizations can supply signals to be used in the detection and isolation of sensor or actuator faults and failures (bias), and to estimate external disturbances as well. Indeed, for a given controller, several observer-based realizations involving different on-board models (each of them taking into account a particular condition) can be devised; for each on-board model, one have to choose the best closed-loop eigenvalue distribution to satisfy given indexes on maximum estimation error and noise levels. One intends to cover the following subjects :• The section II presents briefly the new techniques for determining the observer-based realization of any controller with arbitrary order. • The section III presents the decoupled full pitch plane launcher model used in this study, the general H oo standard problem adopted for the attitude control and the design procedure combining H oo control and computational intelligence (CI). • In the section IV, the CI-designed H 00 controller and an on-board model are used to redesign the original controller as an observer, providing estimates not only of the plant states but also of the angle of attack and plant output bias, when noise and external disturbance are simultaneously acting on the system. • In the section V, simulation results are supplied to validate the overall approach. • The last section states the main conclusions and the next steps toward non-linear digital and hardware-in-the-Ioop simulations.