Fault Tolerant Control of a Variable Pitch Quadrotor

In this paper, we solve the fault tolerant tracking problem for a variable pitch quadrotor. Following the Disturbance Observer-Based Control design paradigm, we face the observation problem for the blade pitch, and consequently the related fault and failure diagnosis problems. The control allocation algorithm solves the optimal redistribution problem of the control effort among the propellers in case of actuation failure. The optimization problem takes into account the energy consumption, the health condition of the actuators, and the presence of input constraints, such as saturation and rate of change limits. In particular, we face the specific problem of lock-in-place servo failures, which reduce the number of control actions in the system. Indeed, together with a fault detection and isolation module, lock-in-place servo failures can be managed in the control allocation algorithm while keeping the tracking capabilities. The proposed optimal fault tolerant accommodation algorithm can be coupled with most of the nonlinear control laws commonly applied to conventional multirotor systems. Numerical simulations with noise and constraints show the capability of the scheme to handle this class of failures.