Development of a reconfigurable flight control law for tailless aircraft

A neural-network-based direct adaptive control approach to the problem of r econfigurable flight control is described. The control law was first tested using a model of a tailless fighter aircraft configured with multiple and redundant control actuation devices and subsequently tested both in piloted simulation and in a flight test on the X-36 aircraft. The model aircraft w as used for design and to illustrate the level to which handling qualities can be maintained in the presence of failures in the actuation channels. Of significance is the speed with which recovery and maintenance of handling qualities can take place. The main advantage lies in eliminating the need f or parameter identification during the recovery phase and limiting the pote ntial need for parameter identification in the problem of control reallocat ion following a failure. A second by-product is that the need for an accura te aerodynamic database for the purpose of flight control design can be sig nificantly reduced. Moreover, the need for extensive offline analysis, in-f light tuning and validation of gain schedules, and contingency coding neces sary to handle a large set of possible failure modes is substantially reduc ed. Thus, the overall approach may also be viewed as a direct path to subst antially reducing the cost associated with the development of new aircraft.