Reconfigurable flight control via multiple model adaptive control methods

Abstract
A multiple model adaptive control (MMAC) algorithm with seven elemental command generator tracker proportional-plus-integral Kalman filter controllers (designed for an aircraft functioning correctly, for one of three failed sensors, or for one of three failed actuators) demonstrates consistently effective reconfiguration capabilities when subjected to hard and soft failures of sensors or actuators. Performance can be enhanced considerably by use of maximum a posteriori versus Bayesian form of the MMAC (or a modified combination of both), alternate commutation of the MMAC hypothesis probabilities, and reduction of identification ambiguities through scalar residual monitoring. This latter additional voting technique provides unequivocal declarations that a particular sensor has or has not failed, and it may be used to supplant the MMAC methodology for sensor failures.

This publication has 6 references indexed in Scilit: