Adaptive Fuzzy Neural Attitude Tracking Control Of A Quadrotor

For attitude tracking control of a quadrotor with the nonlinear of system uncertainties,external disturbances and input saturation,Applying the backstepping control technique,adaptive control technique,fuzzy system and neural network,this paper has proposed several new attitude tracking methods.The main contents of this thesis are as follows:Considering attitude tracking control scheme of a quadrotor,taking the problem of input saturation,model uncertainties and external disturbances into account,A novel scheme which is based on single hidden layer neural network compensation and approximation error feedback is proposed.Then,the online estimation of the unknown nonlinear dynamics and external disturbances and remarkable control performance and stronger robustness is maintained.The closed-loop system has been proved to be stable and tracking errors ultimately uniformly bounded through the Lyapunov stability analysis.Numerical simulations are performed to demonstrate the superiority and excellent performance of this method.Concerned with the problem of external disturbances and model uncertainties involved in the attitude scheme of a quadrotor,an adaptive fuzzy controller with variable contration-expansion factors is proposed.By virtue of the contraction-expansion factors updating automatically,the dynamically fuzzy partition of input space of fuzzy systems is maintained,which improved the accuracy of fuzzy approximation without increasing the number of fuzzy rules.Enhanced robustness and stronger ability to resist external disturbances is obtained.The closed-loop system has been proved to be asymptotically stable and all system states uniformly bounded through the Lyapunov stability analysis.Simulation results demonstrate the superiority and effectiveness of this method.In traditional fuzzy systems,the number of fuzzy rules is difficult to determin.In order to simplify the fuzzy neural system,a self-constructing adaptive robust fuzzy neural attitude tracking control of quadrotor scheme with uncertainties and unknown disturbances is proposed.Then,the adaptive number of fuzzy rules is realized.Besides,A robust controller is designed to eliminate the negative effect of fuzzy approximation error and external disturbances on closed loop system.Through the Lyapunov stability analysis,the closed-loop system has been proved to be global ultimately uniformly bounded and all system states global ultimately uniformly bounded as well.Simulation studies demonstrate the remarkable performance and superiority of this control scheme in terms of attitude tracking and nonlinear approximation.