Output Back Recursive Unknown Frequency Disturbance Estimation And Compensation

Sinusoidal disturbance is a common phenomenon in actual control systems,which affects the control precision and even leads to the system instability.This paper presents a estimation method for sinusoidal disturbance based on output feedback.An auxiliary filter and a low-pass filter are constructed to excite the disturbance characteristics and estimate the disturbance frequency information.Then,the impact of disturbance to the control system is rejected by feedforward compensation.The input disturbance can be described by a series of dynamic form of virtual disturbance,and the back recursive reduced-order method is presented to estimate the virtual disturbance.Together,the converging characteristic of the input sinusoidal disturbance estimation is provided.The static output feedback controller and observer parameters are adjusted within the disturbance observer framework,and the stability of composite closed-loop system is analyzed to provide the convergence and robustness analysis.The research work of this thesis are listed as follows:A disturbance estimation method based on back recursive disturbance observer is proposed for a class of unknown frequency sinusoidal disturbances.The low-pass filter and the observer are designed to excite the disturbance characteristic and estimate disturbance frequency,with which the virtual disturbance can be reconstructed.The estimation of the input disturbance is realized by the back recursive reduced-order method,which is completely independent of the controller structure.It can directly obtain the dynamic relationship between the input disturbance and the measurable information as long as the frequency information is estimated.The disturbance compensation for non-minimum phase system suffers by unknown frequency sinusoids is analyzed.The relationship between disturbance and measurable information cannot be directly described because of the unstable zero point of the system.Therefore,firstly,the back recursive method is used to reduce the order of the controlled system.Then introduce the linear transformation,and then design the observer to estimate the disturbanceequivalent states and reject the influence of the sinusoidal disturbance signal on the non-minimum phase system,and the stability analysis for the system is provided;A disturbance rejection approach is presented for a class of SISO uncertain second-order systems.The sinusoidal signal can be described by a exogenous system and the un-modeled dynamic represents a nonlinear function that satisfies the Lipschitz condition.According to the separation principle,the disturbance observer and the output feedback controller can be designed separately.The composite closed-loop system is constructed and the controller and disturbance observer parameters are adjusted via linear matrix inequality(LMI)to guarantee the asymptotical stability of the composite system.