Abstract

With the increasing demand for unmanned aerial vehicles (UAVs) in both military and civilian applications, critical safety issues need to be specially considered in order to make better and wider use of them. UAVs are usually employed to work in hazardous and complex environments, which may seriously threaten the safety and reliability of UAVs. Therefore, the safety and reliability of UAVs are becoming imperative for development of advanced intelligent control systems. The key challenge now is the lack of fully autonomous and reliable control techniques in face of different operation conditions and sophisticated environments. Further development of unmanned aerial vehicle (UAV) control systems is required to be reliable in the presence of system component faults and to be insensitive to model uncertainties and external environmental disturbances.

This thesis research aims to design and develop novel control schemes for UAVs with consideration of all the factors that may threaten their safety and reliability. A novel adaptive sliding mode control (SMC) strategy is proposed to accommodate model uncertainties and actuator faults for an unmanned quadrotor helicopter. Compared with the existing adaptive SMC strategies in the literature, the proposed adaptive scheme can tolerate larger actuator faults without stimulating control chattering due to the use of adaptation parameters in both continuous and discontinuous control parts. Furthermore, a fuzzy logic-based boundary layer and a nonlinear disturbance observer are synthesized to further improve the capability of the designed control scheme for tolerating model uncertainties, actuator faults, and unknown external disturbances while preventing overestimation of the adaptive control parameters and suppressing the control chattering effect. Then, a cost-effective fault estimation scheme with a parallel bank of recurrent neural networks (RNNs) is proposed to accurately estimate actuator fault magnitude and an active fault-tolerant control (FTC) framework is established for a closed-loop quadrotor helicopter system. Finally, a reconfigurable control allocation approach is combined with adaptive SMC to achieve the capability of tolerating complete actuator failures with application to a modified octorotor helicopter. The significance of this proposed control scheme is that the stability of the closed-loop system is theoretically guaranteed in the presence of both single and simultaneous actuator faults.