Modified state observer based adaptation of a general aviation aircraft - simulation and flight test

Abstract

This paper presents some issues encountered while flight testing a Model Reference Adaptive dynamic inverse controller (MRAC), on a General Aviation (GA) fly-by-wire test bed. Adaptation was replaced with a Modified State Observer (MSO) method which shows promising results in simulation. The flight test results of the MRAC summarized in this paper show PLA surging issues due to adaptation during the commanded flight path angle and airspeed. To overcome this problem, MSO based adaptation methodology is adopted to control the longitudinal dynamics of a typical general aviation aircraft. The advantage of MSO is that it adapts to estimation error, not modeling or tracking error. The controller performance is evaluated for elevator and engine dynamics along with the thrust saturation limits. The elevator model considered in this study deals with a very large delay (200 msec) in the elevator servo. The thrust saturation depends on engine shaft horse power limits that simulate the flight test condition. A controlled flight simulation is carried out including the turbulence effects observed during flight. Controller design using the saturated thrust is used in flight test conditions, where the thrust generated by the engine is not an observable state. The simulation results showing the controller adaptability to flight path angle and airspeed commands are promising. The random PLA surge seen during the simulation and flight test of the baseline MRAC controller is not observed in the simulation results of MSO adaptation controller. Hardware-In-Loop (HIL) Ground Tests and Flight test of MSO based controller is scheduled in the near future.