Adaptive Sliding Mode Control Of Nonlinear Servo System Based On Parameter Identification

At present,the performance requirements of nonlinear servo system control has gradually changed from stability and reliability to high steady-state accuracy,rapidity,strong robustness and anti-disturbance performance.Among many robust control methods,sliding mode control is widely studied because of its strong robustness to uncertainties such as system modeling errors,parameter perturbations and external disturbances,etc.However,sliding mode control depends on the accurate system model,and there is a certain chattering problem,which limits its application in practical engineering.Adaptive parameter identification can quickly feedback the parameter perturbation caused by external changes,and the deterministic equivalence principle shows that if the adaptive parameters identification converges to the true value quickly and accurately,the tracking control performance can be effectively improved.Therefore,the adaptive parameter identification and sliding mode control theory are studied in this paper to improve the accuracy of parameter identification and steady-state tracking performance of the system,and a nonlinear AC servo motor control platform is built for experimental verification.Therefore,the research works of this paper have important academic value and application prospects.The main works and achievements of this paper are as follows:1.Based on the traditional linear sliding mode control,a glowworm swarm optimization(GSO)based adaptive nonlinear sliding mode control method is proposed and applied to the electromechanical servo system with nonlinear dynamic friction.Firstly,the LuGre friction model of electromechanical servo system is established,and the friction parameters are divided into static parameters and dynamic parameters according to high-speed steady-state and low-speed.Because GSO has the advantages of high speed,efficiency and versatility in searching extreme range,the static and dynamic parameters of friction are identified offline by employing GSO algorithm to obtain more accurate model parameters.Then,to slove the problem that the dynamic parameters of friction are susceptible to the influence of external environment and disturbance,an adaptive finite-time parameters identification law is constructed by designing the dynamic equation of state predictor and prediction error,which can reflect the real-time changes of the dynamic parameters of friction online,and solve the problem that the traditional adaptive parameter identification is difficult to converge to the real parameter values due to the control error.In addition,an adaptive nonlinear sliding mode controller is designed based on the identification results by constructing a nonlinear exponential function,which can effectively improve the steady-state accuracy and error convergence speed.2.In view of the system unknown states and uncertainties,a finite-time full-order sliding mode control method based on extended state observer is proposed to effectively suppress the chattering problem of sliding mode control,and the effectiveness of the method is verified on a flexible joint manipulator system.Firstly,an extended state observer based on adaptive parameter tuning is designed to estimate the unknown system states and uncertainties.Since the parameters of the extended state observer are difficult to be obtained,the observer parameters can effectively approach the ideal values by constructing an adaptive parameter learning law,which can improve the accuracy of the state observation and avoid the peak phenomenon caused by high gain.Then,a finite-time full-order sliding mode control strategy with the same order as the system is proposed.The controller does not contain switching function term that can cause system chattering by designing a first-order low-pass filter,thus the chattering problem of sliding mode control can be effectively weakened without the need for all state measurability of the system.3.Based on the analysis of the generation mechanism and the suppression techniques of sliding mode chattering,a novel continuous double hyperbolic reaching law in investigated.When the sliding mode variable reaches a neighborhood of the equilibrium point,the change rate of the sliding mode variable is less than the current value in a unit sampling time.Therefore,the sliding mode variable converges in the way of infinite approaching the equilibrium point without crossing the equilibrium point,thus the chattering-free property and fast convergence property are achieved.Besides,in view of the unknown parameters in electromechanical servo system,an adaptive parameter identification law is constructed by filtering the known regression matrix and designing virtual dynamic variables,so that the parameters can converge to the true value effectively.Then,a nonlinear disturbance observer is designed to effectively observe the uncertainties including disturbances and parameter identification errors in the system,which can improve the steady-state control accuracy.4.On the basis of above researches,to solve the influence of extra variables including disturbance,disturbance observation error and parameter error on the chattering-free characteristics of double hyperbolic reaching law,an improved double hyperbolic reaching law is proposed.By filtering the sliding mode variable and the original double hyperbolic reaching law,the compensation term of the changing rate caused by the extra variable is constructed to improve the convergence performance of the reaching law.In addtion,considering that the electric-driven manipulator is a multi-input multi-output,highly nonlinear and strongly coupled complex system,which contains many unknown parameters and system uncertainties,an adaptive parameter identification method is designed by introducing error information about the parameters themselves into the adaptive law,which realizes the online parameter identification and converges effectively to the true parameter values.Besides,a high-order disturbance observer is designed to reduce the influence of parameter identification errors and disturbances on system performance,and then,the observation errors are compensated by the improved double hyperbolic reaching law to achieve better tracking control performance.5.A nonlinear AC servo motor control platform is built,and the proposed double hyperbolic reaching law sliding mode control algorithm is verified and compared with other reaching law sliding mode control algorithms and PD control by the experiments.In addition,an adaptive parameter identification method is designed to estimate the upper bounds of platform disturbances including friction and load and son on,and to verify the effectiveness and superiority of the proposed method.The adaptive parameter identification method studied in this paper can make the system unknown parameters quickly converge to the true values of the parameters,and based on the deterministic equivalence principle,the method can effectively improve the tracking control performance of the system.In addition,by studying the problem of suppressing sliding mode chattering,the proposed hyperbolic reaching law and improved double hyperbolic reaching law can effectively weaken the chattering and large control gain of sliding mode control,which is conducive to its practical engineering application.