Robust quadrotor actuator fault detection and isolation in presence of environmental disturbances

This paper examines the problem of designing a module for the detection and isolation of faults affecting quadrotors' actuation subsystem. Furthermore, differently from other diagnosis scheme available in literature, in this paper (not necessary small) environmental disturbances are also taken into account. In particular, this paper proposes a new two-steps strategy, based on the NonLinear Geometric Approach, for fault detection and isolation: in the first step, the application of the NonLinear Geometric Approach leads to coordinate changes able to highlight new systems affected by the wind and by a part of all possible faults; the second step exploits wind estimators designed on the equations of the new systems found at the end of the step one. These estimators are combined to generate a residuals set structured for fault detection and isolation. Simulation results demonstrate and validate the performance and the capabilities of the proposed fault detection and isolation algorithm.