Data-Driven Technique-Based Fault-Tolerant Control for Pitch and Yaw Motion in Unmanned Helicopters

This article proposes a fault-tolerant control (FTC) algorithm to monitor the anomalies and control them for unmanned helicopters. This overcomes the drawbacks due to heteroscedasticity and restrictions on the input variables in conventional FTCs and improves the training and testing efficiencies with enhanced control. This is achieved by adopting the support vector data descriptor (SVDD) to learn the operating states of the unmanned helicopter during normal and motor fault operations. Furthermore, a neural integrated fuzzy (NiF) controller is trained to cope up with the states classified by the developed classifier. To realize the development of the proposed control algorithm, the motor faults in the unmanned helicopter are developed by observing the pitch and yaw motor operation of unmanned helicopters. A two-class classification is developed using SVDD for identifying the motor faults and the NiF is trained for controlling the plant during the anomaly. The results depicted 98.6% training accuracy and 98.96% prediction accuracy along with efficient control when tested for a faulty condition on a helicopter test rig.