A Robust and Self-Scheduled Longitudinal Flight Control System: a Multi-Model and Structured H-infinity Approach

Hugo Lhachemi (1) , David Saussie (1) , Guchuan Zhu (1)

1 EPM - École Polytechnique de Montréal

This paper presents a novel approach to design robust and smooth self-scheduled controllers in the framework of structured H-infinity control. The newly available MATLAB-based tool, hinfstruct, is used to tune fixed-structure linear controllers to meet H-infinity constraints. Moreover, multi-model and multi-channel capabilities are exploited in the design of self-scheduled controllers. In this context, we define a priori both controller architecture and gain-scheduling structure. This approach is successfully applied to the longitudinal control of a F-16 Fighting Falcon over the ight envelope under mass and center of gravity uncertainties. After casting the control synthesis into a structured H-infinity problem, the scheduled gains are directly obtained as quadratic polynomial functions of altitude and Mach number. The performance of the designed control systems is validated by numerical simulations.

Domaines: Systèmes et contrôle [cs.SY]

Hugo Lhachemi, David Saussie, Guchuan Zhu. A Robust and Self-Scheduled Longitudinal Flight Control System: a Multi-Model and Structured H-infinity Approach. AIAA Scitech Guidance, Navigation and Control Conference, 2014, National Harbor, United States. pp.1377-1383, ⟨10.2514/6.2014-0601⟩. ⟨hal-03133679⟩