On Adaptive Output Feedback Control For Stochastic Nonlinear Systems Via Command Filter

With the increasing complexity of industrial processes,there are many uncertain factors in the system,such as unmodeled dynamics,input quantization,input dead-zone,actuator failure and stochastic disturbance.Due to the improvement of the quality of life,people's requirements for the performance of the control system are also increasing day by day.The traditional control method only considers the stability of the system,which can not meet the actual needs,while the prescribed performance control considers both the stability and transient performance of the system.In recent years,adaptive control of stochastic output feedback nonlinear systems with various uncertainties,such as unmodeled dynamics,input quantization,input dead-zone and actuator failure,has become one of the hotspots of control theory research.This research has important theoretical significance and potential application value.For stochastic nonlinear systems with unmodeled dynamics,input quantization,input dead-zone,actuator failure and prescribed performance,this paper uses high-gain observer and K-filter to estimate the unmeasured state of the system,organically combines command filtering technology with traditional dynamic surface control(DSC)method and introduces compensation signal,and proposes three adaptive output feedback DSC strategies.The main contents are described as follows:(1)A new adaptive output feedback DSC scheme based on command filter is proposed for stochastic nonlinear systems with input quantization,prescribed performance and unmodeled dynamics.A new quantizer combining the advantages of hysteretic quantizer and uniform quantizer is used to process the input signal,the radial basis function neural networks(RBFNNs)are used to approximate the unknown smooth function,the unmodeled dynamics are processed with the help of dynamic signals,and a high gain observer is designed to estimate the unmeasurable state of the system.Hyperbolic tangent function and time-varying function are applied to the design of prescribed performance function.Based on DSC method,the second-order command filter is introduced and the compensation term is added in each step of traditional DSC,which can not only effectively avoid the "complexity explosion" problem in backstepping design,but also eliminate the filtering error caused by the first-order filter in traditional DSC.Through Lyapunov stability analysis,it is proved that all signals in the controlled system are semi-globally uniformly ultimately bounded(SGUUB)in the sense of probability.Finally,two simulation experiments are used to verify the effectiveness of the control scheme.(2)A new finite-time adaptive output feedback control strategy is proposed by constructing nonlinear command filter for stochastic nonlinear systems with unmodeled dynamics,prescribed performance,input dead-zone and actuator failure.The unknown function is approximated by RBFNNs,dynamic signals are introduced to process unmodeled dynamics,and the prescribed performance function is constructed by hyperbolic tangent function and time-varying function,so that the tracking error of the system converges well within the prescribed range of time-varying in the sense of probability.The unmeasurable state of the system is estimated by a high gain observer,and the input dead time model and actuator fault model are linearized into a composite input model.Based on the improved DSC method and stochastic finite time theory,a new virtual control law without"singularity" is designed to simplify the controller design and accelerate the convergence speed of the system.Through Lyapunov stability method,it is proved that all signals in the controlled system are semi-globally finite-time stability in probability(SGFSP).Finally,two simulation experiments are used to verify the effectiveness of the proposed control strategy.(3)A new decentralized adaptive neural output feedback DSC scheme based on command filter is proposed for nonstrict-feedback stochastic nonlinear interconnected systems with input quantization,prescribed performance,actuator failure and unmodeled dynamics.The unknown smooth function is approximated by RBFNNs,the unmeasurable state of the system is estimated by decentralized K filter,the measurable dynamic signal is introduced to offset the influence of unmodeled dynamic,and the hyperbolic tangent function and time-varying function are introduced into the construction of prescribed performance function.The property of Gaussian function is used to deal with non-strict feedback system,the hysteresis quantizer model and actuator fault model are linearized,and the smoothing function is introduced to compensate the influence of quantization and bounded stuck fault.Based on DSC method,the first-order command filter is used to replace the first-order filter in traditional DSC,and the error compensation term is introduced in each step of DSC design to eliminate the influence of filtering error on the system,improve the tracking performance of the system and simplify the design process.Through Lyapunov stability method,it is proved that all signals in the controlled system are SGUUB in the sense of probability.Finally,two simulation experiments are used to verify the effectiveness of the decentralized adaptive control method.