A fault detection and reconfigurable control architecture for unmanned aerial vehicles

The past decade has seen the development of several reconfigurable flight control strategies for unmanned aerial vehicles. Although the majority of the research is dedicated to fixed wing vehicles, simulation results do support the application of reconfigurable flight control to unmanned rotorcraft. This paper develops a fault tolerant control architecture that couples techniques for fault detection and identification with reconfigurable flight control to augment the reliability and autonomy of an unmanned aerial vehicle. The architecture is applicable to fixed and rotary wing aircraft. An adaptive neural network feedback linearization technique is employed to stabilize the vehicle after the detection of a fault. Actual flight test results support the validity of the approach on an unmanned helicopter. The fault tolerant control architecture recovers aircraft performance after the occurrence of four different faults in the flight control system: three swash-plate actuator faults and a collective actuator fault. All of these faults are catastrophic under nominal conditions.