Fault Diagnosis and Location of Quadrotor UAV Actuator based on ESO-DF

The reliability of the quadrotor UAV is the basic guarantee for its mission execution. Under the condition of frequent maneuvering, the actuators of quadrotor UAV are prone to fault. For the fault diagnosis of the quadrotor UAV, the existing schemes either only consider the diagnostic information of the fault type, or only consider the information of the fault location. In this paper, a data-driven fault diagnosis and location scheme based on the combination of Extended State Observer (ESO) and Deep Forest (DF) algorithm is proposed to solve the problem of actuator fault diagnosis and location that has received less attention. First, under maneuvering conditions, the state of the quadrotor UAV can be accurately extracted by ESO, thereby providing more credible fault information. Then combined with the structure of the quadrotor UAV, a fault data generation method including the actuator fault location information is designed. Finally, the fault data is injected into the DF model for fault diagnosis, and the information of the actuator fault location and fault type of the quadrotor UAV can be obtained at the same time. This scheme has the advantages of simple modeling, few hyperparameters, adaptability to small sample datasets, and high accuracy. The simulation experiment shows that the accuracy of the scheme for fault diagnosis and location of the actuator of the quadrotor UAV is over 99.9% under maneuvering conditions.