Reconfigurable Fault-Tolerant NMPC for Y6 Coaxial Tricopter with Complete Loss of One Rotor

Abstract

Safe and reliable operation of unmanned aerial vehicles (UAVs) is a high priority as most of the UAVs must be operated in humanly populated environments. In this study, we exhibit a nonlinear model predictive control (NMPC) approach, as an active fault-tolerant control method, for a Y6 coaxial tricopter UAV in the presence of an actuator fault respecting the actuator constraints. A cascaded closed-loop control methodology, incorporating reconfigurable low-level controller, is proposed which eradicates the need to design multiple controllers for normal and faulty operations. The efficacy of the presented methodology is illustrated for tracking an 8-shaped trajectory with a complete loss of one rotor. From the sequential failure cases of two different rotors, it is shown that even with a fault-detection delay of 0:5s, the maximum altitude divergence is 0:59m that smoothly diminishes over the span of few seconds. In addition to the adequate tracking performance, thanks to ACADO toolkit, a combined average execution time of 2:1 milliseconds is achieved which is promising for its real-time applicability for fast aerospace applications.