Actuator fault detection in uavs

Abstract

Future unmanned aerial vehicles (UAVs) will be designed to achieve their missions with increased efficiency, safety, and security. To this end, an efficient fault detection and isolation (FDI) system should be capable of monitoring the health status of the aircraft. Fault-tolerant control systems for small and low-cost UAVs should not increase significantly the number of actuators or sensors needed to achieve the safer operation. This chapter is dedicated to actuator fault detection systems for UAVs, with two main requirements: realtime capability and modularity. After defining the terminology employed in this field, this chapter reviews some commonly used techniques in FDI systems. The chapter continues by presenting briefly the mathematical model of a UAV which will serve as a basis for the design of two actuator FDI systems. The first method presents and illustrates the multiple-model approach, whereas the second method presents an FDI system which is based on a single model. Both methods have been enhanced by a mechanism that actively tests actuators in order to efficiently detect and isolate actuator faults and failures. This chapter explains the advantages and drawbacks of each method and discusses issues of robustness against model uncertainties and external perturbation. In addition, aspects of computational load are addressed. Finally, the FDI systems of this chapter are applied to a realistic model of an unmanned aircraft, and the performance of the methods is shown in simulation.