Disturbance Compensation And Adaptive Control For Nonlinear Servo System

The high performance servo mechanisms have been widely applied in industry applications.The friction dynamics,parameter uncertainties and external disturbance exist in servo system,which can affect the control performance of the servo system,and even lead to the instability of the servo system.Consequently,the controller design for the servo system with friction,parameter uncertainties,external disturbances and unknown parameters has great many attentions in the control field,which is also of significant theoretical and practical values.This thesis discusses the disturbance compensation and adaptive control for the servo system with the friction,parameter uncertainties and external disturbance.The adaptive funnel control,adaptive neural dynamics surface control,robust adaptive control and adaptive control based on the parameter estimation error are all proposed.The main contributions of this thesis are listed as follows:(1)An adaptive funnel controller based on the extended state observer(ESO)is developed for the nonlinear servo mechanisms with friction,unknown nonlinear dynamics and external disturbances.To improve the transient and steady-state performance of the servo system,a modified funnel function is proposed by releasing the imposed assumption in conventional funnel controls(e.g.,systems with relative degree one or two)and avoids the potential singularity problem in prescribed performance control.Moreover,the friction,parameter uncertainties and external disturbance is lumped as total disturbance,an ESO is designed based on the system bandwidth to estimate the system states and total disturbance.Then,an adaptive funnel controller is proposed for the servo system,and the parameter tuning rules of the adaptive controller is given.Comparative simulations and experimental results are conducted based on a practical turntable servomechanisms show that the proposed control scheme can guarantee the tracking error within given boundary,and improve the transient and steady-state performance of the servo mechanism.(2)An adaptive neural dynamic surface control method with error constraint is proposed for the two-mass system with mechanical vibration and friction.A modified prescribed performance function(PPF)is proposed and incorporated into control design to improve the dynamic performance and steady-state performance of the control system.The modified PPF avoids the singularity exists in the original PPF.Moreover,a novel adaptive neural dynamic surface controller(DSC)is designed by using a high-gain tracking differentiator(HGTD)replaces the traditional first-order filter.Moreover,the friction,parameter uncertainties and external disturbance are lumped as total disturbance,which is compensated by using echo state neural networks(ESNs).ESNs has the function approximation capability of Recurrent Neural Networks(RNNs),but requires simpler training than RNNs.Compared with RNNs,the ESNs can easily be trained without adjusting the weights between the input layer and the hidden layer,and the connection weights of the reservoir network are not altered during the training phase.The adaptive neural dynamic surface controller is designed based on the recursive feedback technique to damp the torsional vibration.The effectiveness of proposed control scheme is verified by simulation and experiment results.(3)The robust adaptive asymptotic tracking controller is proposed based on the robust integral of the sign of the error(RISE)for the servo system with friction,parameter uncertainties and external disturbances.A continuously differentiable friction model is adopted to account for the friction nonlinearities.A high-order neural network(HONN)is used to accommodate the unknown nonlinearities.An improved prescribed performance function is developed to guarantee the transient and steady-state performance of the tracking error.The residual approximation error and other bounded disturbances are compensated by using a robust integral of sign of the error term.The stability of the servo system is proved by using Lyapunov function.Comparative experiments based on a practical turntable servo mechanism are conducted to validate the effectiveness of the proposed control scheme.(4)The robust adaptive finite-time control design based on estimation error is proposed for nonlinear servo system with unknown parameters.An auxiliary filter is firstly used to derive the information of parameter estimation error,which is then used as a new leakage in adaptive law.Compared with other estimation methods,this estimation algorithm does not calculate the derivative of the system states,and can achieve true values.Moreover,a newly developed continuously differentiable friction model is adopted to address the friction nonlinearities.This model is incorporated into an augmented neural network to account for the unknown nonlinearities.Both the weights and primary friction coefficients are updated via the suggested adaptive law.The proposed parameter update law are further improved via a sliding mode technique to achieve the finite-time convergence.Hence,the exponential convergence of the control error and estimation error is achieved simultaneously.Finally,the effectiveness of the proposed methods are validated by simulations and practical experiments based on a turntable servo system.