Modeling and control of a quadcopter. Robust strategies for the detection and handling of equipment failures

Due to their versatility, Unmanned Aerial Vehicles (UAVs) are increasingly used in the modern aviation industry, and many applications include autonomous trajectory tracking. Given the absence of human control during this type of mission, an emergency event, such as a loss of power in the propulsion system or a fault in the sensors, can strongly deteriorate the vehicle’s stability and trajectory tracking capabilities. The goal of this thesis is to develop a fail-safe logic to address critical component faults. To this end, a comprehensive model of an autonomous UAV was developed in MATLAB Simulink. This model is composed of three main parts: (i) a physics-based model of the UAV and its integrated sensors, (ii) a set of Kalman filters for the sensor fusion algorithm, and (iii) a set of PD controllers responsible for stabilization and trajectory tracking tasks. By deactivating or modifying specific parameters of this complete model, we simulate the failure of certain components in the real system and determine their consequences on the vehicle’s behavior during normal operation. Finally, when a failure is estimated to be critical for the flight, we implement a fault detection module that includes: (i) an emergency descent protocol and (ii) an automated parachute deployment system for unrecoverable attitudes to safely bring the vehicle to the ground.