Nonlinear Adaptive Intelligent Control For A Small Unmanned Helicopter

The unmanned helicopter has the advantages of VTOL(vertical take-off and landing),hovering and flying at a low altitude,but it also has the characteristics of strong nonlinearity,static instability,strong uncertainty and so on.These characteristics make the controller design of the unmanned helicopter difficult.This thesis is concerned with the problem of adaptive control for a small-size unmanned helicopter,and puts forward two control schems.The stability of the proposed control algorithm is proved by using Lyapunov stability theorem.Numerical simulation and real-time flight experiments are presented to illsurate the good performance of the proposed control schemes.Firstly,the definition of coordinate system and the definition of the state variables are made clear in this thesis.The main rotor's and the tail rotor's dynamic characteristics are analyzed in detail,and proper simplification of the original nonlinear dynamic model is carried out,which is the basis for the design and verification of the subsequent nonlinear control law.Secondly,this thesis presents a novel asymptotic tracking controller for a smallsize unmanned helicopter using NN(neural network)and a continuous nonlinear robust control strategy called RISE(robust integral of the signum of the error).The NN in the control law approaches the uncertain parts of the helicopter's dynamics using online network weight tuning,while the NN approximation errors and the external disturbances are compensated by the RISE approach.Semi-global asymptotic stability of the error signals and the boundedness of the closed-loop system signals are ensured via Lyapunov based stability analysis.Finally,real-time experiments are performed on a helicopter attitude control test-bed.The experimental results show that the proposed controller achieves a good performance and the closed loop systems obtain good robustness with respect to system uncertainties and external disturbances.Lastly,this thesis is concerned with the problem of adaptive control for a small-size unmanned helicopter with external disturbances and unknown nonlinear terms through the RL(reinforcement learning control)method.The RL method with critic and actor NNs is presented to learn the optimal control for the system uncertainties online,while the NN approximation errors and the external disturbances are compensated by the robust term.Semi-global exponentially stability of the error signals are ensured via Lyapunov based stability analysis.Finally,Real-time experiments are performed on helicopter hardware-in-loop control testbed,and the results validate the good performance of the proposed control scheme.