Study On Novel Low-order Uncertainty And Disturbance Estimators

To address the high precision and high-performance control problem of motion control systems under multiple constraints,it is of great scientific and practical significance to carry out the construction and performance verification research of a robust control method with clear physical concepts,so as to tackle the challenge of various performance trade-off issues,which cannot be dealt with by the existing control frameworks.The uncertainty and disturbance estimator(UDE)proposed by Chinese scholar Prof.Q.C.Zhong in 2004 has the advantages of simple structure,clear concept in frequency domain,and fewer design parameters,etc.The UDE-based robust control is a typical active disturbance rejection control method.In recent years,more and more academic and industrial researchers have paid attention to the analysis,design,comprehensive verification and application of this control method.However,the comprehension and application of UDE design flexibility,performance limits and the potential to deal with more complex problems are not comprehensive enough; the existing UDE-based robust control design methodology framework,still has more shortcomings and limitations.To this end,this dissertation focuses on the improvement of UDE design,performance verification and new application exploration,and carries out the design methodology from two perspectives.First,to address the performance limitations of fixed-parameter UDE,a novel time-varying UDE(TV-UDE)is proposed,which breaks through the timedomain design paradigm based on time-varying filters by introducing mathematical tools such as integration by parts,which broadens the application range of UDEs by introducing the constraints of fixed-parameter filters and enhances the flexibility of the UDE design,and provides a new idea to solving the problem of balanced design of control systems with multiple performance trade-offs.Secondly,to address the demand for improving the estimation performance under the constraint of limited bandwidth,the uncertainty and disturbance estimator with phase-Lead Compensation(UDE-PLC)is proposed,and the optimization technique of filter structure and parameters is developed to reduce the phase lag property,which can help to improve the estimation performance under the constraint of limited bandwidth.Specifically,the main works of this dissertation are as follows:1.For a normal case where the system includes state-dependent model uncertainties,input-dependent model uncertainties,and external disturbance,a lowest-order UDE-based robust control approach under full-state feedback situation is investigated,which consists of a full-state feedback based nominal controller and a first-order classic UDE.The ultimate bounds of estimation error and state tracking error with respect to the design parameters of the monotonic relationship is proved by using the singular perturbation theory.On this basis,the extension design of the novel time-varying UDE is carried out,in which the estimation(filtering)relationship is described by a time-varying differential equation instead of a transfer function,and the computable time-domain expression of the TV-UDE is obtained by using the method of integration by parts.It is pointed out that the classic UDE is a special realization of the TV-UDE.Aiming at the trade-off issue between the transient performance and steady-state performance of high-gain UDE,a control scheme in which the UDE gain transits from a small value to a large one is proposed,and two types of smooth and bounded transition functions are constructed.Based on the simulation and physical experiment results of AERO platform,it is verified that the control scheme can effectively avoid the initial peak and actuator saturation problems of high-gain UDE while realizing high steady-state control.2.A lowest-order UDE based control scheme under the output feedback situation is investigated,and the UDE design scheme with time-varying parameters is proposed.The scheme consists of two parts: the nominal controller and the UDE,in which the nominal controller is designed based on the passivity technique,and the stability and transient performance of the system are ensured by constructing an auxiliary system to inject damping into the original system,which is approximated by the velocity feedback term in PD control; a physically realizable form of the UDE is constructed by designing a special secondorder filtering relationship.The monotonic relationship between the ultimate bounds of the estimation error and trajectory tracking error and the design parameter is proven by the singular perturbation theory.On this basis,a design scheme of UDE with time-varying parameters is proposed.Simulation and experimental results verify that the second-order time-varying UDE is able to address the transient performance issue of the high-gain UDE while guaranteeing the steady-state performance of the estimator.3.To address the practical problem that the allowable UDE bandwidth of a real system is constraint,the phase lag characteristic of the embedded first-order filter under the constraint of finite bandwidth are analyzed,and a UDE with a phase-overcompensation(UDE-PLC)is proposed,whose main idea behind the design is to use the cascade of a first-order phase lead compensator and a first-order Butterworth filter to construct a new second-order estimation relationship.By choosing properly the involved parameters,the obvious phase lag of the non-ideal Butterworth filter is well compensated and both the disturbance estimation error and trajectory tracking error are reduced by the proposed design.The design specifications and the guideline for parameter tuning are provided to make the philosophy behind the UDE-PLC easy to follow.The effectiveness of UDE-PLC is verified by simulations and physical experiments on a 2-DOF AERO attitude control platform,which show that the estimation performance of the UDE-PLC is significantly improved compared with that of the classic first-order UDE,and the UDE-PLC based robust control achieves a higher steady-state tracking accuracy of the attitude angle.4.Aiming at a typical experimental platform such as a three-degree-of-freedom helicopter with under-actuated,nonlinear and obviously inaccurate design model,a UDEbased active disturbance rejection control attitude control scheme is investigated,and a robust control strategy based on inner and outer loops and the lowest-order UDE is proposed.The scheme firstly introduces a virtual controller based on the high-order fully actuated system approach,which is combined with the feedback linearization technique to realize the transformation from an under-actuated system to a full-actuated system.Considering the dynamic property of the model,a control scheme based on inner and outer loops,including a state feedback-based nominal controller and a first-order UDE,is adopted.The potential time-scale property of the whole system under different disturbance frequencies is analyzed.The necessity of disturbance compensation,the effectiveness of UDE,and the convenience of control parameters adjustment are verified by both simulation and experimental results.5.Aiming at the tracking control problem of a class of multi vehicle systems(threedegree-of-freedom helicopter platform)suffering from internal disturbance of the leader and external disturbance of the follower,a distributed robust control scheme based on the integrated compensation design of the follower is proposed.The control scheme includes a nominal controller for stabilizing the system,an ESO for compensating the external uncertainties,and a UDE for compensating the internal uncertainties.The performance of the whole closed-loop system is analyzed.It is proved that the ultimate bounds of tracking errors and closed-loop estimation errors depend monotonically on two key design parameters of the estimators.The effectiveness of the parameter-tuning mechanism and the necessity of compensating both the internal and external disturbances for followers are demonstrated by a simulation comparison on a system including four virtual helicopter models and by an experimental comparison on a system including both virtual and real helicopter models.